Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberCN1441387 A
Publication typeApplication
Application numberCN 02161115
Publication date10 Sep 2003
Filing date27 Nov 2002
Priority date27 Nov 2001
Also published asCN1187716C, CN1277239C, CN1294540C, CN1428742A, CN1438613A, EP1320264A2, EP1320264A3, EP1320264B1, EP2278808A1, EP2278808B1, US7446771, US20030128883
Publication number02161115.7, CN 02161115, CN 1441387 A, CN 1441387A, CN-A-1441387, CN02161115, CN02161115.7, CN1441387 A, CN1441387A
Inventors金道均, 郑锡润, 张义善, 禹相玉, 李信俊, 韩万镇, 张敬子
Applicant三星电子株式会社
Export CitationBiBTeX, EndNote, RefMan
External Links: SIPO, Espacenet
Method and device for coding and decoding oriented internal inserter
CN 1441387 A
Abstract  translated from Chinese
本发明提供一种用于对定向内插器进行编码和译码的方法和装置,该定向内插器指出了在时间轴上关键帧的位置以及在每一个关键帧中的对象的旋转。 The present invention provides a method and apparatus for the directional interpolator is used for encoding and decoding, the orientation interpolator indicates the rotational position of key frames on a time axis and the object in each keyframe of. 用于编码定向内插器的装置包括断点抽取器,从由输入到其中的定向内插器组成的第一动画路径中抽取最小数量的断点,能产生不大于所述第一动画路径与由抽取的断点所生成的第二动画间的预定误差极限;关键字数据编码器,编码从所述断点抽取器输入的关键字数据;关键字值数据编码器,通过生成旋转差分数据来编码从所述断点抽取器输入的关键字值数据,通过该操作,所述对象按几乎等于当前关键帧的旋转变换值和前关键帧的旋转变换值间的差值被旋转变换。 The coding orientation means interpolator including breakpoints extractor for extracting from a minimum number of break from a first animation path which is input to the directional interpolator composition, can produce no more than the first animation path and predetermined limit of errors by second animation drawn between breakpoints generated; key data encoder, from the key input data breakpoints extractor; key value data encoder to data by generating a differential rotation encoding key value data input from the break point extractor, and by this operation, the objects in almost equal to the current value of the difference between the rotational transformation value of the rotational transformation key frame and the previous key frame is between the rotation transformation.
Claims(57)  translated from Chinese
1、一种用于编码定向内插器的装置,所述定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,该装置包括: An interpolator apparatus for encoding the orientation, the orientation interpolator including key data indicating the object represents the position of key frames on a time axis of rotation of key value data, the apparatus comprising:
断点抽取器,从由输入到其中的定向内插器组成的第一动画路径抽取最小数量的断点,能产生不大于所述第一动画路径和由抽取的断点生成的第二动画间的预定误差极限的误差; Break extractor extracts a minimum number of break from a first animation path from the input to the directional interpolator wherein the composition, can produce no more than between the first animation path and a second animation generated by the extracted break predefined error error limit;
关键字数据编码器,编码从所述断点抽取器输入的关键字数据; Key data encoder encodes the input from the break point extractor keyword data;
关键字值数据编码器,通过生成旋转差分数据来编码从所述断点抽取器输入的关键字值数据,通过关键字值数据编码器,所述对象按几乎等于当前关键帧的旋转变换值和前关键帧的旋转变换值间的差值被旋转变换。 Rotational transformation value and the key value data encoder, by generating rotational differential data to said encoded key value data input from the break point extractor, by key value data encoder, said object by almost equal to the current key frame the difference between the rotational transformation value of the previous key frame is between the rotation transformation.
2、如权利要求1所述的装置,进一步包括: 2. The apparatus of claim 1, further comprising:
重新采样器,将所述第一动画路径采样成具有预定时间量间隔的多个预定部分,并输出包括重新采样关键字数据和重新采样关键字值数据的定向内插器;以及 Resampler, the first animation path into a sample having a plurality of predetermined portions of a predetermined amount of time interval, and outputs the resampled key data and including resampled key value data of the orientation interpolator; and
选择器,根据外部输入信号,将输入到其中的定向内插器输出到所述重新采样器或所述断点抽取器。 Selector, according to an external input signal, wherein the input to the directional interpolator to the resampler output or the breakpoint extractor.
3、如权利要求1所述的装置,进一步包括重新采样器,将所述第一动画路径采样成具有预定时间量间隔的多个预定部分,并输出包括重新采样关键字数据和重新采样关键字值数据的定向内插器, 3. The apparatus of claim 1, further comprising a resampler, the sample of the first animation path into a plurality of predetermined portions having a predetermined amount of time interval, and outputs the resampled key data and including resampling keywords directional interpolator value data,
其中所述断点抽取器从由从所述重新采样器输入的定向内插器组成的动画路径抽取断点。 Wherein the breakpoint Extractor break from animation path by directional interpolator input from the resampling composition.
4、如权利要求1所述的装置,进一步包括重新采样器,将由从所述断点抽取器抽取的定向内插器组成的动画路径采样成具有预定时间量间隔的多个预定部分,并将包括重新采样关键字数据和重新采样关键字值数据的定向内插器输出到所述关键字数据编码器和所述关键字值编码器。 4. The apparatus of claim 1, further comprising a resampler, animation path composed of the extracted from the break point extractor orientation interpolator into a sample having a plurality of predetermined portions of a predetermined amount of time interval, and including re-sampling and re-sampling key data within targeted keyword values interpolated data output to the key data encoder and the key value encoder.
5、如权利要求2至4中任何一个所述的装置,其中所述重新采样器将由定向内插器的关键字数据和关键字值数据组成的动画路径划分成具有预定时间量间隔的多个预定部分,并输出每个部分的终点作为将被编码的关键字数据,并输出存在于每个部分中的动画路径上的关键字值数据以作为将被编码的关键字值数据。 Animation key data and key value path data consisting of 5, as claimed in 2 to 4 apparatus according to any of claims, wherein the resampler by directional interpolator having a predetermined amount of time is divided into a plurality of intervals predetermined portion, and outputs the end of each section as the key data to be encoded, and outputs present on the animation path in each section of the key value data to be encoded as a key value data.
6、如权利要求1至4中任何一个所述的装置,其中所述断点抽取器包括: 6, such as 1 to 4, an apparatus according to any preceding claim, wherein the breakpoint extractor comprising:
线性内插器,抽取输入动画路径的开始路径点和结束路径点,在所述开始和结束路径点间选择路径点,以及利用所选择的路径点和所抽取的路径点来球性线性内插还没有被选择的其它路径点; Linear interpolator, the animation path input to extract the beginning path point and an end point of the path, the path between the start and end point selection route points, and using the selected path points and the extracted path points to the ball of the linear interpolation No other path points to be selected;
误差计算器,计算所述输入动画路径与由所述线性内插器使用内插所生成的内插动画路径间的误差; Error calculator to calculate the input animation path interpolation error between the interpolated animation path generated by the linear interpolation using Nene;
确定单元,抽取断点,通过该操作,能最小化所述输入动画路径与所述内插动画路径间的误差,并且如果所述相应误差不大于预定误差极限,则输出抽取的断点。 Determining means for extracting a breakpoint, through this operation, that minimizes the error of the input animation path and the interpolated animation path between the inside, and if the corresponding error is not greater than a predetermined error limit, then the output extracted breakpoint.
7、如权利要求6所述的装置,其中如果所述输入动画路径与由抽取断点所组成的所述内插动画路径间的误差大于所述预定误差极限,则线性内插器依次选择除从所述确定单元输入的断点外的所有路径点并在所选择的路径点上执行球性线性内插。 7. The apparatus of claim 6, wherein if the input animation path and composed by extracting break interpolation error between the animation path is greater than the predetermined error limit, the linear interpolator select inter determining all the paths from the point of breakpoints outside the unit input and performs linear interpolation of the ball on the selected path points.
8、如权利要求6所述的装置,其中基于构成它们的路径点的一个基准分量,所述误差计算器将所述输入动画路径和通过球性线性内插所生成的动画路径分割成多个预定部分,并通过测量每个部分的区域来计算所述输入动画路径和在每个部分中所生成的动画路径间的误差。 8. The apparatus of claim 6, wherein based on one reference component constituting their path points, the error calculator and the input animation path generated by the ball of the linearly interpolated animation path into a plurality of predetermined portion, and an error of said input animation path and an animation path generated in each section between the regions by measuring each part is calculated.
9、如权利要求1所述的装置,其中关键字数据编码器使用预定量化位量化从所述断点抽取器输入的关键字数据,通过在已量化关键字数据上执行预定DPCM操作来生成差分数据,以及编码所述差分数据。 9. The apparatus of claim 1, wherein the key data encoder using a predetermined quantization bit quantization of the breakpoint key data input from the extractor, by performing a predetermined DPCM operation on the quantized key data to generate a difference data, and encodes the differential data.
10、如权利要求1所述的装置,其中所述关键字数据编码器包括: 10. The apparatus of claim 1, wherein the key data encoder comprises:
第一量化器,使用预定量化位来量化定向内插器的关键字数据; A first quantizer, using predetermined quantization bits for quantizing interpolator targeting keyword data;
第一DPCM处理器,生成所述量化关键字数据的差分数据; A first DPCM processor, generates differential data of the quantized key data;
DND处理器,根据所述差分数据与其中的最大值和最小值间的关系,在所述差分数据上执行DND操作; DND processor, based on the relationship between the differential data and maximum and minimum values among them, perform DND operation on the differential data;
第一熵编码器,熵编码从所述DND处理器输入的所述差分数据。 First entropy encoder entropy encoding from the DND processor input to the differential data.
11、如权利要求1所述的装置,其中所述关键字数据编码器进一步包括线性关键字编码器,其识别和编码所有输入其中的关键字数据都线性增长的区域。 11. The apparatus of claim 1, wherein the key data encoder further comprises a linear key encoder, which identifies and encodes the input region wherein all of the key data are linear growth.
12、如权利要求10所述的装置,其中所述关键字数据编码器进一步包括: 12. The apparatus of claim 10, wherein the key data encoder further comprises:
移位器,获得在从所述第一DPCM处理器所输入的差分数据中具有最高频率的差分数据(模式)并从所述差分数据减去该模式;以及 Shifter, obtained with the highest frequency in the difference data from the first DPCM processor of the differential data input (mode) and subtracts the mode from the differential data; and
折叠处理器,将所述移位差分数据转换成正数或负数,以及DND处理器根据要求用于编码的位的数量,从所述移位器输入的差分数据、从所述折叠处理器输入的差分数据以及DND差分数据中选择一个,并输出选择的差分数据。 Folding processor, converts the shifted differential data into positive or negative, and the DND processor according to the requirements of the number of bits for encoding the differential data input from the shifter, the input from the folding processor differential data and differential data DND select one, and outputs the selected differential data.
13、如权利要求1所述的装置,其中所述关键字值编码器包括: 13. The apparatus of claim 1, wherein the key value encoder comprises:
旋转差分数据生成器,使用当前关键帧的旋转变换值和前关键帧的复原旋转变换值来生成旋转差分数据,所述旋转差分值被用来按几乎等于通过关键字值数据而应用到当前关键帧中的所述对象的旋转变换与通过关键字值数据而应用到前关键帧中的所述对象的旋转变换间的差值来旋转所述对象,并通过量化所述旋转差分值来输出旋转差分数据;以及 Rotational differential data generator using the rotational transformation value of the restored rotational transformation value of the current key frame and the previous key frame to generate a rotational difference data, the difference value is used to press the rotation is almost equal to the key value of the data by the application to the current key rotational transformation of the object frame and the difference between the keyword and the value applied to the data object before the keyframe rotation transformation between the object to rotate, and by quantizing the rotational differential value output rotation differential data; and
熵编码器,熵编码所述旋转差分数据。 Entropy encoder entropy encoding the rotational differential data.
14、如权利要求13所述的装置,其中所述旋转差分数据生成器包括: 14. The apparatus of claim 13, wherein said rotational differential data generator comprises:
量化器,通过量化所述旋转差分值的三个分量值来生成旋转差分数据; Quantizer to generate rotational differential data by quantizing the rotational differential value of the three component values;
量化数据调节器,调节输入到其中的旋转差分数据的三个分量值; Three quantized data controller, to adjust the input to which the rotating component values of the differential data;
逆量化器,逆量化所述调节分量值; An inverse quantizer, an inverse quantizing the adjusted component values;
旋转差分值复原器,使用所述三个逆量化分量值来复原未被量化的一个分量值,从而生成复原旋转差分值;以及 Rotational differential value restorer, using the three inversely quantized component values are not to recover a quantized component values, thereby generating a restored rotational differential value; and
误差测量单元,测量输入到所述量化器中的旋转差分值和所述复原旋转差分值间的误差,并输出具有已调节分量值的旋转差分数据,以便能最小化该误差。 Error measuring means for measuring the input to the quantizer and the restored rotational differential value error between the rotational differential value and outputs adjusted rotational differential having component values of the data, so as to minimize the error.
15、如权利要求13所述的装置,其中所述旋转差分数据生成器包括: 15. The apparatus of claim 13, wherein said rotational differential data generator comprises:
第一四元数乘法器,使用所述当前关键帧的旋转变换值和所述前关键帧的复原旋转变换值来生成所述旋转差分值; Restored rotational transformation value of a first quaternion multiplier, using the rotational transformation value of the current key frame and the front of the key frame to generate rotational difference value;
量化器,通过量化所述旋转差分值来生成旋转差分数据; A quantizer quantizing the rotational differential value by generating rotational differential data;
逆量化器,通过逆量化所述旋转差分数据来生成复原旋转差分值;以及 An inverse quantizer, the differential rotation data restored by the inverse quantization to generate rotational difference value; and
第二四元数乘法器,通过将所述复原旋转差分值四元数乘以所述前关键帧的旋转变换值来生成所述当前关键帧的复原旋转差分值。 The second quaternion multiplier, by restoring the quaternion rotation difference value by multiplying the rotational transformation value of the front of the key frame to generate current restored rotational differential value of key frames.
16、如权利要求13所述的装置,其中所述旋转差分数据生成器包括: 16. The apparatus of claim 13, wherein said rotational differential data generator comprises:
旋转方向误差检测器,基于所述当前关键帧的旋转变换值以及所述前关键帧的复原旋转变换值来检测是否已经出现旋转方向误差,以致所述对象的原始旋转方向与所述对象的译码旋转方向相反; Rotation direction error detector, based on the rotational transformation value of the restored rotational transformation value of the current key frame and key frame of the front to detect whether the rotation direction error has occurred, so that the original rotation direction of translation of the object with the object the opposite direction of rotation code;
旋转方向校正器,调节所述旋转差分值以便所述对象的译码旋转方向能与所述对象的原始旋转方向相同;以及 Rotation direction corrector, adjusting the rotational differential value so that the decoded rotation direction of the rotation of the object with the original object can be the same direction; and
旋转方向选择器,根据从所述旋转方向误差检测器输入的检测结果,选择所述旋转差分值或从所述旋转方向校正器输入的所述旋转差分值以作为将被量化的差分数据。 Rotation direction selector, based on the detection result from the rotation direction error detector input, selecting the rotational differential value or input from the rotation direction corrector to the rotation difference value as the quantized differential data.
17、一种用于编码定向内插器的装置,所述定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,该装置包括: 17, an interpolator means for encoding the orientation, the orientation interpolator including key data indicating the object represents the position of key frames on a time axis of rotation of key value data, the apparatus comprising:
重新采样器,将由输入定向内插器组成的动画路径采样成具有预定时间量间隔的多个预定部分,并输出包括重新采样关键字数据和重新采样关键字值数据的定向内插器; Resampler, by the input orientation interpolator consisting of the sampled animation path into a plurality of portions having a predetermined amount of a predetermined time interval, and outputs the resampled key data and including resampled key value data of the orientation interpolator;
关键字数据编码器,编码从所述重新采样器输入的关键字数据;以及 Key data encoder encodes the input from the resampler keyword data; and
关键字值数据编码器,生成旋转差分值,所述旋转差分值用来按几乎等于通过当前关键帧的关键字值数据而应用到对象的旋转变换和通过前关键帧的关键字值数据而应用到所述对象的旋转变换间的差值来旋转所述对象,从而编码从所述重新采样器所输入的关键字值数据。 Key value data encoder to generate rotational difference value, the rotational differential value used by almost equal to the key value data by the current key frame and rotational transformation applied to the object by key value data and key frame and before application to the difference between the rotational transformation of the object to the rotation of the object, thereby encoding key value data from the re-sampler input.
18、如权利要求17所述的装置,其中所述重新采样器将由定向内插器的关键字数据和关键字值数据所组成的动画路径划分成具有预定时间量间隔的多个预定部分,并输出每个部分的终点以作为将被编码的关键字数据,并输出存在于每个部分中的动画路径上的关键字值数据以作为将被编码的关键字值数据。 Animation key data and key value path data 18. The apparatus of claim 17, wherein the resampler by orientation interpolator consisting of divided into a plurality of predetermined portions having a predetermined amount of time interval, and the output end of each section as the key data to be encoded, and outputs present on the animation path in each section of the key value data to be encoded as a key value data.
19、如权利要求17所述的装置,其中所述关键字数据编码器包括线性关键字编码器,其识别和编码所有输入其中的关键字数据都线性增长的区域。 19. The apparatus of claim 17, wherein the key data encoder comprises a linear key encoder, which identifies and encodes the input region wherein all of the key data are linear growth.
20、如权利要求17所述的装置,其中所述关键字值编码器包括: 20. The apparatus of claim 17, wherein the key value encoder comprises:
旋转差分数据生成器,使用当前关键帧的旋转变换值和前关键帧的复原旋转变换值来生成旋转差分值,所述旋转差分值被用来按几乎等于通过关键字值数据而应用到当前关键帧中的所述对象的旋转变换和通过关键字值数据而应用到前关键帧中的所述对象的旋转变换间的差值来旋转所述对象,并通过量化所述旋转差分值来输出旋转差分数据;以及 Rotational differential data generator using the rotational transformation value of the restored rotational transformation value of the current key frame and the previous key frame to generate a rotational differential value, the rotation difference value is used by almost equal to the key value of the data by the application to the current key the difference between the rotation transformation of the object in the frame and through the key value data applied to the object before the keyframe between the rotation transformation to rotate the object, and by quantifying the differential value of the rotating output rotation differential data; and
熵编码器,熵编码所述旋转差分数据。 Entropy encoder entropy encoding the rotational differential data.
21、如权利要求20所述的装置,其中所述旋转差分数据生成器包括: 21. The apparatus of claim 20, wherein said rotational differential data generator comprises:
量化器,通过量化所述旋转差分值的三个分量值来生成旋转差分数据; Quantizer to generate rotational differential data by quantizing the rotational differential value of the three component values;
量化数据调节器,调节输入到其中的旋转差分数据的三个分量值; Three quantized data controller, to adjust the input to which the rotating component values of the differential data;
逆量化器,逆量化所述调节分量值; An inverse quantizer, an inverse quantizing the adjusted component values;
旋转差分值复原器,使用所述三个逆量化分量值来复原未被量化的一个分量值,从而生成复原旋转差分值;以及 Rotational differential value restorer, using the three inversely quantized component values are not to recover a quantized component values, thereby generating a restored rotational differential value; and
误差测量单元,测量输入到所述量化器中的旋转差分值和所述复原旋转差分值间的误差,并输出具有已调节分量值的旋转差分数据以便能最小化该误差。 Error measuring means for measuring the input to the quantizer and the restored rotational differential value error between the rotational differential value and outputs rotational differential data having adjusted component values so as to minimize the error.
22、如权利要求20所述的装置,其中所述旋转差分数据生成器包括: 22. The apparatus of claim 20, wherein said rotational differential data generator comprises:
第一四元数乘法器,使用所述当前关键帧的旋转变换值和所述前关键帧的复原旋转变换值来生成所述旋转差分值; Restored rotational transformation value of a first quaternion multiplier, using the rotational transformation value of the current key frame and the front of the key frame to generate rotational difference value;
量化器,通过量化所述旋转差分值来生成旋转差分数据; A quantizer quantizing the rotational differential value by generating rotational differential data;
逆量化器,通过逆量化所述旋转差分数据来生成复原旋转差分值;以及 An inverse quantizer, the differential rotation data restored by the inverse quantization to generate rotational difference value; and
第二四元数乘法器,通过将所述复原旋转差分值四元数乘以所述前关键帧的旋转变换值来生成所述当前关键帧的复原旋转差分值。 The second quaternion multiplier, by restoring the quaternion rotation difference value by multiplying the rotational transformation value of the front of the key frame to generate current restored rotational differential value of key frames.
23、如权利要求20所述的装置,其中所述旋转差分数据生成器包括: 23. The apparatus of claim 20, wherein said rotational differential data generator comprises:
旋转方向误差检测器,基于所述当前关键帧的旋转变换值以及所述前关键帧的复原旋转变换值来检测是否已经出现旋转方向误差,以致所述对象的原始旋转方向与所述对象的译码旋转方向相反; Rotation direction error detector, based on the rotational transformation value of the restored rotational transformation value of the current key frame and key frame of the front to detect whether the rotation direction error has occurred, so that the original rotation direction of translation of the object with the object the opposite direction of rotation code;
旋转方向校正器,调节所述旋转差分值,以便所述对象的译码旋转方向能与所述对象的原始旋转方向相同;以及 Rotation direction corrector, adjusting the rotational differential value so that the decoded rotation direction of the object with the original rotation direction of the object of the same; and
旋转方向选择器,根据从所述旋转方向误差检测器所输入的检测结果,选择所述旋转差分值或从所述旋转方向校正器输入的所述旋转差分值以作为将被量化的差分数据。 Rotation direction selector, based on the detection result from the rotation direction error detector input, selecting the rotational differential value or input from the rotation direction corrector to the rotation difference value as the quantized differential data.
24、一种用于译码位流的装置,其中将定向内插器编码成该位流,该定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,该装置包括: 24. An apparatus for decoding a bitstream, wherein the orientation interpolator is encoded into the bit stream, in that rotational orientation interpolator including key data indicates the location of key frames on a time axis indicating object key value data, the apparatus comprising:
关键字数据译码器,译码来自输入位流的关键字数据; Key data decoder decodes key data from an input bit stream;
关键字值数据译码器,译码来自所述输入位流的关键字值数据;以及 Key value data decoder decodes the input bit stream from the key value data; and
定向内插器合成器,通过合成已译码的关键字值数据和使用所述译码关键字值数据而球性线性内插的关键字值数据来生成定向内插器。 Interpolation of orienter synthesizer, key value data and key value data decoded by synthesis using the decoded key value data and linear interpolation of the ball to generate the directional interpolator.
25、如权利要求24所述的装置,其中如果不存在与当前正在受到定向内插器合成的关键字数据相对应的已译码的关键字值数据,则所述定向内插器合成器利用与先前合成的关键字数据相对应的译码关键字值数据以及与下一个将被合成的关键字数据相对应的译码关键字值数据来内插与当前正在受到定向内插器合成的所述关键字数据相对应的关键字值数据。 25. The apparatus of claim 24, wherein if the corresponding key value data are currently being subjected to orientation interpolator synthesized within the decoded key data is not present, then the orientation interpolator synthesizer uses a previously synthesized key data corresponding to the decoding key data and key value data corresponding to the decoding key value data with the next to be synthesized by interpolation and are in the interior orientation interpolator by synthesizing described key data corresponding to key value data.
26、如权利要求24所述的装置,其中所述关键字数据译码器包括: 26. The apparatus of claim 24, wherein the key data decoder comprises:
熵译码器,通过熵译码所述输入位流来生成差分数据; Entropy decoder, the input bit stream by entropy coding the difference data is generated;
逆DPCM处理器,通过在所述差分数据上执行预定逆DPCM操作来生成量化关键字数据; An inverse DPCM processor, by performing a predetermined inverse DPCM operation on the differential data to generate the quantized key data;
逆量化器,通过逆量化所述量化关键字数据来生成复原关键字数据。 An inverse quantizer, the quantization by the inverse quantization to generate restored key data is key data.
27、如权利要求24所述的装置,其中所述关键字值译码器包括: 27. The apparatus of claim 24, wherein the key value decoder includes:
熵译码器,通过熵译码来自所述位流的关键字值数据来生成循环的DPCM旋转差分数据或量化旋转差分数据; Entropy decoder, by entropy decoding key value data from the bit stream to generate the loop DPCM data or quantized rotational differential rotational differential data;
逆循环DPCM运算符,通过在从所述熵译码器输入的旋转差分数据上依据从所述位流译码的DPCM操作的次数来执行逆循环DPCM操作,以生成量化旋转差分数据; Reverse cycle DPCM operator, by rotating in the differential data input from the entropy decoder according to the number of the bit stream from the decoded DPCM operation to perform inverse DPCM loop operation, in order to generate a quantized rotational differential data;
逆量化器,通过逆量化所述量化旋转差分数据来生成旋转差分数据,所述旋转差分数据被用来按几乎等于通过每个关键帧的四元数关键字值数据而应用到对象的旋转变换间的差值来旋转所述对象;以及 An inverse quantizer, the inverse quantized rotational differential data by quantizing the rotational differential data to generate a differential rotation data is used by almost equal by quaternion key value data of each keyframe and rotational transformation applied to the object to the difference between the rotation of the object; and
四元数乘法器,通过将当前关键帧的旋转差分值四元数乘以前关键帧的复原旋转变换差分值来生成所述当前关键帧的旋转变换值。 Quaternion multiplier, by the difference between the current rotation keyframes scores quaternion multiplication previous keyframe recovery rotation transformation to generate a differential value of the current value of the rotational transform keyframes.
28、一种用于编码定向内插器的方法,所述定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,所述方法包括: 28. A directional interpolator coding method for the directional interpolator including a representation of data and objects that represent key position keyframes in the timeline of the rotating key value data, the method comprising:
(b)通过从由所述定向内插器组成的第一动画路径抽取最少量的断点来生成将被编码的关键字数据和关键字值数据,能产生不大于所述第一动画路径和将由抽取的断点生成的第二动画间的预定误差极限的误差; (B) by extracting the minimum amount of the breakpoint from a first animation path by said directional interpolator composition generates key data to be encoded, and key value data, to produce no more than the first animation path and predetermined error error limits will be drawn among the second generation of animation breakpoint;
(d)编码在步骤(b)中生成的所述关键字数据; (D) encoding the key data generated in step (b); and
(e)通过生成旋转差分数据来编码在步骤(b)中生成的关键字值数据,通过该操作,所述对象按几乎等于当前关键帧的旋转变换值和前关键帧的旋转变换值间的差值旋转变换。 (E) by generating rotational differential data are encoded in step (b) the key value data generated by this operation, the objects in almost equal between the rotational transformation value of the current rotational transformation value of the key frame and the previous key frame the difference between the rotation transformation.
29、如权利要求28所述的方法,其中步骤(b)包括: 29. The method of claim 28, wherein step (b) comprises:
(b1)抽取所述第一动画路径的开始路径点和结束路径点; (B1) extracting the start point of the first path animation path and end waypoints;
(b2)利用所选择的路径点和所提取的路径点来依次选择开始和结束路径点之间的路径点,并球性线性内插仍未被选择的的其它路径点; (B2) using the selected path points and the extracted path points to select the start and end point of the path between the path points and the ball of the other linear interpolation point of the path has not yet been selected; and
(b3)计算所述第一动画路径和通过在步骤(b2)中内插生成的第二动画路径间的误差; (B3) calculating an error of the first animation path and a second animation path by step (b2) is inserted in between the inner generated;
(b)抽取断点,通过该操作,能最小化所述第一动画路径和所述第二动画路径间的误差,检查所述相应误差是否不大于预定误差极限,并确定将被编码的关键字数据和关键字值数据。 Key (b) extracting the breakpoint, this operation can minimize the error of the first animation path and the second animation path between the respective error checking is not greater than a predetermined error limit, and determines to be encoded word data and key value data.
30、如权利要求28所述的方法,进一步包括在步骤(b)之前的(a),通过将所述第一动画路径采样成具有预定时间量间隔的多个预定部分来生成包括重新采样关键字数据和重新采样关键字值数据的定向内插器, 30. The method of claim 28 comprising re-sampling key claim, further comprising prior to step (b), (a), by sampling the first animation path into a predetermined amount of portion having a plurality of predetermined time intervals to generate word data and resampled key value data of the orientation interpolator,
其中在步骤(b)中,从由在步骤(a)中所生成的所述定向内插器构成的动画路径抽取断点。 Wherein in step (b), and extracted from an animation path by the breakpoint within the orientation generated in step (a) is composed of an interpolator.
31、如权利要求30所述的方法,其中在步骤(a)中,所述第一动画路径被分割成具有预定时间量间隔的多个预定部分,每个部分的终点被设置成采样关键字数据,以及存在于每个部分中的第一动画路径上的关键字值数据被设置成采样关键字值数据。 31. The method of claim 30, wherein in step (a), the first animation path is divided into portions having a predetermined amount of a plurality of predetermined time interval, the end of each section is arranged to sample keywords data, and key value data is present in each part of the first animation path is arranged on the sampled key value data.
32、如权利要求28所述的方法,进一步包括在步骤(b)之后的(c),通过将利用所抽取的断点而组成的所述第二动画路径采样成具有预定时间量间隔的多个预定部分来生成将被编码的关键字数据和关键字值数据。 32. The method of claim 28, further comprising after step (b), (c), by the use of the breakpoint and the composition of the extracted second animation path into a sample having a volume of more than a predetermined time interval to generate a predetermined portion of the encoded key data and key value data.
33、如权利要求28所述的方法,其中步骤(d)包括: 33. The method of claim 28, wherein step (d) comprises:
利用预定数量的量化位量化所述关键字数据; Using a predetermined number of quantization bits of the quantized key data;
通过在量化数据上执行预定DPCM操作来生成差分数据;以及 Generating differential data by performing a predetermined DPCM operation on quantized data; and
熵编码所述差分数据。 Entropy encoding the differential data.
34、如权利要求28所述的方法,其中步骤(e)包括: 34. The method of claim 28, wherein step (e) comprises:
(e1)使用所述当前关键帧的一旋转变换值以及所述前关键帧的复原旋转变换值来生成旋转差分值,所述旋转差分值用来按几乎等于通过当前和前关键帧的关键字值数据而应用到该对象的旋转变换间的差值来旋转所述对象,以及通过量化所述旋转差分值来生成旋转差分数据;以及 A rotational transformation value (e1) using the current key frame and the restored rotational transformation value of the front frame to generate a key rotation difference value, the rotational differential value used by almost equal to the current and previous keyframe by keyword value data and rotational transformation applied to the difference between the object to rotate said object, and generating rotational differential data by quantizing the rotational differential value; and
(e2)熵编码所述旋转差分数据。 (E2) entropy encoding of the differential rotation data.
35、如权利要求34所述的方法,其中步骤(e1)包括: 35. The method of claim 34, wherein step (e1) comprises:
(e11)通过量化所述旋转差分值的三个分量值来生成旋转差分数据; (E11) by quantifying the difference value of the rotating three component values to generate a differential rotation data;
(e12)调节所述旋转差分数据的三个分量值; Three (e12) adjusting the rotation differential data component values;
(e13)逆量化所述调节分量值; (E13) inverse quantization of the adjustment component values;
(e14)通过利用所述三个逆量化分量值而复原一个分量值来生成复原旋转差分值; (E14) by using the three inversely quantized component values and restoring a recovery rotating component values to generate a difference value;
(e15)测量所述旋转差分值和所述复原旋转差分值间的误差并确定具有已调节分量值的旋转差分数据,以便所述误差能被最小化为将被熵编码的旋转差分数据。 (E15) measuring the rotational differential value and the restored rotational differential value between the error and determine having adjusted component values of the rotational differential data so that the error can be minimized to be rotational differential data entropy-encoded.
36、如权利要求34所述的方法,其中步骤(e1)包括: 36. The method of claim 34, wherein step (e1) comprises:
(e11)使用所述当前关键帧的旋转变换值和所述前关键帧的复原旋转变换值来生成所述旋转差分值; Restoring rotational transformation value (e11) using the current key frame and the front frame of the rotational transformation value of the key to generate the rotational differential value;
(e12)通过量化所述旋转差分值来生成旋转差分数据; (E12) by quantifying the rotating differential value to generate a differential rotation data;
(e13)通过逆量化所述旋转差分数据来生成复原旋转差分值; (E13) inverse quantization by the rotational differential data to generate the restored rotational differential value;
(e14)通过将所述复原旋转差分值四元数乘以所述前关键帧的旋转变换值来生成所述当前关键帧的复原旋转变换值。 (E14) by said return rotational quaternion difference value is multiplied by the key frame before generating the rotational transformation value of the current rotational transformation value of the restored key frames.
37、如权利要求34所述的方法,其中步骤(e1)包括: 37. The method of claim 34, wherein step (e1) comprises:
(e11)基于所述当前关键帧的旋转变换值以及所述前关键帧的复原旋转变换值来检测是否已经发生旋转方向误差以致所述对象的原始旋转方向与所述对象的译码旋转方向相反; (E11) based on the rotational transformation value of the restored rotational transformation value of the current keyframe and the key frame is detected before decoding the original rotation direction is the rotation direction error has occurred so that the object and the rotational direction opposite to the object ;
(e12)调节所述旋转差分值,以便所述对象的原始旋转方向能与所述对象的译码旋转方向相同;以及 (E12) adjusting the rotational differential value so that the original rotation direction of the object with the decoding of the same rotation direction of the object; and
(e13)根据在步骤(e11)中执行的检测结果,选择所述旋转差分值或在步骤(e12)中的所述调节旋转差分值作为将被量化的差分数据。 (E13) in accordance with the detection result performed in step (e11), select the rotation or the difference value in step (e12) adjusting the rotational differential value in as differential data to be quantized.
38、一种用于编码定向内插器的方法,所述定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,所述方法包括: 38. A method for interpolator for coding orientation, the orientation interpolator including key data indicating the object represents the position of key frames on a time axis of rotation of key value data, the method comprising:
(a)将由所述定向内插器所构成的动画路径采样成具有预定时间量间隔的多个预定部分,从而生成包括重新采样关键字数据和重新采样关键字值数据的定向内插器; Animation path (a) by the orientation interpolator consisting of sampled into portions having a predetermined amount of a plurality of predetermined time intervals, thereby generating the resampled key data and including resampled key value data of the orientation interpolator;
(d)减小在步骤(a)中采样的所述关键字数据的范围并编码所述关键字数据;以及 (D) reducing the range of samples in step (a), the key data and said coded key data; and
(e)通过生成和编码旋转差分值来编码在步骤(a)中采样的所述关键字值数据,所述旋转差分值被用来按几乎等于通过当前关键帧的关键字值数据应用到所述对象的旋转变换和通过前关键帧的关键字值数据而应用到所述对象的旋转变换间的差值来旋转所述对象。 (E) by generating rotational differential value and encoding in the encoding step (a) in the sampled key value data, the rotation difference value is used by almost equal to the current keyframe by key value data is applied to the the difference between the object and the rotation transformation described by key value data before keyframes and applied to the object's rotation transformation between the object to rotate.
39、如权利要求38所述的方法,其中步骤(a)包括将所述动画路径分割成具有预定时间量间隔的多个预定部分,将每个部分的终点设置成所述采样关键字数据,以及将存在于在每个部分的所述第一动画路径上的关键字值数据设置为所述采样关键字值数据。 39. The method of claim 38, wherein step (a) comprises dividing the animation path into a predetermined amount of portion having a plurality of predetermined time interval, the end of each portion of the sample is provided to the keyword data, and key value data will exist in each section on the first animation path is set to the key value data sampled.
40、如权利要求38所述的方法,其中所述步骤(e)包括: 40. The method of claim 38, wherein said step (e) comprises:
(e1)使用所述当前关键帧的一旋转变换值以及所述前关键帧的复原旋转变换值来生成旋转差分值,所述旋转差分值用来按几乎等于通过当前和前关键帧的关键字值数据而应用到该对象的旋转变换间的差值来旋转所述对象,以及通过量化所述旋转差分值来生成旋转差分数据;以及 A rotational transformation value (e1) using the current key frame and the restored rotational transformation value of the front frame to generate a key rotation difference value, the rotational differential value used by almost equal to the current and previous keyframe by keyword value data and rotational transformation applied to the difference between the object to rotate said object, and generating rotational differential data by quantizing the rotational differential value; and
(e2)熵编码所述旋转差分数据。 (E2) entropy encoding of the differential rotation data.
41、如权利要求40所述的方法,其中步骤(e1)包括: 41. The method of claim 40, wherein step (e1) comprises:
(e11)通过量化所述旋转差分值的三个分量值来生成旋转差分数据; (E11) by quantifying the difference value of the rotating three component values to generate a differential rotation data;
(e12)调节所述旋转差分数据的三个分量值; Three (e12) adjusting the rotation differential data component values;
(e13)逆量化所述调节分量值; (E13) inverse quantization of the adjustment component values;
(e14)通过利用所述三个逆量化分量值而复原一个分量值来生成复原旋转差分值; (E14) by using the three inversely quantized component values and restoring a recovery rotating component values to generate a difference value;
(e15)测量所述旋转差分值和所述复原旋转差分值间的误差并确定具有已调节分量值的旋转差分数据,以便所述误差能被最小化为将被熵编码的旋转差分数据。 (E15) measuring the rotational differential value and the restored rotational differential value between the error and determine having adjusted component values of the rotational differential data so that the error can be minimized to be rotational differential data entropy-encoded.
42、如权利要求40所述的方法,其中步骤(e1)包括: 42. The method of claim 40, wherein step (e1) comprises:
(e11)使用所述当前关键帧的旋转变换值和所述前关键帧的复原旋转变换值来生成所述旋转差分值; Restoring rotational transformation value (e11) using the current key frame and the front frame of the rotational transformation value of the key to generate the rotational differential value;
(e12)通过量化所述旋转差分值来生成旋转差分数据; (E12) by quantifying the rotating differential value to generate a differential rotation data;
(e13)通过逆量化所述旋转差分数据来生成复原旋转差分值; (E13) inverse quantization by the rotational differential data to generate the restored rotational differential value;
(e14)通过将所述复原旋转差分值四元数乘以所述前关键帧的旋转变换值来生成所述当前关键帧的复原旋转变换值。 (E14) by said return rotational quaternion difference value is multiplied by the key frame before generating the rotational transformation value of the current rotational transformation value of the restored key frames.
43、如权利要求40所述的方法,其中步骤(e1)包括: 43. The method of claim 40, wherein step (e1) comprises:
(e11)基于所述当前关键帧的旋转变换值以及所述前关键帧的复原旋转变换值来检测是否已经发生旋转方向误差以致所述对象的原始旋转方向与所述对象的译码旋转方向相反; (E11) based on the rotational transformation value of the restored rotational transformation value of the current keyframe and the key frame is detected before decoding the original rotation direction is the rotation direction error has occurred so that the object and the rotational direction opposite to the object ;
(e12)调节所述旋转差分值,以便所述对象的原始旋转方向能与所述对象的译码旋转方向相同;以及 (E12) adjusting the rotational differential value so that the original rotation direction of the object with the decoding of the same rotation direction of the object; and
(e13)根据在步骤(e11)中执行的检测结果,选择所述旋转差分值或在步骤(e12)中的所述调节旋转差分值作为将被量化的差分数据。 (E13) in accordance with the detection result performed in step (e11), select the rotation or the difference value in step (e12) adjusting the rotational differential value in as differential data to be quantized.
44、一种计算机可读记录介质,记录可实现权利要求28至43中任何一个的方法的计算机可读程序代码。 44. A computer-readable recording medium recording a computer 28 to 43 can be realized in a method as claimed in any one of claim-readable program code.
45、一种用于译码位流的方法,定向内插器被编码成该位流,该定向内插器包括表示在时间轴上关键帧的位置的关键字数据和表示对象的旋转的关键字值数据,所述方法包括: 45. A method for decoding a bit stream, the orientation interpolator is encoded into the bit stream, the rotation of the orientation interpolator including key data and key object that represents the position of key frames on the timeline representation value data word, said method comprising:
(a)译码来自输入位流的关键字数据; (A) decoding the input bit stream from the key data;
(b)译码来自所述输入位流的关键字值数据;以及 (B) decoding the input bit stream from the key value data; and
(c)通过合成已译码的关键字值数据和利用所述已译码关键字值数据所球性线性内插的关键字值数据来生成定向内插器。 (C) by synthesizing the decoded key value data and use the decoded key value data for the key value data of the ball linear interpolation to generate internal directional interpolator.
46、如权利要求45所述的方法,其中在步骤(c)中,如果没有与当前正在受到定向内插器合成的关键字数据相对应的译码关键字值数据,则利用与先前已合成的关键字数据相对应的译码关键字值数据以及与下一个将被合成的关键字数据相对应的已译码关键字值数据来内插与当前正在受到定向内插器合成的所述关键字数据相对应的关键字值数据。 46. The method of claim 45, wherein in step (c), if no current is being synthesized within the orientation interpolator key data corresponding to the decoding key value data, the use of the previously synthesized The key data corresponding decoding key value data, and will be combined with the next keyword data corresponding to the decoded key value data interpolated with the key inserted in the directed synthesis is currently being word data corresponding to the key value data.
47、如权利要求45所述的方法,其中步骤(a)包括 47. The method of claim 45, wherein step (a) comprises
通过熵译码所述输入位流来生成差分数据; Said input bit stream by entropy coding the difference data is generated;
通过在所述差分数据上执行预定DPCM操作和逆DND操作来生成量化关键字数据; Generating quantized key data by performing a predetermined DPCM operation and an inverse DND operation on the differential data;
通过逆量化所述量化关键字数据来生成复原关键字数据。 The quantized key data by inverse quantization to generate restored key data.
48、如权利要求45所述的方法,其中步骤(b)包括: 48. The method of claim 45, wherein step (b) comprises:
(b)通过熵译码来自所述位流的关键字值数据来生成循环DPCM旋转差分数据或量化旋转差分数据; Key value data (b) by the entropy decoding from the bit stream to generate a loop DPCM data or quantized rotational differential rotational differential data;
(b)通过在从所述熵译码器输入的旋转差分数据上依据从所述位流译码的DPCM操作的次数来执行逆循环DPCM操作以生成旋转差分数据; (B) by the rotation of the differential data input from the entropy decoder according to the number of times of decoding the bit stream to perform inverse DPCM operation DPCM loop operates to generate rotational differential data;
(b)通过逆量化所述量化旋转差分数据来生成旋转差分值,所述旋转差分值被用来按几乎等于通过每个关键帧的四元数关键字值数据而应用到对象的旋转变换间的差值来旋转所述对象;以及 Inter (b) by inverse quantization of the quantized data to generate differential rotation of the rotation differential value, the rotating difference value is used by almost equal by quaternion key value data for each keyframe and applied to the object's rotation transformation The difference to the rotation of the object; and
(b)通过将当前关键帧的旋转差分值四元数乘以前关键帧的译码旋转变换差分值来生成所述当前关键帧的旋转变换值。 (B) the difference between the current rotation keyframe quaternion score by multiplying the previous keyframe decoded difference value to generate a rotational transform rotational transformation value of the current key frame.
49、一种计算机可读记录介质,记录可实现权利要求45至49中任何一个的方法的计算机可读程序代码。 49. A computer-readable recording medium, the recording can be achieved in 45 to 49 The method of any one of the computer-readable program code claims.
50、一种位流,定向内插器被编码成该位流,该定向内插器包括表示在时间轴上关键帧位置的关键字数据和表示对象的旋转的关键字值数据,所述位流包括: 50, a bit stream is encoded in the directional interpolator into the bit stream, within the targeted keywords interposer including a representation of data and objects that represent the position keyframes on the timeline of the rotating key value data, the bit stream include:
关键字数据编码/译码信息,关键字数据和译码所述关键字数据所需的信息被编码成该信息;以及 Key data encoding / decoding information, key data and decode the information key data is encoded into the desired information;
关键字值数据编码/译码信息,关键字值数据和译码所述关键字值数据所需的信息被编码成该信息, Keyword value data encoding / decoding information, the information needed to decode the key value data and key value data to be encoded into the message,
其中所述关键字数据编码/译码信息包括: Wherein the key data encoding / decoding information comprises:
逆DND操作信息,包括表示将在差分数据上执行的逆DND的多个预定周期的逆DND的次数,所述差分数据是通过熵译码所述位流生成的,以便扩展所述差分数据和用在每个逆DND操作周期中的差分数据中的最大和最小值的范围; Inverse DND operation information includes information indicating the number of times the predetermined period of a plurality of difference data in the execution of an inverse DND inverse DND, and the difference data is a bit stream generated by the entropy coding, to extend the differential data and used in each cycle of inverse DND operation of difference data in the range of maximum and minimum values;
第一逆DPCM操作信息,包括将在所述逆DND差分数据上执行的逆DPCM操作的次数,以便将逆DND差分数据转换成用于每个逆DPCM操作周期的量化关键字数据和内部关键字数据;以及 First inverse DPCM operation information including the number of inverse DPCM operation to be performed on the inverse DND differential data, the inverse DND to convert differential data into quantized key data and the internal key for each of the inverse DPCM operation cycles data; and
第一逆量化信息,用在逆量化中,以便通过逆量化所述量化关键字数据来生成复原关键字数据;以及 First inverse quantization information used in inverse quantization to the quantized key data by inverse quantization to generate restored key data; and
关键字数据编码/译码信息,包括: Key data encoding / decoding information, comprising:
通过量化旋转差分值熵编码的旋转差分数据,所述旋转差分值被用来按几乎等于通过每个关键帧的四元数关键字值数据而应用到该对象的旋转变换间的差值来旋转所述对象; The difference between the rotational differential value by quantizing the rotational differential data entropy-encoded, the rotation difference value is used by almost equal by quaternion key value data of each keyframe and applied to the object to rotate the rotation transformation between said object;
熵译码信息,包括表示将在所述旋转差分数据上执行的熵译码方法的熵译码模式; Entropy coding information including entropy coding mode indicates the method of entropy coding that will be executed on the differential rotation data;
逆循环DPCM操作信息,包括逆循环DPCM操作的次数,表示是否将在熵译码旋转差分数据上依据所述熵译码模式执行逆循环DPCM操作;以及 Reverse circulation DPCM operation information including the number of reverse circulation DPCM operation indicating whether or inverse DPCM loop operating in accordance with the entropy coding mode on entropy coding rotating differential data; and
第二逆量化信息,包括多个预定逆量化位,用来通过逆量化所述量化关键字值数据复原原始关键字值数据。 The second inverse quantization information, including a plurality of predetermined inverse quantization bits to the quantized key value data restore the original key value data by inverse quantization.
51、如权利要求50所述的位流,其中所述逆DND操作信息进一步包括表示是否将在经逆DND操作的差分数据上执行上移操作的标志位。 51. The bitstream of claim 50, wherein the inverse DND operation information further comprises a flag indicating whether the shifting operation performed on the difference data by an inverse DND operation.
52、如权利要求50所述的位流,其中所述第一逆量化信息包括当逆量化所述量化关键字数据时所使用的逆量化位大小以及在量化关键字数据中的最大和最小值。 52. The bitstream of claim 50, wherein the first inverse quantization information comprises an inverse quantization of the quantization when inverse quantization bit size and the key data used in the maximum and minimum values of the quantized key data .
53、如权利要求52所述的位流,其中调节在所述量化关键字数据中的最大和最小值,以便最小化所述量化关键字数据的量化误差。 53. The bitstream of claim 52, wherein adjusting the maximum and minimum values of the quantized key data in order to minimize the quantization error of the quantized key data.
54、如权利要求50所述的位流,其中所述关键字数据编码/译码信息进一步包括在对包括在位流中的线性关键字区域进行译码而使用的线性关键字译码信息,并且所述线性关键字译码信息包括表示是否存在所述线性关键字区域的标志位,包括在所述线性关键字区域中的关键字数据的数量、以及所述线性关键字区域的开始和结尾关键字数据,其中该线性关键字区域中的关键字数据线性增加。 54. The bitstream of claim 50, wherein the key data encoding / decoding information further comprises linear key decoding information comprises a bit stream of a linear key region and used for decoding, and the linear key decoding information includes a flag indicating whether the linear key region, including the start and end of the number of keywords in the linear key data region, and the linear key region key data, including key data in the linear region of the linear increase of the keyword.
55、如权利要求50所述的位流,其中所述旋转差分数据被编码,以便由四元数表示的旋转差分数据的仅三个分量被量化。 55. The bit stream of claim 50, wherein said rotational differential data are encoded so that only three components represented by a quaternion are quantized rotational differential data.
56、如权利要求50所述的位流,其中所述熵编码信息进一步包括: 56. The bitstream of claim 50, wherein said entropy coding information further comprises:
关键字值标志位,表示所述关键字值数据的每个分量的旋转差分数据是否具有相同值;以及 Key value flag indicates the rotational differential data of each component of key value data have the same value;
预定旋转差分数据,当所述关键字值标志位表示所述关键字值数据的每个分量的旋转差分数据具有相同值时,将被译码成所有关键字值数据的旋转差分数据的每个分量。 Predetermined rotational differential data, when the key value flag indicates that the rotational differential data of each component of key value data having the same value, to be decoded into a rotational differential data of all key value data of each of the components.
57、如权利要求50所述的位流,其中逆循环DPCM信息进一步包括用于在所述旋转差分数据上执行逆循环DPCM操作的内部旋转差分数据。 57. The bitstream of claim 50, wherein the reverse cycle DPCM information further comprises means for performing an inverse DPCM operation cycles on said internal rotary differential rotation data differential data.
Description  translated from Chinese
用于对定向内插器进行编码和译码的方法和装置 Method and apparatus for the orientation interpolator is encoded and decoded

技术领域 FIELD

本发明涉及一种用于编码和译码三维动画数据的方法和装置,更准确地说,涉及一种用于对定向内插器进行编码和译码的方法和装置,该定向内插器表示有关动画中的对象的旋转的信息。 The present invention relates to a method and apparatus for encoding and decoding three-dimensional animation data, and more particularly, to a method and apparatus for the directional interpolator for encoding and decoding, the representation of the orientation interpolator For information about rotating animation object.

背景技术 BACKGROUND

MPEG-4BIFS,是国际多媒体标准之一,利用具有动画地关键字和关键字值的内插器节点支持一种基于关键帧的动画。 MPEG-4BIFS, is an international multimedia standards, the use of interposer has keywords and keyword values animate node supports a keyframe-based animation.

为尽可能自然和平滑地使用这样一种基于关键帧动画技术来表示动画,要求相当多的关键字和相当多的关键字值数据,并且关键帧间的现场数据是由内插器填充的。 Natural and smoothly as possible using such a keyframe-based animation techniques to represent animation, require a considerable number of keywords and a considerable amount of key value data, key frames and scene data is filled in by the interpolator. 在虚拟现实造型语言(VRML)中的内插器包含线性或球性内插。 The interpolator in the virtual reality modeling language (VRML) in linear interpolation or contains Spherial.

在时间轴上关键字和关键字值近似原始的动画曲线。 On the timeline keywords and keyword values approximate original animation curves. 图1是根据在三维球性的表面上的时间推移由定向内插器节点所表示的动画数据的二维轨迹的图。 Figure 1 is a time on the surface of the ball of the three-dimensional diagram of a locus of the passage by the two-dimensional orientation interpolator node represented by animation data. 如图1所示,常规的MPEG-4BIFS支持关键帧间的球性线性内插器,并且动画路径看起来与表示该动画数据的变化的一组片段类似。 As shown in Figure 1, the conventional MPEG-4BIFS supports keyframes Spherial linear interpolator and an animation path looks similar to changes in the representation of a set of moving image data segment.

在由BIFS提供的定向内插器节点中,关键字数据使用-8和8间的间断数字表示在动画所处的时间轴上的预定时间瞬间。 In directional interpolation provided by BIFS node, the keyword data using intermittent digital -8 and 8 represents a predetermined time instant at which the animation timeline. 关键字值数据表示在由关键字数据所表示的预定的时间瞬间在合成图像中有关该对象的旋转的信息。 Keyword value data representation at a predetermined time instant represented by the keyword data rotation information about the objects in the composite image. 在不是由关键字数据所表示的在另一预定时间瞬间上的对象的旋转信息是通过球性线形内插来利用与两个时间瞬间相对应的、与预定时间瞬间最邻近的关键字数据而获得的。 In the rotation information is not represented by the keyword data on another predetermined time instant object is achieved by linear interpolation within the sphere of the use of the two time instants corresponding with the predetermined time instant nearest keyword data obtained.

在球性线性内插器中,由旋转轴和旋转角来表示旋转信息。 In glomerular linear interpolator, the rotation angle by the rotation axis and the rotation information is represented. 象虚拟现实造型语言(VRML)一样,MPEG4-BIFS利用定向内插器节点来支持由旋转轴和旋转角所表示的旋转信息。 Like Virtual Reality Modeling Language (VRML), like, MPEG4-BIFS use within directional interpolation node to support the rotation information from the rotation axis and the rotation angle represents. 当利用在球性线性内插器中的关键字值数据来生成平滑动画时,关键帧间的关键字值数据的微分值彼此非常相关,这导致了数据间的冗余。 When the use of linear ball inserted in the device key value data to generate a smooth animation, the key values of the inter-frame differential value of the key data is very related to each other, which causes redundancy among data. 因此,使用数据的微分值来编码关键字值数据的一种方法是很有效的。 Therefore, the use of the differential value data to a method of encoding key value data is very effective.

MPEG-4 BIFS提供用于编码由定向内插器节点的关键字和关键字值数据所表示的现场数据的两种不同的方法。 Both MPEG-4 BIFS provide keywords and keyword values used to encode data from within the directional interpolator node represented by the field data in different ways. 一种是利用脉码调制(PCM)来编码现场数据的方法,另一种是利用差分脉码调制(DPCM)和熵编码来编码现场数据的方法。 One is the use of pulse code modulation (PCM) method to encode field data, and the other is the use of differential pulse code modulation (DPCM) and entropy coding to encode field data.

在利用PCM来编码现场数据的方法中,仅在将被编码的关键字数据和关键字值数据上执行量化处理。 PCM encoded using the method of field data, quantization processing performed only on the data to be coded keywords and key value data. 由于将被编码的数据的特征在本方法中不被考虑,该方法被认为无效。 Since feature data to be encoded are not considered in this method, this method is considered ineffective. 在利用PCM来编码现场数据的方法中,输入定向内插器节点的现场数据,并且现场数据的关键字值数据被变换成在四元数空间中的值。 PCM is encoded using the method of field data, the input node is inserted in the orientation of the field data, field data and key value data is converted into a value in quaternion space. 接着,量化关键字和关键字值数据。 Then, to quantify the value of the keywords and keyword data. 量化的现场数据以二进制数据的形式被输出。 Quantizing field data in the form of binary data is output. 为测量与原始现场数据相比四元数变换的结果可视失真的程度,二进制数据被复原成由旋转轴和旋转角组成的关键字值数据。 To measure the extent of the field data and the results compared to the original quaternion transformation of visual distortion, binary data is restored to the rotation axis and the rotation angle of the composition of key value data. 存储并在屏幕上输出定向内插器节点的复原后的现场数据。 And outputting the stored directional field data interpolator node after restoring the screen. 利用复原后的数据测量由四元数误差所引起的图像的可视失真的程度是可能的。 Measured using the image data from the restored quaternion error caused by the degree of visual distortion is possible. 图像的失真可用以下等式(1)来计算。 Distortion of the image using the following equation (1) is calculated.

在等式(1)中,N表示现场数据的数量,以及εi表示在一四元数时间中复原编码关键字值数据Qi以及关键字值数据 In equation (1), N represents the number of field data, and εi represents recovery encoded key value data Qi and key value data in a quaternion time

间的一差分值。 A difference value between.

另一方面,在用于使用DPCM和熵编码来编码现场数据的方法中,连续数据间的一关联被考虑,因此在编码效率方面,这种方法被认为比利用PCM来编码现场数据的方法更有效。 On the other hand, in the method for using DPCM and entropy coding to encode field data, a correlation between successive data is considered, so the coding efficiency, this method is considered compared to the use of PCM encoded field data approach is more effective. 在这种方法中,在将被编码的先前复原关键字值和关键字值数据间的一差分值在一量化处理前被计算,然后量化该差分值,从而通过利用在该差分值中所示的数据的特性提高编码效率。 In this method, will be encoded in a previously restored key values and the difference value between the key value data is calculated before a quantization process, and then quantizes the difference value, so that by utilizing the difference value shown in characteristic data to improve the encoding efficiency.

图2A和2B分别是利用线性DPCM和熵编码的一MPEG-4PMFC以及利用线性衰减DPCM和熵编码的一MPEG-4PMFC译码器的框图。 2A and 2B are using linear DPCM and entropy encoding an MPEG-4PMFC decoder and the use of the linear attenuation DPCM and entropy coding a block diagram of one-4PMFC MPEG. 在图2A中所示的一线性DPCM运算符用以下等式(2)来计算当前关键字值数据和先前复原关键字值数据间的差分数据 In FIG linear DPCM operator shown in 2A with the following equation (2) to calculate the current key value data and previously restored key value data differential data between

在等式(2)中,Qi表示在一预定时间瞬间(t)的原始关键字值数据,其是由一四元数来表示,以及 In equation (2), Qi said that at a predetermined time instant (t) of the original key value data, which is composed of a quaternion to represent, and

表示在一预定时间瞬间(t-1)的关键字值数据,其是从一误差补偿电路复原。 Represents key value data at a predetermined time instant (t-1), which is to recover from an error compensation circuit.

然而,在图2A中所示的在用于编码关键字值数据的装置中执行的编码方法没有一高的编码效率。 However, the encoding method performed in the apparatus for encoding key value data shown in FIG. 2A in none of the high coding efficiency. 通过分析关键字值数据很容易指出该编码方法的缺点是什幺是可能的,其确定在一四元数时间中一对象的旋转。 By analyzing the key value of the data it is easy to point out the shortcomings encoding method is Jinping is possible, it is determined in a quaternion rotation time of an object. 关键字值数据由以下等式中的一四元数来表示。 Key value data represented by the following equation in a quaternion number.

例如,当一个四元数的分量具有相同的绝对值以作为在一四元数时间中不同标记的另一四元数的它们的相应的分量,如等式(3)中所示时,两个四元数被认为是相同的。 For example, when a component of the quaternion have the same absolute value as shown in their respective components of a quaternion time another quaternion different tags, as shown in equation (3) at the time, two quaternion is considered to be the same. 换句话说,这两个四元数在一3D时间的一对象的旋转变换方面提供相同的效果,这表示影响一对象的旋转变换的因素是一旋转轴的一方向以及一旋转角,而不是该旋转轴的矢量。 In other words, the two quaternions provide the same effects in terms of the rotational transformation of a 3D time of an object, which means that factors affecting the rotation transformation of an object is a direction of a rotating shaft and a rotation angle, rather than The rotation axis vector. 然而,象在MPEG4BIFS中,如果关键字值数据由使用等式(3)的一四元数表示以及一差分值通过计算连续关键字值数据间的矢量中的差值被线性计算,该差分值不为0,这表示线性差分值没有很好地反映在旋转变换中的冗余。 However, like in MPEG4BIFS, if the key value data by using equation (3) represents the number of a quaternion, and a differential value calculated by the vector between successive key value data of the difference is a linear calculation, the difference value is not 0, which means that linear differential values not well reflected in the rotational transformation of redundancy. 因此,使用如等式(1)中所示用于测量图像的失真程度的方法精确地测量图像质量是不可能的。 Therefore, as shown in equation (1) shown in a method for measuring the degree of distortion of the image accurately measure the image quality is not possible.

发明内容 SUMMARY

为解决上述和其它问题,本发明的一个方面是提供一种用于编码和译码一定向内插器的方法和装置,编码和译码由从一原始定向内插器抽取的断点组成的一抽取定向内插器以便防止抽取定向内插器与原始定向内插器间的一误差大于一允许误差极限,从而提供具有一高压缩比的高质量动画。 To solve the above and other problems, an aspect of the present invention is to provide a method for encoding and decoding interpolator inwardly certain methods and apparatus for encoding and decoding by the interpolator extracted from an original orientation breakpoint composition inserted inside a directional extraction in order to prevent an error in the decimation directional interpolator within the original orientation interpolator is greater than a tolerance limit, thus providing a high-quality animation with a compression ratio.

本发明的另一方面是提供一种用于编码和译码一定向内插器的方法和装置,通过计算一旋转差分比,能提供具有一高压缩比的高质量动画,能充分地反映在旋转变换中的冗余,并能使用该旋转差分值来编码一定向内插器的关键字值数据。 Another aspect of the present invention is to provide a method for encoding and decoding method and apparatus must inwardly interpolator, and by calculating a ratio of the rotation difference, can provide high-quality animation with a high compression ratio, can be adequately reflected in rotational transformation redundant, and can be used to encode the rotational differential value interpolator inwardly certain key values of data.

本发明的另一方面提出了一种由根据本发明的对一定向内插器进行编码和译码的方法和装置来编码和译码的一位流,这能提供具有一高压缩比的高质量的动画。 Another aspect the present invention proposes a method and apparatus of the present invention must be inwardly interpolator encoding and decoding according to the encoding and decoding of a bit stream, which can provide a high compression ratio with high quality animation.

因此,为实现本发明的上述和其它目的,提供一种对一定向内插器进行编码的装置,该定向内插器包括用于表示在一时间轴上关键帧位置的关键字值以及用于表示一对象的旋转的关键字值数据。 Therefore, in order to achieve the above and other objects of the invention, to provide a certain inward interpolator means for encoding, within this orientation interpolator includes a key value used to represent the key frames in a time axis and a position for represents the rotation of key value data of an object. 该装置包括一断点抽取器、一关键字数据编码器、一关键字值数据编码器。 The apparatus includes an break point extractor, a key data encoder, a key value data encoder. 该断点抽取器从由输入到其中的一定向内插器所组成的第一动画路径中抽取一最少数的断点,这能产生不大于所述第一动画路径和由抽取断点所生成的一第二动画路径间的一预定误差极限的一误差。 The breakpoint extractor extracts a minimum number of break from a first animation path from the input to which a certain inwardly interpolator composition, which can result in no more than the first animation path generated by the extracted break point and a second one of a predetermined limit of errors error animation paths. 一关键字数据编码器,编码从断点抽取器输入的关键字数据。 A key data encoder, extractor from the breakpoint keywords entered data. 一关键字值数据编码器,编码从所述断点抽取器输入的关键字值数据。 A key value data encoder, extractor from the breakpoint input key value data.

最好,本发明进一步包括一重新采样器以及一选择器。 Preferably, the present invention further comprises a re-sampler, and a selector. 该重新采样器将第一动画路径采样到具有一预定时间量间隔的多个预定部分中,并输出包括重新采样关键字数据以及重新采样关键字值数据的一定向内插器。 The re-sampler sampling the first animation path into a plurality of portions having a predetermined amount of a predetermined time interval and outputs the resampled key data and including resampling interpolator inwardly certain key value data. 该选择器响应一外部输入信号,将输入到其中的一定向内插器输出到该重新采样器以及该断点抽取器中。 The selector in response to an external input signal, wherein the input to the interpolator output to a certain inwardly resampler and the decimator in the breakpoint.

为实现本发明的上述和其它目的,提供一种对一定向内插器进行编码的装置,该定向内插器包括用于表示在一时间轴上关键帧位置的关键字值以及用于表示一对象的旋转的关键字值数据。 To achieve the foregoing and other objects of the present invention, there is provided an inwardly interpolator for certain encoding means, within this orientation interpolator includes a time axis for representing the key values and the key frame for indicating a position rotation of key value data objects. 该装置包括一重新采样器、一关键字数据编码器、一关键字值数据编码器。 The apparatus includes a resampler, a key data encoder, a key value data encoder. 该重新采样器将由一输入定向内插器组成的一动画路径采样到具有一预定时间量间隔的多个预定部分中并输出包括采样关键字数据和采样关键字值数据。 The re-sampler by an input orientation interpolator consisting of the animation path is sampled a plurality of predetermined portions having a predetermined amount of time intervals and outputs sampled key data and sampled includes a key value data. 该关键字数据编码器,编码从该重新采样器输入的关键字数据。 The key data encoder, from the re-sampling the input of key data. 该关键字值数据编码器,生成一旋转差分值,该差分值用来按几乎等于通过一当前关键帧的关键字值数据应用到该对象的旋转变换以及通过一前关键帧的关键字值数据应用到该对象的旋转变换间的一差值来旋转该对象,并因此编码从该重新采样器输入的关键字值数据。 The key value data encoder, generates a rotational differential value, the difference value is used by almost equal rotational transformation applied to the object by key value data of a current keyframe by key value and the key frame data of a prior rotation transformation is applied to a difference between the object to rotate the object, and thus encode input from the resampling the key value data.

最好,断点抽取器包括一线性内插器、一误差计算器、以及一确定单元。 Preferably, the breakpoint extractor included in the linear interpolator, an error calculator, and a determination unit. 该线性内插器抽取一输入动画路径的一开始路径点以及一结束路径点,在开始和结束路径点间的选择路径点,以及利用所选择的路径点内插还没有被选择的其它路径点。 Interpolator extract an input animation path of the linear path beginning point and an end point of the other path path points within the path between the start and end point of the selection path points, and the path interpolation using the selected point has not yet been selected . 该误差计算器,计算输入动画路径和由线性内插器利用内插所生成的一内插动画路径间的一误差。 The error calculator for calculating the input animation path and the interpolated by linear interpolation using a device inserted within the generated an error between the animation path. 该确定单元抽取断点,通过该操作,能最小化输入动画路径与内插动画路径间的一误差,并输出选择的断点,如果相应误差小于一预定误差极限。 The determination unit extraction break through this operation, to minimize input animation path and the interpolated animation path between an error and outputs the selected breakpoint, if the corresponding error is less than a predetermined limit of errors.

最好,关键字值编码器包括一旋转差分数据生成器以及一熵编码器。 Preferably, the key value encoder includes a rotational differential data generator, and an entropy encoder. 该旋转差分数据生成器利用一当前关键帧的一旋转变换值以及一前关键帧的一复原旋转变换值来生成一旋转差分值,该差分值被用来按几乎等于通过关键字值数据而应用到在该当前关键帧中的该对象的旋转变换以及通过关键字值数据而应用到前关键帧中的该对象的旋转变换间的一差值来旋转该对象,并通过量化该旋转差分值来输出旋转差分数据。 The differential rotation data generator using a current value of a rotation transformation keyframe and a restored rotational transformation of a former key value to generate a rotating frame difference values, the difference value is used by almost equal to the value of data and applications by keyword to a difference in the current rotation transformation keyframe of the object as well as by key value data and applications to the previous keyframe rotation transformation between the object to rotate the object, and by quantifying the rotation differential value differential data output rotation. 该熵编码器,熵编码该旋转差分数据。 The entropy encoder entropy encoding the rotational differential data.

最好,旋转差分数据生成器包括一第一四元数乘法器,一量化器,一逆量化器,以及一第二四元数乘法器。 Preferably, the rotational differential data generator includes a first quaternion multiplier, a quantizer, an inverse quantizer, and a second quaternion multiplier. 该第一四元数乘法器利用该当前关键帧的旋转变换值以及前关键帧的复原旋转变换值来生成旋转差分值。 The first quaternion multiplier value by using the current rotational transform and rotation transformation keyframe values recover before keyframes to generate rotational difference score. 该一量化器通过量化该旋转差分值来生成旋转差分数据。 A quantizer quantizing the rotational differential value of the difference data is generated by the rotation. 该逆量化器通过逆量化该旋转差分数据来生成一复原旋转差分值。 The inverse quantization by an inverse quantizer which generates rotational differential data a restored rotational differential value. 该一第二四元数乘法器通过将复原旋转差分值四元数乘以前关键帧的一旋转变换值来生成当前关键帧的一复原旋转变换值。 The second quaternion multiplier A by the restored rotational differential value multiplied by a quaternion rotational transformation value of the previous key frame to generate a restored rotational transformation value of the current key frame.

最好,关键字数据编码器包括一第一量化器,一第一DPCM处理器,一DND处理器,以及一第一熵编码器。 Preferably, the key data encoder includes a first quantizer, a first DPCM processor, a DND processor, and a first entropy coder. 该第一量化器利用预定量化位来量化一定向内插器的关键字数据。 The first use of a predetermined quantization bit quantizer for quantizing certain keywords data inwardly interpolator. 该第一DPCM处理器,生成量化关键字数据的差分数据。 The first DPCM processor generates differential data to quantify the keyword data. 该DND处理器,根据差分数据与一最大值的关系以及它们间一最小值在该差分数据上执行一DND操作。 The DND processor, based on the relationship between the differential data between a maximum value and a minimum value are performed on the differential data of a DND operation. 该第一熵编码器,熵编码从DND处理器输入的有效期分数据。 The first entropy encoder entropy coding from DND processor input data valid points.

为实现本发明的上述和其它目的,提供一种用于对一位流进行译码的装置,将包括用于表示在一时间轴上关键帧的位置的关键字数据以及用于表示一对象的旋转的关键字值数据的一定向内插器进行编码。 To achieve the foregoing and other objects of the present invention, there is provided a method for decoding a bit stream of means for indicating the key data includes a key frame on a time axis, and the position of an object used to represent rotating inwardly interpolator certain key value data are encoded. 该装置包括一关键字数据译码器,一关键字值数据译码器,一定向内插器合成器。 The apparatus includes a key data decoder, a key value data decoder, a certain inward interpolator synthesizer. 该关键字数据译码器译码来自一输入位流的关键字数据。 The key data decoder decodes key data from an input bit stream. 该关键字值数据译码器,译码来自该输入位流的关键字值数据。 The key value data decoder decodes the data from the key values of the input bit stream. 该定向内插器合成器,通过合成译码关键字值数据和利用该译码关键字值数据球性线性内插的关键字值数据来生成一定向内插器。 The directional interpolation Synthesizer, be generated by synthesizing the decoded key value data and key value data utilizing the decoded key value data linearly interpolated ball of a certain inward interpolator.

最好,关键字值数据译码器包括一熵译码器,一逆循环DPCM运算符,一逆量化器,以及一四元数乘法器。 Preferably, the key value data decoder includes an entropy decoder, an inverse DPCM loop operator, an inverse quantization, and a quaternion multiplier. 该熵译码器通过熵译码来自该位流的关键字值数据来生成循环DPCM旋转差分数据或量化旋转差分数据。 Key value data by the entropy decoder entropy decoding from the bit stream to generate DPCM loop rotational differential data or quantized rotational differential data. 该逆循环DPCM运算符,通过在从该位流译码的DPCM的次数操作后在从熵译码器输入的旋转差分数据上执行一逆循环DPCM操作来生成量化旋转差分数据。 The reverse cycle DPCM operator, by the number of times after the operation from the bit stream decoded DPCM performed an inverse DPCM operation on the rotational cycle of the differential data input from the entropy decoder generates quantized rotational differential data. 该逆量化器,按几乎等于通过逆量化量化旋转差分数据的每个关键帧的四元数关键字值数据应用到该对象的旋转变换间的一差值来旋转该对象。 The inverse quantization, press almost equal application to a difference between the rotation transformation of the object to rotate the object quantify key value data quaternion rotation difference data of each keyframe by inverse quantization. 该四元数乘法器,通过将当前关键帧的一旋转差分值四元数乘以一前关键帧的一复原旋转变换值来生成一当前关键帧的一旋转变换值。 The quaternion multiplier, by a rotational difference between the current keyframe quaternion value by multiplying a restored rotational transformation value of a front key frame to generate a rotational transformation value of a current keyframe.

为实现本发明的上述和其它目的,提供一种对包括用于表示在一时间轴上关键帧的位置的关键字数据和表示一对象的旋转的关键字值数据的一定向内插器进行编码的方法。 To achieve the foregoing and other objects of the present invention, there is provided a pair comprising key data for representing an object representing a certain inward rotation interpolator key value data on the position of a key frame is encoded timeline method. 该方法包括(b)通过从由该定向内插器组成的第一动画路径抽取一最少的断点来生成将被编码的关键字数据和关键字值数据,能产生小于第一动画路径和通过抽取断点生成的一第二动画间的一预定误差极限的一误差,(d)编码在步骤(b)中生成的关键字数据,以及(e)编码在步骤(b)中生成的关键字值数据。 The method includes (b) generating key data to be encoded, and key value data by extracting a minimal break from a first animation path by the composition of the orientation interpolator, can produce less than a first animation path and by a predetermined error limit of an error of a second animation generated between the extracted break, (d) encoding generated in step (b), the key data, and (e) encoding generated in step (b), the keyword value data.

最好,步骤(b)包括(b1)选择第一动画路径的一开始路径点和一结束路径点,(b2)依次选择开始和结束路径点间的路径点并利用所选择的路径点内插仍未被选择的其它路径点,(b3)计算第一动画路径与通过内插在步骤(b2)中生成的一第二动画路径间的一误差,以及(b4)选择断点,通过该操作第一动画路径和第二动画路径间的一误差能被最小化,核对相应的误差是否小于一预定误差极限,并确定将被编码的关键字数据和关键字值数据。 Preferably, the step (b) includes (b1) to select a start point of the first animation path and a path to end the path point, (b2) in order to select the path points between the start and end points of the path and use within the selected route point interpolation Other yet selected path points, (b3) calculating an error of the first animation path and by interpolation in step (b2) produced in a second animation paths, and (b4) selecting break, by the operation an error of the first animation path and the second path between the animation can be minimized, check the corresponding error is less than a predetermined limit of errors, and to determine the encoded key data and key value data.

最好,用于编码一定向内插器的方法可进一步包括在步骤(b)之前的(a),通过将第一动画路径采样到具有一预定时间量间隔的多个预定部分中,生成包括重新采样关键字数据以及重新采样关键字值数据的一定向内插器,或可能进一步包括步骤(b)之后的(c),通过将利用抽取断点所构成的第二动画路径采样到具有一预定时间量间隔的多个预定部分中来生成将被编码的关键字数据和关键字值数据。 Preferably, for encoding a certain inward interpolator method may further comprise prior to step (b), (a), by sampling the first animation path into a plurality of portions having a predetermined amount of a predetermined time interval, the generating comprises resampled key data and resampled key value interpolator inwardly certain data, or may further comprise the step (b) after (c), by the use of the extracted second animation path constituted breakpoint sample having a a plurality of predetermined portions of a predetermined amount of time interval to generate the key data to be encoded, and key value data.

为实现本发明的上述和其它目的,提供一种对包括用于表示在一时间轴上关键帧的位置的关键字数据和表示一对象的旋转的关键字值数据的一定向内插器进行编码的方法。 To achieve the foregoing and other objects of the present invention, there is provided a pair comprising key data for representing an object representing a certain inward rotation interpolator key value data on the position of a key frame is encoded timeline method. 该方法包括(a)将由定向内插器所构成的一动画路径采样到具有一预定时间量间隔的多个预定部分中,并因此生成包括重新采样关键字数据和重新采样关键字值数据的一定向内插器,(d)减小在步骤(a)中采样的关键字数据的范围并编码关键字数据,以及(e)通过生成和编码一旋转差分值来编码在步骤(a)中采样的关键字值数据,该旋转差分值被用来按几乎等于通过一当前关键帧的关键字值数据而应用到该对象的旋转变换与通过一前关键帧的关键字值数据而应用到该对象的旋转变换间的一差值来旋转该对象。 The method comprises (a) sampling an animation path by an orientation interpolator consisting of the plurality of portions having a predetermined interval in the predetermined amount of time, and thus generating comprises resampled key data and resampled key value data of a certain inwardly interpolator, (d) reducing the range of samples in step (a), the key data and coded key data, and (e) to encode in step (a) by generating and encoding a rotational differential value sampling The key value data, which is used to rotate the differential value by almost equal to the current key value data by a keyframe and rotation transformation is applied to the object by key value data with a key frame before being applied to the object One difference between the rotation transformation to rotate the object.

最好,步骤(d)包括用一预定数据的量化位来量化关键字数据,通过在量化数据上执行一预定DPCM操作来生成差分数据,以及熵编码该差分数据。 Preferably, step (d) comprises using a predetermined quantization bit quantized key data to the data, generates differential data by performing a predetermined DPCM operation on quantized data, and entropy coding the difference data.

最好,步骤(e)包括(e1)生成一旋转差分值,该差分值用来通过利用当前关键帧的一旋转变换值以及前关键帧的一复原旋转变换值按几乎等于通过当前和前关键帧的关键字值数据而应用到该对象的旋转变换间的一差值来旋转该对象并通过量化该旋转差分值来生成旋转差分数据,(e2)在该旋转差分数据上有选择地执行一线性DPCM或一循环DPCM操作,以及熵编码该旋转有效期分数据。 Preferably, step (e) comprises (e1) generating a rotational differential value, which is used by almost equal to the difference value by the current and the previous key value by a restored rotational transformation with the current keyframe and a rotational transformation value of the former key frame key value data frame applied to a difference between the rotational transformation of the object to rotate the object and generates the rotational differential data by quantizing the rotational differential value, (e2) selectively on the rotational differential data to execute a or a linear DPCM DPCM loop operation, as well as the validity of the rotation entropy coding data points.

最好,步骤(e1)包括(e11)利用当前关键帧的一旋转变换值以及前关键帧的一复原旋转变换值生成旋转差分值,(e12)通过量化该旋转差分值来生成旋转差分数据,(e13)通过逆量化旋转差分数据来生成一复原旋转差分值,以及(e14)通过将复原旋转差分值四元数乘以前关键帧的一复原旋转变换值来生成当前关键帧的一复原旋转变换值。 Preferably, step (e1) includes (e11) using a rotational transformation value of the current keyframe and a restored rotational transformation value of the previous key frame generating rotational differential value, (e12) generating rotational differential data by quantizing the rotational differential value, (e13) generating rotational differential data by the inverse quantizing a restored rotational differential value, and (e14) the restored rotational differential value by quaternion multiplying a restored rotational transformation value of the previous key frame to generate a current key frame restored rotational transformation value.

为实现本发明的上述和其它目的,提供对一位流进行译码的方法,将包括有用于表示在一时间轴上关键帧的位置的关键字数据和用于表示一对象的旋转的关键字值数据的一定向内插器编码成该位流。 To achieve the foregoing and other objects of the present invention, there is provided a stream of the decoding method, including key data indicating key frame's location on the time axis and a rotation of an object used to represent keywords certain inward interpolator value data encoded into the bit stream. 本发明包括(a)译码来自一输入位流的关键字数据,(b)译码来自该输入位流的关键字值数据,以及(c)通过合成译码关键字值数据以及利用该译码关键字值数据球性线性内插的关键字值数据来生成一定向内插器。 The present invention includes (a) decoding key data from an input bit stream, (b) decoding key value data from the input bit stream, and (c) by synthesizing the decoded key value data and the use of the translation key value data code key value data balls of linear interpolation to generate a certain inward interposer.

最好,在步骤(c)中,如果不对与关键字数据相对应的关键字值进行译码,该关键字数据当前附属于定向内插器合成,与关键字数据相对应的关键字值数据利用与当前合成关键字数据相对应的译码关键字值数据和与下一个将被合成的关键字数据相对应的译码关键字值数据被合成。 Preferably, in step (c), if not the key corresponding to key value data is decoded, the keyword data is currently attached to the interposer synthesis of orientation, and keyword data corresponding key value data Utilization of this synthetic key data corresponding to the decoding key value data and the next key data to be synthesized corresponding to the decoding key value data are synthesized.

最好,步骤(a)包括通过熵译码该输入位流来生成差分数据以及通过在差分数据上执行一预定的DPCM操作或一逆DND操作来生成量化关键字数据以及通过逆量化量化关键字值数据来生成复原关键字数据。 Preferably, step (a) comprises the input bit stream by entropy coding the difference data is generated and by performing a predetermined DPCM operation or an inverse DND operation on the differential data to generate the key data is quantized by the inverse quantization and quantization keywords value data to generate restored key data.

最好,步骤(b)包括(b1)通过熵译码来自该位流的关键字值数据来生成循环DPCM旋转差分数据或量化旋转差分数据,(b2)通过在从该位流译码的DPCM次数操作后在熵编码旋转差分数据上执行一逆循环DPCM操作来生成旋转差分数据,(b3)生成一旋转差分值,该差分值被用来按几乎等于通过逆量化旋转差分数据的每个关键帧的四元数关键字值数据而应用到该对象的旋转变换间的一差值来旋转该对象,以及(b4)通过将当前关键帧的一旋转差分值四元数乘以一前关键帧的一译码旋转变换值来生成一当前关键帧的一旋转变换值。 Preferably, step (b) comprises (b1) by the entropy decoding key value data from the bit stream to generate DPCM loop differential data or quantized rotational differential rotation data, (b2) through the bit stream from the decoded DPCM After the number of operations performed on the rotational differential data entropy encoding an inverse DPCM operation cycle to generate a differential rotation data, (b3) generating a rotational differential value, the difference value is used by almost equal to the quantized rotational differential data by the inverse of each of the key quaternion key value data frame and a difference between rotational transformation applied to the object between the object to rotate, and (b4) by a rotational difference between the current key frame is multiplied by a quaternion value before keyframe of a decoded rotational transformation values to generate a rotational transformation value of a current keyframe.

为实现本发明的上述和其它目的,提供一位流,将包括有用于表示在一时间轴上关键帧的位置的关键字数据和用于表示一对象的旋转的关键字值数据的一定向内插器编码成该位流。 To achieve the foregoing and other objects of the present invention, there is provided a stream, including key data indicating key frame's location on the time axis for representing a certain inward rotation of a key value data object interposer encoded into the bit stream. 该位流包括关键字数据编码/译码信息以及关键字值数据编码/译码信息。 The bit stream data includes key encoding / decoding information and key value data encoding / decoding information. 将该关键字数据编码/译码信息编码成关键字数据和译码该关键字数据所必须的信息。 The key data encoding / decoding information encoding and decoding key data into the key data necessary for the information. 将关键字值数据编码/译码信息编码成关键字值数据以及译码该关键字值数据所必须的信息被编码。 The key value data encoding / decoding information encoded into decoded key value data and key value data of the necessary information is encoded. 在这里,关键字数据编码/译码信息包括逆DND操作信息,第一逆DPCM操作信息,第一逆量化信息。 Here, the key data encoding / decoding information includes inverse DND operation information, first inverse DPCM operation information, the first inverse quantization information. 该逆DND操作信息包括表示将在通过熵译码该位流所生成的差分数据上执行的逆DND的多个预定周期的逆DND的次数以便扩展该差分数据的范围以及在每个逆DND操作周期中所使用的差分数据的最大和最小值。 The inverse DND operation information includes information indicating the number of a plurality of predetermined periods in the entropy coding the difference data by the bit stream generated by the implementation of the inverse DND inverse DND to extend the range of the differential data, and in each inverse DND operation The maximum and minimum cycle used in the difference data. 第一逆DPCM操作信息包括将在逆DND差分数据上执行的逆DPCM操作操作的次数以便将逆DND差分数据转换成每个逆DPCM操作周期的已量化关键字数据以及内部关键字数据。 First inverse DPCM operation information including the operation of an inverse DPCM operation to be performed on the differential data of the inverse DND to convert the number of differential data to the inverse DND operation cycles each inverse DPCM quantized key data and internal key data. 第一逆量化信息用在逆量化中以通过逆量化已量化关键字数据来生成复原关键字数据。 First inverse quantization information used in inverse quantization to the quantized by the inverse quantization to generate restored key data key data. 关键字值数据译码/编码信息包括由量化用于通过几乎等于通过每个关键帧的四元数关键字值数据而应用到该对象的旋转变换间的一差值来旋转该对象的一旋转差分数据熵编码的旋转差分数据,熵译码信息包括表示将在旋转差分数据上执行的一熵译码方法的一熵译码模式,逆循环DPCM操作信息包括逆循环DPCM操作的次数,其表示是否将在熵译码旋转差分数据上用熵译码模式执行一逆循环DPCM操作,以及第二逆量化信息包括多个预定逆量化位,用来通过逆量化量化关键字值数据来复原原始关键字值数据。 Keyword value data decoding / encoding information includes quantification by almost equal to a difference between the rotation and the application to convert between the object by quaternion key value data for each keyframe to rotate the object of a rotating rotational differential data entropy-encoded difference data, entropy coding information includes information indicating an entropy coding mode an entropy decoding method to be executed on a rotary differential data, an inverse DPCM operation information including the number of loop cycles reverse DPCM operation, which indicates whether entropy coding is performed on the rotational differential data by entropy coding mode an inverse DPCM operation cycle, and a second inverse quantization information comprising a plurality of predetermined inverse quantization bits for quantization by the inverse quantization to restore the original key value data key word value data.

最好,关键字数据编码/译码信息进一步包括线性关键字译码信息,用于对包括在该位流中的一线性关键字部分进行译码,并且该线性关键字译码信息包括一特征位,该特征位表示是否存在线性关键字区域,包含在该线性关键字部分中的关键字数据的数量以及线性关键字部分的开始和结束关键字数据。 Preferably, the key data encoding / decoding information further comprises linear key decoding information, for including the linear key part in the bit stream is decoded, and the linear key decoding information comprises a feature bit, the flag indicates whether there is a linear key region, containing the start and end key data in the linear key part of the key part of the number, and the linear key data. 上述区域中的关键字数据线性的增加到关键字数据中。 Above area increased linearly keyword data keyword data.

附图说明 Brief Description

本发明的上述目的和优点将通过参考附图详细地描述其优选实施例变得清楚,其中: The above objects and advantages of the present invention will become apparent from preferred embodiments thereof described in detail with reference to the drawings, wherein:

图1是根据一三维球面上的时间推移,描述由一定向内插器节点表示的动画数据的二维轨迹图; Figure 1 is a time lapse of a three-dimensional spherical surface, two-dimensional trajectories described by the interpolator node represents a certain inwardly moving image data;

图2A和2B分别是使用线性DPCM和熵编码的一MPEG-4PMFC编码器以及使用线性衰减DPCM和熵译码的一MPEG-4PMFC的框图; 2A and 2B are a MPEG-4PMFC block diagram of an encoder and a MPEG-4PMFC using linear DPCM and entropy decoding attenuation using linear DPCM and entropy encoding;

图3A是根据本发明的一优选实施例,一种用于编码一定向内插器的装置的框图,以及图3B是根据本发明的一优选实施例,一种用于编码一定向内插器的方法的流程图; Figure 3A is a preferred embodiment of the present invention, a block diagram of an apparatus for encoding a certain inward interpolator, and Figure 3B is a preferred embodiment of the present invention, a method for encoding a certain inwardly interpolator a flowchart of a method;

图4A至4C是根据本发明的优选实施例,一分析器的例子的框图; 4A to 4C is a block diagram showing an example of the analyzer of the present invention according to a preferred embodiment;

图5A是如图3B所示的一步骤S320的流程图; 5A is a flowchart of a step S320 of FIG. 3B;

图5B是根据本发明的一优选实施例,一重新采样方法的流程图; 5B is accordance with a preferred embodiment of the present invention, a flowchart of a method of re-sampling;

图5C是根据本发明的一优选实施例,抽取断点的一方法的一流程图; 5C is in accordance with a preferred embodiment of the present invention, a flow diagram of a method of extracting break;

图6A是描述原始关键字数据以及重新采样关键字数据的图以及图6B是描述一原始动画路径以及一重新采样动画路径的图。 6A is a keyword describing the original data and resampled key data, and FIG. 6B is a diagram describing an original animation path and a resampled animation path in FIG.

图7A至7F是根据本发明的一优选实施例,描述抽取断点的一方法的一例子的图; 7A to 7F is accordance with a preferred embodiment of the present invention, describes a method for extraction of an example of the breakpoints map;

图8是描述在一断点生成模式中从一断点抽取器输出的关键字数据和关键字值数据的图; Figure 8 is a keyword describing the data breakpoint generating mode output from a break extractor and key value data of the map;

图9A是根据本发明的一优选实施例,一关键字数据编码器的框图; 9A is a block diagram of a key data encoder according to a preferred embodiment of the present invention,;

图9B是如图9A所示的一DND处理器的框图; 9B is a block diagram shown in Figure 9A a DND processor;

图10A至10E是根据本发明的一优选实施例,编码关键字数据的一方法的流程图; 10A to 10E are a flowchart of a method for coding key data in accordance with a preferred embodiment of the present invention;

图11是描述一函数encodeSignedAAC的例子的图; FIG 11 is a description of an example of FIG function encodeSignedAAC;

图12A至12J是根据本发明的一优选实施例,描述在执行编码关键字数据的不同步骤后获得的关键字数据的图; 12A to 12J in accordance with a preferred embodiment of the present invention, after performing the various steps described in the coded key data of the key data of FIG obtained;

图13A是根据本发明的一第一实施例,一关键字值数据编码器的框图,以及图13B是根据本发明的一第一实施例,编码关键字值数据的一方法的流程图; 13A is a block diagram of a key value data encoder, and Figure 13B is a flowchart of a method according to the encoding key value data according to a first embodiment of the present invention, a first embodiment of the present invention,;

图14A是描述在一旋转差分值的每个分量中的一概率分布函数(PDF)的典型例子的图; 14A is a description of a typical example of a rotation of each component in the difference value in a probability distribution function (PDF) of the diagram;

图14B是用于非线性量化的反正切曲线; 14B is used to quantify the arc tangent nonlinear curve;

图15A是根据本发明的一优选实施例,从包括在一关键字值数据译码器中的一量化器输出的旋转差分数据的例子;图15B是描述在如图15A所示的差分数据上执行一线性DPCM操作的结果的图;以及图15C是描述在如图15B所示的线性DPCM差分数据上执行一循环DPCM操作的结果的图; 15A is in accordance with a preferred embodiment of the present invention, examples of which include the rotational differential data from a key value data decoder in a quantizer output; Figure 15B is a description of the differential data shown in FIG. 15A perform a linear DPCM operation result; and Figure 15C is a description of the results of performing a DPCM loop linear DPCM operation on the differential data shown in FIG. 15B in FIG;

图16是描述用于熵编码的一函数UnaryAAC()的一个例子的图; FIG 16 is a description of a function UnaryAAC for entropy coding () is an example of a graph;

图17是描述在使用一旋转差分值来编码四元数旋转变换值期间产生的一旋转方向误差的图; FIG 17 is a description of a rotation direction using a rotational differential value during encoding quaternion rotational transformation values generated error map;

图18A是根据本发明的第二实施例,一关键字值数据编码器的一旋转DPCM运算符的框图,以及图18B是如图18A所示的一旋转方向误差计算器的框图; 18A is a second embodiment according to the present invention, the value of a block diagram of a data encoder rotational DPCM operator of a key, and the block diagram shown in FIG. 18B is a rotation direction error calculator shown. 18A;

图19A是根据本发明的第二实施例,一旋转DPCM操作的流程图;图19B是描述如图9A是所示的一旋转方向误差计算器、一旋转方向误差检测器、以及一旋转方向校正器的操作的流程图; 19A is a second embodiment of the present invention, a flowchart of a rotational DPCM operation; FIG. 19B is depicted in Figure 9A is a rotation direction error calculator shown, a rotation direction error detector, and a rotation direction correction a flowchart of the operation;

图20A是根据本发明的第三实施例的一关键字值数据编码器的一量化器的框图,以及图20B是根据本发明的第三实施例的一量化器的操作的流程图; 20B is a flowchart of FIG. 20A is a quantizer according to a third embodiment of the present invention operating in accordance with a block diagram of a quantizer of a key value data encoder to a third embodiment of the present invention, and FIG;

图21A是根据本发明的一优选实施例的用于译码一定向内插器的装置的框图,以及图21B是根据本发明的一优选实施例,一种用于译码一定向内插器的方法的流程图; 21A is a block diagram of apparatus in accordance with a preferred embodiment of the present invention for decoding a certain inward interpolator, and Figure 21B is accordance with a preferred embodiment of the present invention, a method for decoding a certain inwardly interpolator a flowchart of a method;

图22是根据本发明的一优选实施例的一关键字数据译码器的框图; 22 is a block diagram of a key data decoder according to a preferred embodiment of the present invention embodiment of FIG;

图23A和23B是根据本发明的一优选实施例的译码关键字数据的一种方法的流程图; 23A and 23B are a flowchart of a method embodiment of the decoding key data in accordance with a preferred embodiment of the present invention;

图24A是根据本发明的一优选实施例的一关键字值数据译码器的框图,以及图24B是根据本发明的一优选实施例的译码关键字值数据的方法的流程图; A flowchart of a decoding key value data 24A is a block diagram of a method according to a preferred embodiment of the present invention, a key value data decoder embodiment, and FIG 24B is in accordance with a preferred embodiment of the present invention;

图25是根据本发明的一优选实施例,描述输入到一关键字值数据译码器的一熵译码器的一位流的结构的图; Figure 25 is a preferred embodiment of the present invention, showing the structure described is input to an entropy decoder of a key value data decoder of the bit stream;

图26是根据本发明的一优选实施例,合成一定向内插器的关键字数据和关键字值数据的方法的流程图; Figure 26 is a preferred embodiment of the present invention, a flowchart of the synthesis of certain keywords data inwardly key value data and a method of interpolator;

图27是描述计算将被编码的一定向内插器以及一译码定向内插器间的一误差的方法的例子的图; FIG 27 is a description of an error calculation method will be encoded within a certain inwardly interpolator and a decoded orientation interpolator between the example of FIG;

图28是用于将根据本发明用于编码一定向内插器的方法的性能与用于编码一定向内插器的常规方法的性能相比的图; FIG 28 is a performance of a conventional method of a certain inward interpolator according to the present invention for encoding a certain inward interpolator performance and a method for encoding as compared to Fig;

图29A至29J是SDL语言程序代码,通过该代码,根据本发明的一优选实施例,用于译码一定向内插器的装置,实现译码关键字数据和关键字值数据。 29A to 29J are SDL-language program code, through the code, according to a preferred embodiment of the present invention, the decoding must inwardly interpolator means for, implement a decoding key value data and key data.

具体实施方式 DETAILED DESCRIPTION

根据本发明的一优选实施例,将参考附图来更详细地描述一种用于编码一定向内插器的方法和装置。 According to a preferred embodiment of the present invention, will be described with reference to the drawings a method and apparatus for encoding a certain inward interpolator in more detail.

图3A根据本发明的一优选实施例给出了一种用于编码一定向内插器的装置的框图。 Figure 3A according to a preferred embodiment of the present invention shows a block diagram of an apparatus for encoding a certain inward interpolator. 参考图3A,用于编码一定向内插器的装置包括一分析器40、一关键字值编码器200、一关键字值数据编码器300以及一标题编码器400。 3A, the apparatus for encoding a certain inward interpolator comprises a parser 40, a key value encoder 200, a key value data encoder 300, and a header encoder 400.

图3B根据本发明的一优选实施例给出了用于编码一定向内插器的方法的流程图。 Figure 3B according to a preferred embodiment of the present invention shows a flowchart of the encoding method must inwardly interpolator used. 参考图3B,在步骤S300中将将被编码的一定向内插器输出到分析器40中。 With reference to Figure 3B, at step S300 will be coded must inwardly interpolator output to the analyzer 40. 在步骤S320中,分析器40从由定向内插器的x,y,z及θ(旋转角)分量的关键字值数据组成的一第一动画路径抽取将被编码的关键字数据和关键字值数据,将抽取关键字数据输出关键字数据编码器200,并将抽取关键字值数据输出到关键字值数据编码器300。 Extracting a first animation path to be coded keywords and keyword data in step S320, the analyzer 40 by the orientation interpolator from the x, key value data y, z and θ (rotation angle) components of the composition the value of the data, extract key data output key data encoder 200, and outputs the data to extract the key value key value data encoder 300.

关键字数据编码器200利用多个预定已量化位来量化从分析器40所输入的关键字数据,通过在已量化关键字数据上执行一预定DPCM操作来生成差分数据以及在步骤S340中熵编码该差分数据。 200 with the keyword data encoder has a plurality of predetermined quantization bits for quantizing the key data input from the analyzer 40, generates differential data by performing a predetermined DPCM operation on quantized key data, and entropy coding in step S340 The differential data.

关键字值数据编码器300使用多个预定已量化位来量化从分析器40所输入的关键字值数据,通过在量化数据上执行一预定DPCM操作来生成差分数据以及在步骤S360中编码该差分数据。 Key value data encoder 300 uses a plurality of predetermined quantized bits quantized key value data input from the analyzer 40, generates differential data by performing a predetermined DPCM operation on quantized data, and in step S360, the differential encoding data.

标题编码器400从关键字数据编码器200和关键字值数据编码器300接收译码关键字数据和关键字值数据所必须的信息并在步骤S380中编码该信息。 400 header encoder 300 receives the decoding key data and key value data from the key information necessary for data encoder 200 and the key value data encoder and encoded in step S380 that information.

在下文中,将参考附图更详细地描述分析器40、关键字数据编码器200以及关键字值数据编码器300的结构和操作。 Hereinafter, will be described with reference to the drawings in greater detail the analyzer 40, the key data encoder 200 and the key value data encoder 300 structure and operation.

图4A根据本发明的第一实施例给出了分析器40的一个例子的框图。 Figure 4A shows a block diagram of an example of the analyzer 40 according to a first embodiment of the present invention. 即使使用分析器40在关键字值数据的所有分量(x,y,z及θ)上可执行抽取将被编码的关键字数据和关键字值数据的处理,该处理将在下面的图中,为了方便说明仅考虑关键字值数据(定向内插器)的分量中的一个。 Even with all the components of the analyzer 40 in key-value data (x, y, z, and θ) perform decimation processing on the coded key data and key value data, this process will be in the following diagram, For convenience of description only consider key value data (in the directional interpolator) components of a.

参考图4A,根据本发明的第一实施例,该分析器包括一重新采样器43、一断点抽取器42和一选择器41。 4A, the first embodiment of the present invention, the analyzer includes a resampler 43, a break point extractor 42 and a selector 41. 重新采用器基于一输入定向内插器,将一第一动画路径采样成具有一预定时间量的间隔的多个部分并将所采样的动画路径输出到一关键字编码器200、一关键字值编码器300、以及一标题编码器400中。 Re-use of the device based on an input interpolator orientation, the sample of a first animation path into a plurality of portions having a predetermined amount of time and the interval of the output of the sampled animation path to a key encoder 200, a keyword value encoder 300, and a header encoder 400. 该断点抽取器42抽取一最小量的断点,通过该操作,第一动画路径与基于从第一动画路径抽取的断点生成的一第二动画路径间的一误差能防止超过一预定误差极限,并将抽取断点输出到关键字数据编码器200、关键字值数据编码器300以及标题编码器400中。 The breakpoint extractor 42 extracts a minimum amount of breakpoints, through this operation, an error of the first animation path and a second animation path based on the first animation path generated by the extracted break point between more than one predetermined error can be prevented limit, and outputs to the keyword extraction breakpoint data encoder 200, key value data encoder 300 and a header encoder 400. 该选择器41响应一外部输入信号,将输入定向内插器输出到重新采样器43或断点抽取器42中。 The selector 41 in response to an external input signal, the input to the output of the directional interpolator resampler 43 or the break point extractor 42. 断点抽取器42包括一线性内插器42a、一误差计算器42b以及一确定单元42c。 Breakpoint extractor 42 includes an inner linear interpolator 42a, an error calculator 42b, and a determining unit 42c.

图5A根据本发明第一实施例给出了分析器40的操作流程图。 Figure 5A shows a flowchart of the operation analyzer 40 according to the first embodiment of the present invention. 参考图5A,选择器41在步骤S325从外部接收一定向内插器以及一设定信号。 5A, the selector 41 in step S325 receive an inwardly interpolator and a setting signal from the outside. 该设定信号包括用来确定将被编码的关键字数据和关键字值数据的方法的一生成方法设置信息以及用来确定用于生成将被译码的关键字数据和关键字值数据的模式的一生成模式设置信息。 The setting signal includes a generation method for determining the method to be encoded key data and key value data and means for determining the setting information to be used to generate the key data and decoded key value data pattern A generation mode settings.

首先在下面的段路中描述用于生成关键字数据和关键字值数据的模式。 First described in the following section of road used to generate key data and key value data model.

分析器40通过减小输入到其中的一定向内插器的关键帧的数量来降低将被编码的关键字数据和关键字值数据量。 By reducing the input to the analyzer 40 where a certain number of inwardly interpolator key frames to reduce the amount of key data and key value data to be encoded. 根据从外面输入其中的一模式设定信号,假定分析器40具有一动画路径保存模式或一动画关键字保存模式。 Set according to the input signal from the outside of which a model is assumed analyzer 40 has an animation path-saving mode or a movie keyword save mode.

在一动画路径保存模式中,一定向内插器仅被用来描述一动画路径的内插,且不必随机存取关键帧。 In an animation path save mode, only certain inward interposer is used to describe an animation path within the plug, and do not have random access keyframes. 为有效编码在动画路径保存模式中的一定向内插器,可移除在一预定误差范围内与一动画路径共存的一定向内插器的关键字数据和与该关键字数据相对应的关键字值数据。 For efficient coding in the animation path-saving mode in a certain inward interpolator, can be removed within a predetermined error range with a certain animation path coexist inwardly interpolator and key data and key data corresponding to the keyword word value data.

另一方面,在一动画关键字保存模式中,使用MPEG-4BIFS命令,如'replace','delete'或'insert'随机存取关键帧是必要的。 On the other hand, in a movie by keyword Save mode, using MPEG-4BIFS commands such as 'replace', 'delete' or 'insert' random-access key frame is necessary. 在动画关键字保存模式中,不能改变一定向内插器的关键字数据的数量。 Keyword saving mode in the animation, you can not change a certain number of inward interposer keyword data. 下面将更详细地描述动画路径保存模式以及动画关键字保存模式。 Described in greater detail below animation path-saving mode to save mode and animations keywords.

再参考图5A,选择器41在从外面输入一生成模式后,选择用于生成将被编码的关键字数据和关键字值数据的模式。 Referring again to Figure 5A, the selector 41 in a generation mode input from the outside, the selection for generating key data to be encoded, and key value data pattern. 在步骤S330中,选择器41将输入定向内插器输出到断点抽取器42中,如果输入生成模式是一动画关键字保存模式。 In step S330, the input selector 41 is inserted in the output directed to break the extractor 42, if you enter the keyword generation mode is a movie-saving mode. 如果输入模式是一动画路径保存模式,向应从外面输入的一生成方法设定信号,选择器41在步骤S330中将输入定向内插器连同生成关键字数据和关键字值数据所必需的信息输出到重新采样器43或断点抽取器42中。 If the input mode is an animation path to save the model, a method of generating the signal should be set outside input, the input selector 41 is inserted inside the directional information, together with the output data to generate keywords and keyword values necessary data in step S330 will to the resampler 43 or the break point extractor 42.

特别地,在通过重新采样生成将被编码的关键字数据和关键字值数据的情况下,选择器41将关键字数据的数量(即时间间隔)和一生成模式连同定向内插器输出到该重新采样器43中。 In particular, in the case generated by re-sampling will be coded key data and key value data, select the number of 41 key data (ie, the time interval), and within a generation mode together with directional interpolator output to the resampler 43. 在通过抽取断点生成将被编码的关键字数据和关键字值数据的情况下,选择器41将在一原始动画路径和通过抽取断点生成的一路径间的一临界误差以及该生成模式输出到断点抽取器42。 In breakpoints generated by extracting key data to be encoded and key-value data, the selector 41 outputs a critical error in one of the original animation path and a path generated by extracting and break between the generation mode to break extractor 42.

在步骤S335中,重新采样器43在多个预定时间间隔通过采样由从选择器41输入的定向内插器生成的一动画路径生成的采样关键字数据和采样关键字值数据,以及断点抽取器42抽取一最少断点,通过该操作,能防止由输入定向内插器生成的一动画路径和由抽取断点生成的一动画路径间的一误差超过一预定误差极限。 In step S335, the resampler 43 at predetermined time intervals by a plurality of sampled key data and sampled key value data by sampling an animation path generated from the orientation interpolator generated from the input selector 41, and the breakpoint extraction 42 extracts a minimum break, through this operation, an error can be prevented by the input orientation interpolator and an animation path generated by a an animation path generated by the extracted break point between the error exceeds a predetermined limit.

图5B根据本发明的一优选实施例给出了重新采样器43的操作流程图。 Figure 5B according to a preferred embodiment of the present invention shows the flowchart of the operation resampler 43. 参考图5B,重新采样器43在步骤S502中从选择器41接收将被重新采样的一定向内插器和关键字数据的数量(m)。 5B, a re-sampler 43 in step S502 receives the number (m) to be resampled must inwardly of the interpolator and key data from the selector 41. 将被重新采样的关键字数据的数量(m)可由一用户任意设置或预先被设置为一预定值上。 The number (m) of key data to be resampled is arbitrarily set by a user or previously set at a predetermined value.

在步骤S504中,重新采样器43选择由输入定向内插器生成的一原始动画路径的一第一路径点和一最终路径点并将将被重新采样的该关键字数据的初始值(i)设置为1。 The initial value in step S504, the resampler 43 selects a first path point and a final path point by the input orientation interpolator generated within an original animation path and will be re-sampled data to the keyword (i) is set to 1.

此后,在步骤S506,重新采样器43在多个预定时间的一间隔上生成第I个关键字数据。 Thereafter, in step S506, the resampler 43 generates the I-th key data at a plurality of predetermined time intervals.

图6A是描述原始关键字数据和重新采样关键字数据的图。 6A is a description of the original key data and resampled key data of FIG. 由于该输入定向内插器的关键字数据表示在一时间轴上关键帧的位置,如图6A所示,关键字数据单调地增加,而该关键字数据中的间隔却没有规律。 Since the input orientation interpolator to the key data in the key frame represents the position of a time axis, as shown in Figure 6A, the key data monotonously increase, but the key data in the interval was not studied.

因此,如图6A所示,重新采样器43通过按将被重新采样的关键字数据的数量来划分分别表示在步骤S504中选择的第一路径点和最终路径点的关键字数据间的一差值来获得多个预定时间的一间隔并因此在多个预定时间的间隔上重新采样将被重新采样的关键字数据。 Thus, as shown, resampler 43 by pressing the number of the key data to be resampled to divide the keyword data represent the first path point selected in step S504, and the final points of the path difference between a respectively 6A value to obtain a plurality of predetermined time interval and a re-sampling will therefore be resampled key data on a plurality of predetermined time intervals.

在步骤S508中,重新采样器43通过使用原始动画路径按线性内插器重新采样生成的与该关键字数据相对应的关键字值数据。 In step S508, the resampler 43 by using the original animation path by linear interpolation re-sampling the data generated by the key corresponding to the key value data. 换句话说,与重新采样关键字数据相对应的关键字值数据使用正好在重新采样关键字数据后与关键字数据相对应的关键字值数据以及正好在重新采样关键字数据前与关键字数据相对应的关键字值数据被线性内插。 In other words, with the re-sampling key data corresponding to key value data just after the re-sampled using keywords and keyword data corresponding to the data key value data, and just before the key data and resampled key data corresponding key value data is linear interpolation.

此后,在步骤S510中,重新采样器43验证是否已经在所有将被重新采样的关键字数据上执行重新采样处理并重复执行步骤S506和S508直到重新采样所有关键字数据以及他们相应的关键字值数据。 Thereafter, in step S510, the resampler 43 verify resampling processing has been performed on all the key data to be resampled and steps S506 and S508 until all the key data and resampled key values of their corresponding repeat data.

图5C根据本发明的第一实施例给出了一种抽取断点的方法流程图,图7A至7F根据本发明的一优选实施例给出了从一定向内插器抽取断点的每个步骤的框图。 Figure 5C shows a flow chart of a method of extracting break point according to the first embodiment of the invention, Figs. 7A to 7F accordance with a preferred embodiment of the present invention gives a certain inward interpolator extracted from each breakpoint block diagram of the steps.

参考图4A,5C以及7A至7F,断点抽取器42的线性内插器42a在步骤S520从选择器41接收一定向内插器和一临界误差eth。 With reference to FIG. 4A, 5C, and 7A through 7F, the linear breakpoint extractor 42 receive an interpolator 42a inwardly of the interpolator and a critical error eth from the selector 41 in step S520. 由输入定向内插器所构成的一动画路径如图7A所示。 7A, by an animation path within the input directional interpolator posed.

在步骤S522中,线性内插器42a抽取由输入定向内插器所组成的动画路径的第一路径点Q0以及最终路径点Qn,如图7A所示,并将一计数器(i)设置为1。 In step S522, the linear interpolator 42a extracts from the input orientation interpolator consisting of the first path point Q0 and the final path of the animation path point Qn, 7A, and a counter (i) is set to 1 .

在步骤S524中,线性内插器42a在第一Q0和最终路径点Qn间依次地任意或顺序选择路径点。 In step S524, the linear interpolator 42a arbitrarily or sequentially selects path points between the first path point Q0 and the final Qn sequentially. 接着,在步骤S526中,线性内插器42a利用所选择的路径点线性地内插还未被选择的路径点并将所选择的路径点和内插路径点输出到误差计算器42b中。 Outputs Next, in step S526, the linear interpolator 42a, the path inside the path using the selected point linearly interpolated has not been selected and the selected path points and the interpolated path points to the error calculator 42b.

在步骤S528中,误差计算器42b计算原始动画路径和由选择路径点和内插路径点所构成的一候选动画路径间的一误差(e)并将该误差(e)输出到确定单元42c中。 In step S528, one error (e) error calculator 42b to calculate the original animation path and a candidate animation path by selecting the path points and interpolating path points between posed and the error (e) to determine the output unit 42c, . 下面将描述计算误差(e)的方法。 Calculating an error (e) of the method will be described below.

误差计算器42b核对在未被线性内插器42a所选择的路径点中,是否仍然存在当计算误差(e)时未被考虑的路径点。 Check in error calculator 42b is not linear interpolation of the selected points 42a, path, the path is still present when the calculation error (e) is not taken into account when points. 如果有当计算误差(e)时未被考虑的路径点,误差计算器42b通过重复执行步骤S524至S528在步骤S530中计算路径点与原始动画路径间的一误差。 If there is a path when calculating the error (e) is not considered a point when the error calculator 42b Repeat steps S524 to S528 to calculate an error path between points and the original animation path in step S530 through.

图7C是描述步骤S524至S530的图。 7C is a description of steps S524 to S530 in FIG. 参考图7C,线性内插器42a抽取与在预定时间瞬间k1的关键字数据相对应的一断点Q1并通过线性内插第一路径点Q0和断点Q1间的路径点来生成一第一候选动画路径。 With reference to FIG. 7C, the linear interpolator 42a extracts a predetermined time instant k1 keyword data corresponding to a break point Q1 and point by linear interpolation path a first path point Q0 and the break point Q1 is generated between a first candidate animation path. 误差计算器42b计算原始动画路径与第一候选动画路径间的一误差e1。 Error calculator 42b to calculate an error e1 original animation path and the first candidate animation paths. 此后,用相同的方式,线性内插器选择另一断点Qk并通过线性内插在第一路径点Q0与断点Qk间和断点Qk和最终路径点Qn间的路径点来生成第k个候选动画路径。 Thereafter, in the same manner, the linear interpolator selects another break point Qk and by linearly interpolating path points between the first path point Q0 and the break point Qk and the break point Qk and the final path is generated between the first Qn k candidate animation path. 误差计算器42b计算原始动画路径与第k个候选动画路径间的一误差(ek)。 Error calculator 42b to calculate an error original animation path and the k-th candidate animation path between (ek).

如果步骤S524至S530已经在未被线性内插器42a选择的所有路径点上执行过,原始动画路径和在步骤S524至S530生成的每个候选动画路径间的误差被输出到确定单元42c。 If all paths point to step S524 S530 has not been inserted in the selected linear 42a performed on the original animation path and error for each candidate animation path generated in step S524 to S530 is output to between determining unit 42c. 然后,在步骤S532中,确定单元42c选择形成具有最小误差和原始动画路径的一候选动画路径的一断点,并将计数器(i)值加1。 Then, in step S532, the determination unit 42c to select a breakpoint is formed of a candidate animation path having the smallest error and the original animation path, and the counter value (i) is incremented.

在步骤S534中,确定单元42c核对原始动画路径和由抽取断点构成的候选动画路径间的一误差(e)是否大于临界误差eth以及计数器(i)的值是否大于关键字数据的数量(n)即第一路径点Q0与最终路径点Qn的路径点的数量。 In step S534, the determination unit 42c to check the original animation path and the candidate animation path constituted by the extracted break point between an error (e) is greater than a critical error eth and the counter (i) is greater than the value of the key data of the number (n ) that the number of the first path point Q0 and the final path points Qn path points.

如果误差(e)小于临界误差eth,它表示已经抽取要求编码的所有断点。 If the error (e) is less than the critical error eth, it said it had extracted all breakpoints require coding. 如果最终选择作为将被编码的断点的数量等于“n”,表示完成抽取所有路径点的断点的处理。 If the final choice will be encoded as equal to the number of breakpoints "n", to indicate the completion of all the path points extracted breakpoint processing.

然而,如果抽取断点的数量小于n以及误差(e)大于临界误差eth,这表示还存在将被抽取的断点,将所选择的断点输出到线性内插器42a,然后再执行步骤S524至S532。 However, if the number is less than n extracted breakpoints and error (e) is greater than the critical error eth, which means that there will be extracted breakpoints, the selected breakpoint output to linear interpolator 42a, and then perform step S524 to S532.

在下文中,在下面的段路中将描述当生成模式是一动画路径保存模式时,假定从重新采样器43和断点抽取器42输出到关键字值数据译码器的数据。 In the following, in the section of road in the following description when generating mode is an animation path save mode, assuming that the output from the re-sampler 43 and 42 to break extractor key value data decoder data.

重新采样器43将采样的关键字数据和采样的关键字值数据分别输出到关键字数据编码器200和关键字值数据编码器300中分别作为将被编码的关键字数据和关键字值数据。 Resampler 43 sampled key data and key value data samples are output to the key data encoder 200 and the key value data encoder 300, respectively, as key data to be encoded, and key value data.

在下文中,将参考图8来描述根据一生成模式从断点抽取器42输出的关键字数据和关键字值数据。 In the following, will be described with reference to FIG. 8 is generated based on keyword data model output from a breakpoint extractor 42 and key value data.

如图8所示,假定最终抽取断点被称为0,3,6和8,与断点0,3,6和8相对应的关键字数据和关键字值数据连同一关键字选择特征位被输出,如下表如示。 8, assume that the final extraction is called breakpoints 0,3,6 and 8, with breakpoints 0,3,6 and 8 the data corresponding to the key data and key value together with a key selection flag is output, as shown in the following table.

表2 Table 2

上面已经根据本发明的第一实施例描述过分析器40的结构。 Have been described above structure analyzer 40 according to a first embodiment of the present invention. 然而,对本领域的技术人员来说,分析器40可仅由断点抽取器42而没有选择器41和重新采样器43或仅由重新采样器43而没有选择器41和断点抽取器42组成。 However, those skilled in the art, the analyzer 40 may be only the breakpoint extractor 42 without the selector 41 and the resampler 43 or only by the resampler 43 without the selector 41 and the break point extractor 42 Composition .

在下文中,将根据本发明的第二实施例来描述分析器40的另一例子。 Hereinafter, the second embodiment according to the present invention will be described another example of the analyzer 40.

参考图4B,根据本发明第二实施例的分析器40包括一重新采样器45以及一断点抽取器46。 4B, the analyzer according to the second embodiment of the present invention 40 comprises a re-sampler 45 and a break point extractor 46. 该重新采样器45接收和重新采样一定向内插器。 The resampler receives and resampling interpolator 45 must inwardly. 该断点抽取器46抽取重新采样的定向内插器的断点并输出将被编码的关键字数据和关键字值数据。 The breakpoint extractor 46 extracts directional interpolator resampling breakpoints and output will be encoded key data and key value data. 在本发明的第二实施例中,断点抽取器46象在本发明的第一实施例中一样,也包括一线性内插器46a、一误差计算器46b、以及一确定单元46c。 In the second embodiment of the present invention, the breakpoint extractor 46 as in the first embodiment of the present invention, also includes an interpolator 46a, an error calculator 46b, and an inner linear determining unit 46c.

当一定向内插器被输入到分析器40中时,重新采样器45将由该定向内插器所组成的一第一动画路径重新采样成彼此之间具有多个预定时间的一间隔的多个预定部分。 When a certain inward interpolator is input into the analyzer 40, the resampler 45 by a first animation path to the interior orientation interpolator consisting of the re-sampling into a plurality of having a plurality of predetermined time intervals between each other a predetermined portion.

重新采样器45将由采样关键字数据和采样关键字值数据组成的定向内插器输出到断点抽取器46的线性内插器46a。 Directional interpolator resampling 45 by sampling key data and key value data sample consisting of a linear interpolator output to break extractor within 46 46a.

线性内插器46a通过执行如图5C所示的步骤S522至S526来内插一定向内插器并将该内插的定向内插器输出到误差计算器46b中。 Step linear interpolator 46a shown in FIG. 5C by performing interpolation must inwardly interpolator S522 to S526 to the interpolation and the directional interpolator output to the error calculator 46b. 误差计算器46b通过执行步骤S528至S530计算第一动画路径和由内插的定向内插器所构成的一第二动画路径间的一误差。 Error calculator 46b to calculate an error a second animation path first animation path and orientation interpolator posed by the interpolation between the steps S528 through to S530. 确定单元46c选择将导致第一和第二动画路径间的一最小误差的一路径点,验证相应的误差是否大于一临界误差eth以及是否已经选择第一动画路径的所有路径点并生成将被编码的关键字数据和关键字值数据。 Will result in the determination unit 46c to select a point of a minimum error path of the first and second animation paths, and verify the corresponding error is greater than a critical error eth and if all path points have been selected and generates a first animation path to be coded The key data and key value data.

如上所述,在根据本发明的第二实施例的分析器40中,除断点抽取器46接收由从重新采样器45输出的关键字数据和关键字值数据构成的一定向内插器以及在由从重新采样器45输入的定向内插器构成的一动画路径上执行抽取断点的处理外,重新采样器45和断点抽取器46的操作与在本发明的第一实施例中的相应组件的操作相同。 As described above, according to the second embodiment of the present invention, the analyzer 40, except a certain breakpoint extractor 46 receives data from the interpolator inwardly key data output from the resampler 45 and the key value and constituted On a motion path from the directional interpolator input from the resampler 45 constituted execution breakpoint extraction processing, the operator re-sampler 45 and the extractor 46 with the breakpoint in the first embodiment of the present invention the same as the corresponding components of the operation.

在下文中,根据本发明的第三实施例,参考图4C来描述分析器40的一个例子。 Hereinafter, according to the third embodiment of the present invention, will be described with reference to FIG. 4C an example of the analyzer 40.

参考图4C,分析器40包括一断点抽取器48,其接收一定向内插器,从由该定向内插器构成的一第一动画路径抽取断点并输入关键字数据和关键字值数据,以及一重新采样器49,其在一预定时间间隔采样由包括从断点抽取器48输入的关键字数据和关键字值数据的一定向内插器构成的一第二动画路径。 With reference to FIG. 4C, the analyzer 40 includes a break point extractor 48, which receive an inwardly interpolator, extracts break from a first animation path by the inner configuration of the orientation interpolator and the input key data and key value data , and a resampler 49, which samples at a predetermined time interval by a second animation path including key data and key value data input from the break point extractor 48 is inserted inwardly of certain constituted. 断点抽取器48,与在本发明的第一和第二实施例中一样,也包括一线性内插器48a、一误差计算器48b以及一确定单元48c。 Breakpoint extractor 48, in the first and second embodiments of the present invention, as also including the linear interpolator 48a, an error calculator 48b, and a determining unit 48c.

断点抽取器48,与本发明第一实施例中的一样,将从第一动画路径中所抽取的关键字数据和关键字值数据输出到重新采样器49中。 Breakpoint extractor 48, and the present invention is the same as the first embodiment, the output from a first animation path data and the extracted keyword data to the keyword values resampler 49.

重新采样器49在预定时间间隔重新采样由包括从断点抽取器48所输入的关键字数据和关键字值数据的一定向内插器所构成的一动画路径并输出将被编码的关键字数据和关键字值数据。 Resampler 49 by a re-sampling interval of the animation path comprising interpolator inwardly from a certain keyword data breakpoint extractor 48 and the input key value data and outputting the composed key data to be encoded at a predetermined time and key value data. 重新采样器49的功能与在本发明的第一和第二实施例中相同,因此这里不再重复描述。 Re-sampler 49 functions in the first and second embodiments of the present invention, the same description is not repeated here.

在本发明的第一到第三实施例中从分析器40输出的关键字数据和关键字值数据被分别输出到关键字数据编码器200和关键字值数据编码器300。 In the first to third embodiments of the present invention is a data analyzer 40 outputs the output key and key value data, respectively, to the key data from the encoder 200 and the key value data encoder 300.

在下文中,将参考图9A至12J来描述关键字数据编码器200的结构和操作。 Hereinafter, with reference to FIG. 9A to 12J keywords to describe the structure and operation of the data encoder 200.

图9A根据本发明的一优选实施例给出了一关键字数据编码器的框图。 Figure 9A in accordance with one preferred embodiment of the present invention shows a block diagram of a key data encoder. 参考图9A,一关键字数据编码器200包括一线性关键字编码器900、一量化器910、一DPCM处理器920、一移位器930、一折叠处理器940、一DND处理器950以及一熵编码器960。 With reference to FIG. 9A, a key data encoder 200 includes a linear key encoder 900, a quantizer 910, a DPCM processor 920, a shifter 930, a folding processor 940, a processor 950 and a DND entropy encoder 960.

线性关键字编码器900识别在整个关键字数据范围内部关键字数据线性增加的一部分并编码该部分。 Keywords linear encoder 900 identifies key data throughout the range of the internal part of the key data and encode linear increase in the part. 量化器910利用能最小化一量化误差的一量化方法来量化输入到其中的关键字数据。 The quantizer 910 can minimize the use of a quantization error of a quantization method for quantizing the key data inputted thereto. DPCM处理器920接收量化关键字数据并生成关键字数据的差分数据。 DPCM processor 920 receives quantized key data and generates differential data of key data. 移位器930从差分数据中减去在所有差分数据中具有最大频率的一差分数据。 Shifter 930 subtracts having the maximum frequency among all differential data from the differential data in a differential data. 折叠处理器940将所有差分数据传送到一正数区或一负数区。 Fold difference processor 940 will transfer all the data to a positive or a negative zone area. DND处理器950通过执行一划分操作然后有选择地执行一分割操作或向下分割操作来减小关键字数据的差分数据的范围。 DND processor 950 by performing a divide operation and then selectively performing a dividing operation or a split operation to reduce the down key data of the differential data range. 熵编码器960在每个位平面上使用一函数SignedAAC或UnsignedAAC来编码差分数据。 The entropy encoder 960 uses a function SignedAAC or UnsignedAAC to encode differential data on each bit plane.

在下文中,将参考图10A来更详细地描述关键字数据编码器200的操作。 Hereinafter, the operation key data encoder 200 will be described in more detail with reference to FIG. 10A. 图10A是根据本发明的一优选实施例的编码关键字数据的方法的流程图。 10A is the invention according to a preferred coding key data flow diagram of a method embodiment. 当将关键字数据输入到用于编码一定向内插器的一装置中时,信息如关键字数据的数量以及每个关键字数据的数字的数量被输入到标题编码器400中并被编码。 When the key data input means for encoding a certain inward when the interpolator, information such as the number of the number of key data and digital data of each keyword is input into the header encoder 400 and is encoded. 在步骤S9000中,线性关键字编码器900在该输入关键字数据中查找在某一时间间隔中存储关键帧的一区域,关键字数据具有相同的差值,以及关键字数据线性改变,以及所查找的线性部分被首先编码。 In step S9000, a linear encoder 900 keywords to find the data in the input keywords in one area at a time interval stored key frames, key data has the same difference, as well as key data linear change, as well as linear portion of the search is first encoded.

著名的3D应用软件,如3DMax或Maya,利用在特定区域间具有预定时间间隔的关键字来生成基于动画的关键帧。 The famous 3D applications, such as 3DMax or Maya, between the use of keywords in a specific area of a predetermined time interval to generate keyframe-based animation. 在这种情况下,有可能很容易使用一线性关键字数据区域的开始和结尾关键字数据以及存在于它们间的关键帧的数量来编码关键字数据。 In this case, it is possible to easily use the beginning and ending key data of a linear key data region and the number of key frames present in between them for coding key data. 因此,线性预测对使用一定向器来编码在某一区域中的关键字非常有用。 Thus, linear prediction using the device to encode a certain region in a keyword is useful.

下面的等式用于线性预测。 The following equation for linear prediction.

在这里,tS表示一部分线性区域开始的一关键字的数据,tE表示该部分线性区域结束的一关键字的数据,S表示tS的一指数,以及E表示tE的一指数。 Here, tS represents a key part of the linear region of the data begins, tE represents a key part of the data for the end of the linear region, S represents an index tS, tE and E represents an index. 在从第S个关键字数据到第E个关键字数据的特定区域中的实际关键字数据与用等式(4)线性预测的关键字数据间的误差能用下面的等式计算。 In a particular area from the S-th key data to the first data E keywords in the actual key data and using Equation (4) linear prediction error between the key data can be calculated following equation.

如果在使用等式(5)计算的误差中的一最大值不大于一预定误差极限,ti能被认为在区域[tS,tE]或在某一误差范围内共线。 If the error in using equation (5) in a calculated maximum value is not greater than a predetermined error limit, ti can be considered in region [tS, tE] or colinear within a margin of error. 使用下述等式(6)能确定最大误差值ti是否与特定区域共线。 Using the following equation (6) able to determine the maximum error value ti is co-linear with the specific region.

如果ti是与区域[tS,tE]共线。 If ti is the region [tS, tE] collinear. 此时,nBits表示用于编码的位的数量。 At this time, nBits represents the number of bits used for encoding.

如果线性关键字编码器900查找部分线性区域,将部分线性关键字数据区域的开始和结尾关键字数据输出到浮点数转换器905中。 If the linear encoder 900 keywords to find part of the linear region, the output of the start and end of the keyword data portion of the linear key data area to the floating-point converter 905. 包含在线性关键字数据区域中的关键字的数量被输出到标题编码器400中并被编码。 The number of data included in the linear region of the keyword keyword is output to the title encoder 400 and encoded. 使用线性编码有可能相当大地降低将被编码的数据量。 Using a linear coding it is possible to considerably reduce the amount of data to be encoded.

使用浮点数变换来编码开始关键字数据和结尾关键字数据,这将在下面描述。 Using the floating point transform coding start and end of the keyword data key data, which will be described below.

浮点数转换器905将用二进制表示的关键字数据转换成十进制以便编码开始关键字数据和结尾关键字数据。 Float keyword data converter 905 that converts binary coded decimal to the beginning and end of key data key data.

一计算机将浮点数存储为32位二进制数。 A computer will be stored as 32-bit binary floating-point number. 如果给出用二进制表示的浮点数,浮点数转换器905将浮点数转换成十进制的一尾数和一指数,该处理用下面的等式来表示。 If given binary floating-point representation, floating point converter 905 converts the floating point decimal and fraction and an exponent, the process is represented by the following equation.

例如,一浮点数12.34能用一计算机转换成一二进制数,如下所示。 For example, one can use a computer float 12.34 converted into a binary number, as shown below.

0 10001010111000010100011 10000010 0 10000010 10001010111000010100011

1 2 3 123

1:特征位 1: flag

2:二进制的尾数 2: binary mantissa

3:二进制的指数 3: binary index

用等式(7)能将一二进制数转换成一十进制数,如下所示。 Using equation (7) can convert a binary number to a decimal number, as shown below.

0 1234 2 012,342

1 2 3 123

1:特征位 1: flag

2:十进制的尾数 2: decimal mantissa

3:十进制的指数 3: Decimal index

为在一位流中包括十进制的一尾数和一指数,必须计算为表示该尾数和指数而所需的位的数量。 Including a mantissa and a decimal exponent in a stream, you must calculate the number of bits representing the mantissa and exponent and needed. 指数具有在-38到38间的一值,因此能用7位来表示它及其特征位。 Index has a value of -38 to 38, so it can be used to represent the seven and flag. 为表示尾数而所需的位的数量由位数而定。 The number of bits required to represent the mantissa and the number of bits may be. 尾数值以及为表示尾数而所需的位的数量如下表所示。 Mantissa values and the number of bits required to represent the mantissa and the following table.

表4 Table 4

在图10B中所示的下述编码处理来编码利用上述处理已经查找过和转换过的线性关键字数据区域的开始和结尾关键字数据,并输出到报文编码器400,且在该位流中存储。 In the following the encoding process shown in FIG. 10B to encode the above-described process has been to find the start and end of the keyword and the converted data of the linear key data region, and output to the message encoder 400, and the bit stream in storage.

图10B表示在浮点数转换器905中执行的编码两个输入浮点数的一处理。 10B shows the encoding performed in floating-point converter 905 to enter a deal with two floating-point numbers. 参考图10B来描述浮点数转换器905编码一浮点数的方法。 Referring to FIG. 10B to describe the floating-point converter 905 encodes a floating-point approach.

浮点数转换器905接收原始关键字数据、开始关键字数据S以及结尾关键字数据E的数字数Kd并在步骤S9040中用等式(7)将转换它们。 Floating-point number converter 905 receives the original key data, beginning key data S, and ending key data E and converts digital data Kd them in step S9040 by the equation (7).

浮点数转换器905首先编码S。 Float encoding converter 905 S. First 具体来说,浮点数转换器905核对S的数字数目是否不同与Kd。 Specifically, the floating-point digital converter 905 to check whether the number of S is different from Kd. 如果S的数字数目与Kd不同,在步骤S9042中获得S的数字数目并输出到标题编码器400。 If the number different from the number of S and Kd, the number of digits to obtain S in step S9042 and output to the header encoder 400. 利用函数Digit(),浮点数转换器905获得S的数字数目。 Using the function Digit (), the number of floating-point digital converter 905 S obtained.

如果S的数字数目大于7,在步骤S9043中使用多个预定位(在本发明中,用IEEE754标准的一浮点数方式使用32位)将S输出到标题编码器以便将能S的数字数目包括在该位流中。 If the number is greater than the number of S 7, the use of a plurality of predetermined bits in step S9043 (in the present invention, with the IEEE754 standard manner using a 32-bit floating-point number) S is output to the header encoder to the number of digits can include S In the bit stream.

在步骤S9100中,量化误差最小化器915使用用于控制一量化范围的方法来预先量化或逆量化该输入关键字数据以便能最小化该量化误差。 In step S9100, the quantization error minimizer 915 controls a quantization range using the method for a pre-quantization or inverse quantization of the input keyword data so as to minimize the quantization error.

具体来说,如果用于量化的一混合最大值用Max表示,用于量化所控制的一最小值用Min表示,一输入值用Xi表示,以及用于量化的位的数量用nQuantBit表示,那幺能使用下面的等式来获得一量化输入值 Specifically, if a mixed maximum value used for quantization represented by Max, a control for quantifying the minimum value represented by Min, an input value is represented by Xi, and the number of bits for quantization is represented by nQuantBit, that Unitary can be obtained using the following equation a quantized input value

一逆量化值 An inverse quantization values

以及一误差ei。 And an error ei.

有两种用于最小化误差和∑ei的方法。 There are two kinds of error minimization method and Σei for. 一种方法是通过连续控制Min来降低误差和直到误差和最小化为止的方法。 One method is to reduce by continuously controlling Min until the error and the error is minimized and a method so far. 另一下如下。 Another down below.

首先,假定Xi=(i+n)ΔX+εi,其中Xi表示一输入关键字数据序列,Δx表示输入数据的基本步长,n是一任意整数,以及εi表示零平均值随机噪声。 First, assume that Xi = (i + n) ΔX + εi, where Xi represents a key input data sequence, Δx represents a basic step input data length, n is an arbitrary integer, and εi represents zero mean random noise.

接着,当di=Xi-Xi-1=ΔX+(εi-εi-1)时,Δ′x=E[di]以及Min=Max-Δ′x*(2nQuantBit-1)。 Next, when di = Xi-Xi-1 = ΔX + (εi-εi-1) when, Δ'x = E [di] and Min = Max-Δ'x * (2nQuantBit-1).

Min能使最小化一量化误差成为可能,以及Max能被输入到量化器910中并被用于关键字值数据的量化。 Min minimize a quantization error to make it possible, and Max can be input to the quantizer 910 and is used to quantify the key value of the data.

在步骤S9200中量化器910接收最大和最小值Max以及能最小化一量化误差的Min并用下述等式(9)量化关键字数据fKeyi。 910 receives maximum and minimum values Max in step S9200 and the quantizer that minimizes a quantization error and Min quantized key data fKeyi by the following equation (9).

在这里,i表示已量化关键字数据的一指数,nQkeyi表示已量化关键字数据的一整数数组,fKeyi表示已量化关键字数据的一浮点数数组,fKeyMax表示从量化误差最小化器915输入的一最大值,fKeyMin表示从量化误差最小化器输入的一最小值,以及nKeyQBit表示一量化步长。 Here, i represents the quantized data of a keyword index, nQkeyi represents an integer array of quantized key data, fKeyi represents a floating-point array of quantized key data, fKeyMax 915 represents input from the quantization error minimizer of a maximum, fKeyMin represents a minimum value from the input quantization error is minimized, and nKeyQBit represents a quantization step. 在等式(9)中,函数floor(v)是输出不大于某一浮点值v的一最大整数的一函数。 In Equation (9), function floor (v) is the output is not greater than a certain floating-point value v is a function of a maximum integer.

本发明的量化器910可能不使用用于降低一量化误差的这种算术,在这种情况下,简单使用输入关键字数据中的最大和最小值fKeyMax和fKeyMin来执行量化。 Quantizer 910 of the present invention may not use for reducing a quantization error of this arithmetic, in this case, simply using the input key data is the maximum and minimum values fKeyMax and fKeyMin to perform quantization.

参考图10C来更全面地描述本发明的量化处理。 With reference to FIG. 10C quantization process more fully described the present invention.

量化器910在步骤S9210中接收关键字数据并在步骤S9220中核对最大和最小值MAX和MIN是否从量化误差最小化器915中输入。 Quantizer 910 receives the key data in step S9210 and check the maximum and minimum is 915 MIN MAX and input from the quantization error is minimized in the step S9220.

如果输入MAX和MIN,在步骤S3230中量化器910设置最大和最小值fKeyMax和fKeyMin分别用作量化的MAx和MIN并将新设置的最大和最小值fKeyMax和fKeyMin输出到浮点数转换器905中。 If you enter the MAX and MIN, in step S3230 quantization 910 set the maximum and minimum fKeyMax and fKeyMin are used as quantitative MAx and MIN and the new set of maximum and minimum fKeyMax and fKeyMin output to the floating-point converter 905. 通过上述浮点数转换方法来转换和编码最大和最小值fKeyMax和fKeyMin并输出到标题编码器400中以便它们能被包括在用在编码中的一关键字标题中。 By the above-mentioned floating-point number conversion method for converting and encoding the maximum and minimum values fKeyMax and fKeyMin and output to the header encoder 400 so that they can be included with the coding of a keyword in the title.

如果没有值从量化误差最小化器915中输入,在步骤S9240中,量化器910设置第一关键字数据fKey0和最终关键字数据fKeyN-1分别作为最小fKeyMin和最大值fKeyMax。 If no value is entered 915 from the quantization error minimizer, in step S9240, the quantizer 910 sets the first key data fKey0 and final key data fKeyN-1, respectively, as the minimum and maximum values fKeyMin fKeyMax. 如果最大值fKeyMax不小于1或不大于0,将最大和最小值fKeyMax和fKeyMin输出到浮点数转换器905中并通过上述浮点数转换方法转换和编码。 If the maximum value fKeyMax not less than 1 or greater than 0, the maximum and minimum fKeyMax and fKeyMin converter output to 905 floating-point conversion and encoding and floating-point conversion by the above method. 接着,在步骤S9260中,已经被转换和编码的最大和最小值fKeyMax和fKeyMin被包括在关键字标题中以便它们能被用在译码中。 Next, in step S9260, the already converted and coded maximum and minimum values fKeyMax and fKeyMin to be included in the key header so that they can be used in decoding.

另一方面,如果最大值fKeyMax小于1且最小值fKeyMin大于0,在步骤S9270中核对表示最大和最小值fKeyMax和fKeyMin是否将被包括在正用在译码中的关键字标题中的一标志位。 On the other hand, if the maximum value fKeyMax is smaller than 1 and the minimum value fKeyMin is greater than 0, in step S9270 the check indicating whether or not the maximum and minimum values fKeyMax and fKeyMin will be included in the decoding is used in the title of a keyword flag . 如果该标志位被设置以致最大和最小值fKeyMax和fKeyMin能被包括在关键字标题中,执行步骤S9260以便最大和最小值fKeyMax和fKeyMin能被输出到标题编码器400中。 If this flag is set so that the maximum and minimum values fKeyMin fKeyMax and can be included in the keyword in the title, in order to step S9260 fKeyMax maximum and minimum values can be output to the title and fKeyMin encoder 400. 如果标志位未被设置,量化器910不允许将最大和最小值fKeyMax和fKeyMin包括在关键字标题中。 If the flag is not set, the quantizer 910 does not allow the maximum and minimum fKeyMax and fKeyMin include the keyword in the title.

在最大和最小值fKeyMax和fKeyMin未被包括在关键字标题中的情况下,关键字数据编码器和关键字数据译码器被假定分别执行编码和译码、分别将最大和最小值fKeyMax和fKeyMin设置为1和0。 The maximum and minimum values fKeyMax and fKeyMin are not included under the heading in the case of the keyword, key data encoder and the key data decoder is assumed to perform encoding and decoding, respectively, are the maximum and minimum values fKeyMax and fKeyMin set to 1 and 0. 在这种情况下,在步骤S9280中,量化器910分别将最大和最小值fKeyMax和fKeyMin设置为1和0。 In this case, in step S9280, the quantizer 910, respectively, the maximum and minimum values fKeyMax and fKeyMin is set to 1 and 0. 对关键字数据译码器来说最大和最小值fKeyMax和fKeyMin已经公知以致它们不需要被包括在关键字标题中。 Keyword data decoder for maximum and minimum fKeyMax and fKeyMin have known that they do not need to be included in the keyword in the title.

在步骤S9290中,量化器910通过将通过上述处理设置的最大和最小值fKeyMax和fKeyMin代入等式(9)中来量化输入关键字数据并将量化关键字数据输出到一DPCM处理器920中。 In step S9290, the quantizer 910 by the maximum and minimum values into equation by substituting the above-described processing fKeyMax and fKeyMin set (9) to quantize the input key data and outputs the quantized key data to a DPCM processor 920.

DPCM处理器920接收量化关键字数据并在多个预定时间在量化关键字数据上执行DPCM。 DPCM processor 920 receives quantized key data and performs DPCM on the quantized key data in a plurality of predetermined time. 接着,DPCM处理器920输出DPCM的次数,通过该操作,在离散的程序中获得一最小值以及在每个DPCM周期中获得的内部关键字数据到标题编码器400中。 Next, the DPCM processor 920 outputs the number of DPCM, by this operation, to give a minimum value and the internal key data obtained in each cycle of DPCM in the header encoder 400 in discrete program. 在步骤S9300中,DPCM处理器920将由DPCM生成的差分数据输出到移位器930中。 In the step S9300, DPCM processor 920 generates differential data by DPCM to the shifter 930 output.

参考图10D,在步骤S9310中,DPCM处理器920在多个预定时间在输入关键字数据上执行DPCM并存储DPCM的周期的数量作为DPCM的次数。 With reference to FIG. 10D, in the step S9310, DPCM processor 920 at a predetermined time in the plurality of DPCM performed on the input key data and the number of storage cycles of DPCM as the number of DPCM. 在本发明的一优选实施例中,DPCM能被执行三次。 In a preferred embodiment of the present invention, DPCM be performed three times.

此后,在步骤S9250中,量化器910核对最大值fKeyMax是否小于1而大于0以及最小值是fKeyMin否大于0。 Thereafter, in step S9250, the quantizer 910 to check the maximum value fKeyMax is smaller than 1 but greater than 0, and whether the minimum value fKeyMin is greater than 0. 在这里,离散度可由离散、标准偏差或四分位差以及在本发明的一优选实施例中,可使用四分位差。 Here, the dispersion may be discrete, or quartile deviation, and the standard deviation in a preferred embodiment of the present invention, quartile deviation may be used.

接着,DPCM处理器920选择一DPCM周期,通过该DPCM周期可获得离散度中的一最小位,并将选择的DPCM次数的结果输出到移位器930中。 Subsequently, DPCM processor 920 selects a cycle of DPCM, DPCM period obtained by the dispersion of a minimum bit, and outputs the result to the number of the selected DPCM shifter 930. 选择的DPCM周期、每个DPCM周期的内部关键字数据以及DPCM所要求的其它信息均在步骤S9330中被输出到标题译码器400中。 Selected DPCM cycle, internal key data in each cycle of DPCM and other information required for DPCM are output in step S9330 in the subtitle decoder 400. 然而,在本发明的一优选实施例中,如果关键字的数量小于5,则仅执行DPCM一次。 However, in a preferred embodiment of the present invention, if the number of the keyword is less than 5, DPCM is performed only once. 例如,DCPM的第一周期用等式(10)来执行。 For example, a first cycle DCPM performed by equation (10).

Δi=nQkeyi+1-nQkeyi …(10) Δi = nQkeyi + 1-nQkeyi ... (10)

在这里,i表示量化关键字数据的一指数,nQKeyi表示一整数数组,以及Δi表示差分数据。 Here, i represents an index of quantized key data, nQKeyi represents an array of integers, and Δi represents the differential data.

DPCM处理器920计算为编码DPCM的选择周期的结果以及在步骤S9340中在一预定存储器(nQStep DPCM0中由DPCM所生成的关键字数据的差分数据而所需的位的数量。为进行编码而所需的位的数量的计算也可在后来的选择将被编码的关键字数据的步骤中执行,这对本领域的技术人员是显而易见的。 DPCM processor 920 calculated as the result of the selection period of DPCM encoding and the number of bits in step S9340 in a predetermined memory (nQStep DPCM0 DPCM difference data by the generated key data and the required order to perform the coding calculating the number of bits required for the step can also be encoded key data is performed in the subsequent choice, which those skilled in the art are apparent.

移位器930从DPCM处理器输入的差分数据中选择具有最高频率的一差分数据(在下文中,称为一模式)。 Shifter 930 having the highest frequency of a differential data (hereinafter, referred to as a pattern) from the differential data DPCM processor inputted. 然后,移位器930在步骤S9400中从所有差分数据中减去该模式以便将被编码的大多数数据排列在0周期以及能减小为进行编码而所需的位的数量。 Then, the shifter 930 subtracts the mode from all the differential data in step S9400 so that the most of the data to be encoded are arranged in cycles, and the number of bits can be reduced to 0 by encoding desired.

通过从所有已量化关键字数据中减去一模式nKeyShift,来执行这种移位操作,用下述等式来表示。 A mode by subtracting from all the quantized key data nKeyShift, to perform this shift operation, represented by the following equation.

shift(nQKeyi)=nQkeyi-nKeyShift …(11) shift (nQKeyi) = nQkeyi-nKeyShift ... (11)

在这里,i表示已量化关键字数据的一指数,nQKeyi表示一整数数组,以及nKeyShin表示一模式值。 Here, i represents the quantized data of a keyword index, nQKeyi represents an array of integers, and nKeyShin represents a mode value. 作为该移位操作的结果,具有最高频率的差分数据变为0以便能相当大地减小为进行编码而所需的位的数量。 Difference data as a result of the shift operation, with the highest frequency becomes 0 in order to considerably reduce the number of bits required for coding a.

通过移位操作的关键字数据被输出到折叠处理器940和DND处理器950中,以及模式值nKeyShift被输出到标题编码器400以便被包括在关键字标题中。 Is output by the shift operation of the key data to the folding processor 940 and the DND processor 950, and a mode value nKeyShift is output to the header encoder 400 so as to be included in the key header.

在步骤S9500中,折叠处理器940在移位器930的输出上执行一折叠操作并将折叠操作的结果输出到DND处理器950中。 In step S9500, the folding processor 940 performs a folding operation and folding operation to the DND processor 950 outputs the result in the output 930 of the shifter.

折叠操作被用来通过将它们集中到正或负数区域中来减小在一正数区域和一负数区域上广泛离散的差分数据的范围。 By folding operation is used to concentrate them to a positive or negative number region to reduce the number of zones on a positive and a negative difference data of discrete area wide range. 在本发明实施例中,折叠操作用等式(12)来执行以在正数区域中集中该差分数据。 In the embodiment of the invention, the folding operation is performed in order to concentrate the differential data in the positive number region by equation (12).

fold(nQKeyi)=2nQKeyi (nQKeyi≥0) …(12) fold (nQKeyi) = 2 nQKeyi (nQKeyi≥0) ... (12)

=2|nQKeyi|-1 (nQKeyi<0) = 2 | nQKeyi | -1 (nQKeyi <0)

在这里,i表示量化关键字数据的一指数,nQKeyi表示一整数数组。 Here, i represents an index of quantized key data, nQKeyi represents an array of integers. 作为折叠操作的结果,正差分数据被转换成偶数,以及负差分数据被转换成奇数。 As a result of the folding operation, positive differential data are converted into an even number, and a negative differential data are converted into odd.

折叠处理器940计算为对经过折叠操作的差分数据进行编码而所需的位的数量并将其存储在一预定存储器nQStep fold。 Folding processor 940 calculates the number of bits of differential data after the folding operation by encoding the desired and stored in a predetermined memory nQStep fold. 在该步骤中,编码所需的位的数量的计算可在后面的选择将被熵编码的差分数据的步骤后执行,这对本领域的技术人员来说是显而易见的。 In this step, the calculation of the number of bits required for encoding may be performed after the step of entropy coding of difference data in the back of choice, which those skilled in the art is apparent. 通过在折叠处理器940中的折叠操作生成的数据被输出给DND处理器950。 By folding operation in the folding processor 940 generates data 950 is output to the DND processor.

为提高熵编码的效率,DND处理器950在关键字数据的输入差分数据上执行一DND操作一预定次数,从而在步骤S9600中减小差分数据的范围。 To improve the efficiency of entropy coding, DND processor 950 performs a DND operation a predetermined number of times on the input differential data of key data, thereby reducing the range of the differential data in step S9600 in.

参考图9B,DND处理器950包括一DND运算符952,一第一差分数据选择器954,一上移运算符956,以及一第二差分数据选择器958。 With reference to FIG. 9B, DND processor 950 includes a DND operator 952, a first differential data selector 954, an upper-shift operator 956, and a second differential data selector 958. 该DND运算符952在差分数据上执行一DND操作。 The DND operator 952 performs a DND operation on the differential data. 该第一差分数据选择器954基于用于编码的位的数量来选择将被熵编码的差分数据。 The first differential data selector 954 for encoding the number of bits to be selected based on the entropy-encoded differential data. 该上移运算符956在通过一DND操作的差分数据上执行一上移操作。 The move by the operator in a DND 956 differential data to perform an operation on a shift operation. 该第二差分数据选择器958从具有仅通过DND操作的差分数据和通过上移操作的差分数据上选择具有一较低离散度的一个并将所选择的差分数据输出到熵编码器960。 The second differential data selector 958 to select one from having a low dispersion and having only by DND operation on the differential data and the difference data by moving the selected operation of the differential data is output to the entropy encoder 960.

下面描述将在DND运算符952中执行的DND操作。 The following describes the operation will be performed in the DND DND operator 952.

当在折叠处理器940中通过折叠操作的差分数据被输入到DND运算符952中时,它们被划分成二组,比另一差分数据组具有一较高范围的一组差分数据通过一划分函数被移动到正数区域中。 When through the folding operation in the folding processor 940 are input to the differential data DND operator 952, they are divided into two groups, other than the difference data set having a higher range of a set of difference data by a partition function is moved to the positive number region. 划分函数由下述等式来定义。 The partition function is defined by the following equation.

divide(nQKeyj,nKeyMax) …(13) divide (nQKeyj, nKeyMax) ... (13)

=nQKeyj-(nKeyMax+1) = NQKeyj- (nKeyMax + 1)

=nQKeyj = NQKeyj

在这里,j表示输入差分数据的一指数,nQKeyj表示一整数数组,以及nKeyMax表示通过折叠操作的差分数据间的一最大值。 Here, j represents an index of the input differential data, nQKeyj represents an array of integers, and nKeyMax by folding operation represents a maximum difference between the data. 特别地,在大多数差分数据沿所有差分数据占用的整个区域的边界被密集地填充的情况下,使用该划分操作有可能相当大地减小所有差分数据的整个区域。 In particular, in the case where most differential data along the boundary of all differential data occupies the entire region is densely populated, it is possible to use the dividing operation considerably reduce the entire region of all differential data.

在划分操作后,计算离散度,在这种情况下,编码所需的位的大小被用作离散度的一度量以便选择用于编码的位的大小中的一最小值。 After the division operation, calculate dispersion, in this case, the size of bits required for encoding is used as a measure of dispersion in order to select a minimum size of bits for encoding the.

在DND操作后,不同类型的DND操作,即一上划操作或一下划操作被进一步执行。 After the DND operation, different types of DND operation, that is a designated operation or at the designated operation is further executed. 根据差分数据的一正范围的大小以及差分数据的一负范围的大小,确定将进一步执行一上划操作还是一下划操作。 Depending on the size of a size range of a negative differential positive range of data and differential data to determine the further implementation of the plan operation or at a designated operation.

如果具有正值的差分数据的范围大于具有负值的差分数据的范围,执行由下述等式定义的一下划操作。 If the range has a positive value of the difference is greater than the range of data has a negative value of the difference data, execute the following operations plan defined by the following equation.

divide-dwon(nQKeyj,nKeyMax) …(14) divide-dwon (nQKeyj, nKeyMax) ... (14)

=-2(nKeyMax-nQKeyj+1)+1 = -2 (NKeyMax-nQKeyj + 1) +1

=nQKeyj = NQKeyj

=2nQKeyj (nQKeyj<0) = 2 nQKeyj (nQKeyj <0)

另一方面,如果具有正值的差分数据的范围大于具有负值的差分数据的范围,执行由下述等式定义的一上划操作。 On the other hand, if the range of the differential data having positive values is larger than a negative value having a range of difference data, perform a defined operation designated by the following equation.

divide-up(nQKeyj,nKeyMin) …(15) divide-up (nQKeyj, nKeyMin) ... (15)

=nQKeyj (nQKeyj≥0) = NQKeyj (nQKeyj≥0)

=2nQKeyj = 2 nQKeyj

=2(nKeyMin-nQKeyj-1)+1 = 2 (nKeyMin-nQKeyj-1) +1

在等式(14)和(15)中,j表示量化关键字数据的一指数,nQKeyj表示一整数数组,nKeyMax表示nQKeyj的一最大值,以及nKeyMin表示nQKeyj的一最小值。 In the equation (14) and (15), j indicates an index of quantized key data, nQKeyj represents an array of integers, nKeyMax indicates a maximum value nQKeyj, and nKeyMin indicates a minimum value nQKeyj of.

下面将参考图10E来描述DND运算符952的操作。 Will now be described with reference to FIG. 10E to operate the DND operator 952.

当从折叠处理器940输入关键字数据的差分数据时,DND运算符952在步骤S9610中获得输入差分数据间的最大值nKeyMax和最小值nKeyMin。 When the 940 differential data input key data from the folding processor, DND operator 952 to get the maximum and minimum nKeyMin nKeyMax differential data input in step S9610 in between. 然后,在步骤S9620中,DND运算符将nKeyMax的绝对值与nKeyMin的绝对值比较。 Then, in the step S9620, DND operator is compared with the absolute value of the absolute value of nKeyMax nKeyMin of. 如果nKeyMax不小于nKeyMin的绝对值,DND运算符952在步骤S9622中将nKeyMax设置为在DND操作的当前周期中的一最大值。 If nKeyMax not less than the absolute value nKeyMin, DND operator 952 is set to the current cycle of DND operation at a maximum in the step S9622 nKeyMax.

DND运算符952核对DND操作的次数是否为1,换句话说,如果DND的次数为1,在步骤S9624中,如果它是,则DND运算符952在步骤S9630中通过将最大值nKeyMax代入等式(13)中来在输入差分数据上执行一划分操作。 DND operator 952 to check whether the number of DND operation is 1, in other words, if the number of DND is 1, in step S9624, and if it is, the DND operator 952 by the maximum value nKeyMax in step S9630 substituting into equation (13) performs a division operation on the input differential data.

此后,在步骤S9640中,使用函数getQBit(),DND运算符952使用划分操作来测量用于编码已经被减小的差分数据范围的位的大小。 Thereafter, in step S9640 using the function getQBit (), DND operator 952 used to measure the size of the division operation has been reduced for encoding the differential data range of bits. 在步骤S9650中,如果DND操作的次数变为1,编码所需的位的大小被存储为表示用于编码的最小位的大小的一值nQBitDND,以及在步骤S9655中将DND操作的次数加1。 In step S9650, the DND operation if the number becomes 1, the size of bits required for encoding is stored as a representation for encoding a bit size of the minimum value nQBitDND, and increase the number of DND operation in step S9655 will be in a .

接着,DND处理器952再次执行步骤S9610至S9622。 Then, DND processor 952 steps S9610 to S9622 again. 如果在步骤S9624中DND操作的次数不为1,则在步骤S9634中,DND运算符952通过将最大值nKeyMax代入等式(14)来执行一下划操作。 If the number of times in the DND operation is not 1 in step S9624, then in step S9634 in, DND operator 952 by substituting the maximum value nKeyMax in equation (14) to execute the following operation plan. 在步骤S9640中,DND运算符952计算用于编码通过下划操作的差分数据的位的数量。 In the step S9640, DND operator 952 calculates the number of bits encoded by the underlined operation of the differential data. 如果该数量小于在先前DND操作周期中存储的最小值nQBitDND,在步骤S9658中,在DND操作后,它替换要求用于编码的最小的位的大小。 If the number is smaller than the minimum value nQBitDND DND operation in the previous cycle is stored in step S9658, after the DND operation, it replaces the minimum size requirements for encoding bits.

如果在步骤S9620中,最小值nKeyMin的绝对值似乎大于最大值nKeyMax的绝对值,在步骤S9623中,在当前DND操作的周期中的最大值被更新为一最小值,然后,在步骤S9628中,通过将最小值nKeyMin代入等式(15)中来执行一上划操作。 If in step S9620, the absolute value of the minimum value nKeyMin appears to be greater than the absolute value of the maximum value nKeyMax in step S9623, the maximum value in the current cycle of DND operation is updated to a minimum value, then, in step S9628, the by the minimum value nKeyMin into equation (15) to execute an operation on a draw. 此后,DND运算符952在步骤S9640中计算用于编码通过上划操作的差分数据的位的数量。 Thereafter, DND operator 952 in step S9640 by the number of bits of the encoded difference data on the designated operation for calculation. 在步骤S9652中,如果计算结果变为小于在先前DND操作周期中存储的nQBitDND,在步骤S9658中,在DND操作后,它替换要求用于编码的位的最小数nQBitDND。 In step S9652, if the calculation result becomes smaller than in the previous cycle of DND operation stored nQBitDND, in step S9658, after the DND operation, it replaces the requirements for the minimum number of bits encoded nQBitDND.

DND处理器952执行预定次数的DND操作,以及DND操作的性能的数量可改变。 DND processor 952 performs a predetermined number of times of DND operation, and the number of performance of the DND operation may vary. 例如,在本发明实施例中,执行7次DND操作。 For example, in the embodiment of the invention, the DND operation performed 7 times. DND运算符952将nQBitDND和与nQBitDND相对应的差分数据输出到第一差分数据选择器954中。 DND operator 952 outputs nQBitDND and the differential data corresponding to nQBitDND to the first differential data selector 954. DND运算符952将通过其生成相应差分数据的DND次数输出到标题编码器400中并允许它们包括在位流中。 DND operator 952 generates differential data by the corresponding number of outputs to which DND header encoder 400 and allows them comprising the bit stream.

第一差分数据选择器954接收经过移位操作的差分数据、经过折叠操作的差分数据以及经过DND操作的差分数据并在它们三个中确定将被熵编码的差分数据。 The first differential data selector 954 receives the differential data after the shift operation, after the folding operation and the differential data after the DND operation and determines which differential data to be entropy-encoded in the three differential data.

参考图10A,在步骤S9710中,如果在DND操作后为进行编码而所需的位的最小数nQBitDND不小于在步骤S9700DPCM操作后用于编码的大小nQBitDND,则第一差分数据选择器954选择DPCM的结果并在它们上执行一移位操作。 With reference to FIG. 10A, in step S9710, and if the operation is performed in the DND coding bits required is not less than the minimum number nQBitDND size nQBitDND after step S9700DPCM operation for encoding, then the first differential data selector 954 to select DPCM results and performs a shift operation on them. 接着,第一差分数据选择器954将位移操作的结果输出给熵编码器960并允许在步骤S9710中编码它们。 Next, the results of the first differential data selector 954 outputs the shift operation to the entropy encoder 960 and allows them encoded in step S9710. 在这种情况下,DND操作的次数被设置为-1,被输出到标题编码器400并被包括在关键字标题中。 In this case, the number of DND operation is set to -1, is output to the header encoder 400 and is included in the key header.

然而,如果在步骤S9720中出现nQBitDND小于nQStep-DPCM并不小于在折叠操作后用于编码的位的大小,第一差分数据选择器954将经过折叠操作的差分数据输出到熵编码器960中并在步骤S9730中允许熵编码它们,在这种情况下,DND操作的次数被设置为0,输出到标题编码器400,从而包括在关键字标题中。 However, if in step S9720 is smaller than nQBitDND appears nQStep-DPCM is not less than the size used for encoding after the folding operation of the bit, the first differential data selector 954 will go through the folding operation of the differential data is output to the entropy encoder 960 and allows entropy coding in step S9730 in which, in this case, the number of DND operation is set to 0, the output to the header encoder 400, so that the keyword is included in the title.

如果在DND操作后用于编码差分数据的位的数量是最小的,第一差分数据选择器954将经过DND操作的差分数据输出到上移运算符956中,然后在步骤S9740中上移运算符956计算从第一差分数据选择器954输入的差分数据的第一离散度。 If the number of bits after the DND operation to the encoded difference data is the smallest, the first differential data selector 954 will be operated via the differential data output to the DND operator 956 moves, and then move the operator in step S9740 calculating a first dispersion 956 from the first differential data selector 954 inputs the difference data. 接着,上移运算符956在步骤S9800在经过DND操作的差分数据上执行用下述等式定义的上移操作并在步骤S9810中计算上移操作的结果的第二离散度。 Next, the shift operator 956 a second dispersion step S9800 shift operations on the DND operation on the differential data performed by the following equation, and defined on the calculation result in the shifting operation of the step S9810.

shift-up(nQKeyj,nkeyMax) …(16) shift-up (nQKeyj, nkeyMax) ... (16)

=nQKeyj (nQKeyj≥0) = NQKeyj (nQKeyj≥0)

=nKeyMax-nQKeyj (nQKeyj<0) = NKeyMax-nQKeyj (nQKeyj <0)

在这里,j表示量化关键值数据的差分数据的一指数,nQKeyj表示一整数数组,以及nKeyMax表示差分数据间的一最大值。 Here, j represents the quantized key value data of the difference data of an index, nQKeyj represents an array of integers, and nKeyMax indicates a maximum value among differential data.

当输入经过DND操作的差分数据以及经过上移操作的差分数据时,在步骤S9900中,第二差分数据选择器958将第一离散度与第二离散度进行比较。 When the input data through the DND operation and the differential through the differential data operations on the move, in step S9900, the second differential data selector 958 will be the first dispersion and the second dispersion compared. 如果第二离散度小于第一离散度,第二离散数据选择器958将经过上移操作的差分数据输出到熵编码器960并在步骤S9910中允许熵编码它们。 If the second dispersion degree is smaller than the first dispersion degree, the second differential data selector 958 discrete data will go through the shifting operation are output to the entropy encoder 960 and allows them to entropy coding in step S9910. 第二差分数据选择器958输出用在DND操作中的最大和最小值nKeyMax和nKeyMin。 The second differential data selector 958 output DND operation in the maximum and minimum values nKeyMax and nKeyMin. 在本发明的一优选实施例中,标准偏差可被用作第一和第二离散度的一度量。 In a preferred embodiment of the present invention, the standard deviation can be used as a measure of the first and second dispersion.

熵编码器960根据差分数据的特点在差分数据上执行两个不同的函数。 Two different functions entropy encoder 960 is executed on difference data according to the characteristics of the differential data. 例如,经过一DPCM操作和一移位操作的差分数据以及仅经过一划分操作的差分数据具有正和负值,因此它要求执行编码每个差分数据的特征位和差分数据本身的一处理。 For example, after a DPCM operation and a shifting operation and differential data after only one division operation of the differential data having positive and negative values, so it requires the implementation of differential encoding of each data bit and the characteristic difference of a processing data itself. 另一方面,由于经过一折叠操作的差分数据仅具有正值,执行仅编码差分数据的一处理。 On the other hand, since differential data via a folding operation only have positive values, a processing execution only encoded difference data.

在本发明的一优选实施例中,函数encodeSignedAAC被用于编码差分数据以及它们的特征位,以及函数encodeUnsignedAAC被用于仅编码差分数据。 In a preferred embodiment of the present invention, the function is used to encode differential data encodeSignedAAC and their flag, and is used only function encodeUnsignedAAC encoded difference data.

图11是函数encodeSignedAAC()的例子的图。 FIG 11 is a function encodeSignedAAC () example of FIG. 参考图11,当一输入值是74以及用于编码该输入值的位的数量为8时,其特征位为0,以及它与二进制数1001010相同。 Referring to Figure 11, when an input value is 74, and for encoding bits of the input value of the number is 8, characterized bit is 0, and it is the same with the binary number 1001010. 特征位以及所有位平面用下述方法编码。 Flag and all bit-plane coded by the following method.

第一步:在每个位平面上按从其最高有效位(MSB)到其最低有效位的顺序编码一二进制数; Step: each bit-plane in the press from which the most significant bit (MSB) to its least significant bit of a binary number encoding sequence;

第二步:核对当前正被编码的位是否为0; Step Two: Check bit currently being encoded is 0;

第三步:如果当前正被编码的位不是0,接着编码该二进制数的特征位;以及 The third step: if the bit currently being encoded is not 0, then the encoding feature bit binary numbers; and

第四步:编码该二进制数据的其余位; Step Four: The remaining bits encode the binary data;

使用与这些值有关的环境,函数encodeUnsignedAAC将不具有一特征位的这些值编码成一自适应算术编码位流。 These values are used with related environmental, function encodeUnsignedAAC will not have a characteristic bit values encoded into an adaptive arithmetic encoding bitstream. 该函数除存在一特征位上下文(sign context)外,几乎与函数encodeSignedAAC()相同。 This function is in addition to the presence of a flag context (sign context), almost the function encodeSignedAAC) the same as (.

图12A至12J是根据本发明的一优选实施例,表示经过操作的关键字数据的图。 12A to 12J in accordance with a preferred embodiment of the present invention, showing the operation after the keyword data in FIG. 在图12A至12J,X轴表示每个关键字数据的指数,以及Y轴表示关键字数据的值。 In FIG. 12A to 12J, X-axis represents the index, and the Y-axis represents the value of each keyword data key data.

图12A是表示输入到本发明的编码器中的原始关键字数据的图。 12A is input to the encoder of the present invention, the original key data in Fig. 如图12A所示的关键字数据被输出到量化器910中,然后与9个量化位一起被量化以便获得如图12B所示的量化关键字数据。 12A, the key data is output to the quantizer 910, and then quantized with 9 bits are quantized to obtain quantized key data together as shown in FIG. 12B. 如果在如图12B所示的量化关键字数据上执行DPCM,获得如图12C所示的差分数据。 If DPCM performed on the quantized key data shown in Figure 12B, and 12C obtain the differential data shown in FIG.

接着,使用约为7的一模式值来移位该量化关键字数据的差分数据以便获得如图12D所示的差分数据。 Next, using a mode value of about 7 to shift the difference data of the quantized key data to obtain difference data 12D shown in Fig. 此后,如果在移位差分数据上执行一折叠操作,获得如图12E所示的仅为正值的数据。 After that, if you perform a folding operation on the shift differential data, obtained only positive data shown in Fig. 12E.

在如图12E所示的折叠数据上执行一DND操作的结果如图12F至12H所示。 Result of performing a DND operations on the data shown in Figure 12E folded to 12H shown in Figure 12F. 具体来说,在折叠数据上执行一划分操作的结果如图12F所示。 Specifically, performing a division operation on the folded data shown in the results shown in Figure 12F. 如图12F所示,正关键字数据值范围从0至28,负关键字数据值范围从-28至0,这表示负关键字数据值的范围大于正关键字值数据的范围。 FIG. 12F, positive key data values range from 0 to 28, the negative key data values range from -28 to 0, which means that the range of negative key data values is greater than the key value of the data range. 因此,要求在如图12F所示的数据上执行一上划操作,以及上划操作的结果如图12G所示。 Therefore, the requirements on the implementation of a data plan shown in Figure 12F operation, and the results of the draw operation is shown in Figure 12G.

作为上划分操作的结果,相当大地减小了负关键字数据值的范围以便它远小于正关键字数据值的范围。 As a result of the division operation, considerably reduces the range of negative key data values so that it is much smaller than the range of positive key data values. 在DND操作的下一周期,在上划操作的结果上执行一下划操作。 DND operation in the next cycle, the implementation of some actions on the results of the draw stroke operation. 图12H是表示在如图12G所示的差分数据上执行一下划操作的结果。 Figure 12H is a diagram showing the execution result about the classified operations on the differential data shown in FIG. 12G. 在如图12H所示的关键字数据上执行一上移操作的结果如图12I所示。 Performing a shift operation on the key data shown in FIG. 12H 12I results are shown in Fig.

如图12A至12G所示,逐渐减小关键字数据和差分数据的范围。 Shown in FIG. 12A to 12G, the range of key data is gradually decreased and the difference data. 然而,如图12H和12I所示,在上移操作后增加的差分数据的范围大于在上移操作前,这表示经过划分操作的差分数据,如图12H所示,是将被最终编码的差分数据,如图12J所示。 However, as shown in FIG. 12H and 12I, the range after the increase in the shifting operation of the differential data is larger than in the former shifting operation, which means that the differential data after dividing operation, as shown in Figure 12H, is to be a final coded difference data, as shown in Figure 12J.

在标题编码器400中编码以及在关键字标题中存储的信息如下所述。 In the header encoder 400 and the encoding information stored in the key header described below.

当输入将被编码的关键字数据时,标题编码器400编码将编码的关键字数据的数字数和关键字数。 When the input key data to be encoded, header encoder 400 encodes the digital data and the number of key data of the encoded key. 接着,标题编码器400从线性关键字编码器900接收有关是否存在经过线性关键字对输入关键字数据进行编码的一线性关键字数据区域以及在该线性关键字数据区域中的关键字数据的数量的信息并从浮点数转换器905接收经过浮点数转换的线性关键字数据的开始和结尾关键字数据。 Then, heading encoder 400 receives the number 900 after about whether to enter a keyword keyword linear encoding data area exists a linear key data and key data in the linear region of the key data from the linear encoder keyword and 905 to receive information through the beginning and end float keyword data conversion from floating-point linear key data converters.

在浮点数转换器接收能带来一最小量化误差的最小和最大值并将它们转换成浮点数的情况下,转换的最大和最小值被从浮点数转换器905输入到标题编码器400以便它们能被再次用于逆量化。 In the floating-point number converter receives bring a minimum quantization error of the minimum and maximum values and converts them into floating-point numbers, the converted maximum and minimum values are input from the floating-point number converter 905 to the header encoder 400 so that they can be re-used for inverse quantization. 另外,量化位的大小也能被输入到标题编码器400中并被包括在关键字标题中。 In addition, the size of quantization bits can also be input to the header encoder 400 and is included in the keyword in the title.

关键字标题编码器400从DPCM处理器920接收DPCM的次数以及在每个DPCM循环中的内部关键字数据并从移位器930接收已经用于一移位操作的一模式值。 Key header encoder 400 receives the number of DPCM DPCM processor 920 and internal key data in each cycle of DPCM from and received from the shifter 930 has a mode value for a shift operation. 另外,标题编码器400从DND处理器950接收有关是否已经执行一上移操作以及DND次数的信息,通过该操作,能减小差分数据的离散度,以及在每个DND操作周期中的最大和最小值。 Further, header encoder 400 receives from the DND processor 950 as to whether the shifting operation has been performed a number of times and DND information, through this operation, the difference data can be reduced dispersion, and in each cycle of DND operation and the maximum Min.

最后,标题编码器从熵编码器960接收用于编码的位的数量并将其编码为一关键字标题。 Finally, the header encoder 960 receives from the entropy encoder for encoding the number of bits and encoded as a title keyword.

在下文中,参考图13A至20B,根据本发明的第一实施例来更全面地描述一关键字值数据编码器300的结构和操作。 Hereinafter, with reference to FIG. 13A to 20B, according to a first embodiment of the present invention will be more fully described with a key value data encoder 300. The structure and operation.

图13A是根据本发明第一实施例的一关键字值数据编码器300的框图。 Figure 13A is a block diagram 300 of a key value data encoder according to a first embodiment of the present invention. 参考图13A,关键字值数据编码器300包括一旋转DPCM运算符1300,一循环DPCM运算符,一熵编码器1450,以及一标题编码器。 With reference to FIG. 13A, key value data encoder 300 includes a rotational DPCM operator 1300, a loop DPCM operator, an entropy coder 1450, and a header encoder. 旋转DPCM运算符1300计算在连续关键帧中一对象的旋转变换值间的旋转差分值,将关键帧的四元数关键字值数据应用到它们各自的对象上,量化该旋转差分值,并输出旋转差分值。 Rotation DPCM operator 1300 calculation continuous rotation keyframes rotational transform differential value between the value of an object, the application quaternion key value data keyframes to their respective objects, to quantify the rotation differential value and outputs rotation differential value. 该循环DPCM运算符在量化旋转差分数据上有选择地执行一线性DPCM或一循环DPCM操作。 The loop DPCM operator to selectively perform a cyclic or linear DPCM DPCM operation on the quantized rotational differential data. 该熵编码器1450对旋转DPCM或循环DPCM旋转差分数据进行熵旋转。 The entropy coder 1450 pairs of rotating DPCM or DPCM loop rotational differential data entropy rotation. 该标题编码器编码为译码已编码的动画内插器节点的关键字值数据所需的信息。 The header encoder encodes information required for decoding key value data encoded animation interpolator node.

旋转DPCM编码器1300包括一第一四元数乘法器1310,一量化器1340,一逆量化器1350,一第二四元数乘法器1370,以及一延迟器1390。 DPCM rotary encoder 1300 includes a first quaternion multiplier 1310, a quantizer 1340, an inverse quantization unit 1350, a second quaternion multipliers 1370, 1390, and a delay. 该第一四元数乘法器1310计算在一前关键帧的一对象的一旋转变换值与在一当前关键帧中该对象的一旋转变换值间的一旋转差分值,通过将在当前关键帧中的该对象的旋转变换值四元数乘以在前关键帧中的该对象的旋转转换值。 The first quaternion multiplier 1310 calculates a difference value between the rotation value of a rotational transform a rotational transformation value of a key frame in front of an object with a current keyframe between the object, by the current keyframe rotary conversion value rotational transform values of the object multiplied by the former quaternion key frame of the object. 该量化器1340通过非线性量化从第一四元数乘法器1310输入的旋转差分数据来生成量化差分数据。 Rotational differential data by the quantization unit 1340 from the first non-linear quantization quaternion multiplier 1310 generates the input quantized differential data. 该逆量化器1350通过逆量化已量化旋转差分数据来生成在当前关键帧中的该对象的复原旋转差分数据。 Restoring rotational differential data by the inverse quantizer 1350 inverse-quantizing the quantized rotational differential data to generate in the current key frame of the object. 该第二四元数乘法器1370通过将在当前关键帧中的该对象的旋转差分值四元数乘以通过累加旋转差分数据所计算的在前关键帧中该对象的旋转变换值来复原在当前关键帧中的该对象的旋转变换值。 The second quaternion multiplier 1370 by the rotational difference in the current keyframe value of the object multiplied by the quaternion rotation transformation previous keyframe value by accumulating data calculated differential rotation of the object to the recovery in The current rotation transformation keyframe value of the object. 该延迟器1390当输入在下一关键帧中的该对象的一旋转变换值时,将在当前关键帧中的该对象的复原旋转变换值输出到第一四元数乘法器。 The delay 1390 when a rotational transform values are entered in the next key frame of the object will be restored rotational transformation value of the current key frame of the object to the first quaternion output multipliers.

在下文中,将更详细地描述根据本发明的在一旋转DPCM运算符中执行的一旋转DPCM操作。 Hereinafter, described in more detail in accordance with the present invention, a rotary DPCM performed on a rotary DPCM operator in operation.

在已经在常规MPEG-4PMFC中采用的一线性DPCM方法中,用下面的等式(17)来计算表示在一当前关键帧中一对象的旋转变换(或表示在一当前关键帧中一对象被旋转变换的程度)的一四元数旋转变换值Q1(Q1=(q1,0,q1,1,q1,2,q1,3))与表示在下一关键帧中该对象的旋转变换的一四元数旋转变换值Q2(Q2=(q2,0,q2,1,q2,2,q2,3))间的一差分值。 In a linear DPCM method has been employed in the conventional MPEG-4PMFC in, the following equation (17) represents the rotation transform to calculate a current keyframe one object (or shown in a current keyframe an object is degree rotation transformation) of a quaternion rotation transformation value Q1 (Q1 = (q1,0, q1,1, q1,2, q1,3)) and represents the next key frame in the object's rotation transformation of a four quaternion rotation transformation value Q2 (Q2 = (q2,0, q2,1, q2,2, q2,3)) a difference value between.

QlinearDPCM=(q1,0-q2,0,q1,1-q2,1,q1,2-q2,2,q1,3-q2,3) …(17) QlinearDPCM = (q1,0-q2,0, q1,1-q2,1, q1,2-q2,2, q1,3-q2,3) ... (17)

然而,线性DPCM方法仅计算四元数分量间的一差分值,这不能反映任何有意义的旋转差分值,即一实际的旋转差分值。 However, the linear DPCM method only calculates a differential value between quaternion components, which does not reflect any meaningful rotational differential value, i.e., an actual rotational differential value. 因此,线性DPCM方法不能有效地减少将被编码的连续关键字值数据间的冗余。 Accordingly, the linear DPCM method can not effectively reduce the redundancy to be encoded among successive key value data. 另外,在线性DPCM方法中,除具有最大值的一个分量之外,编码一四元数的三个分量。 In addition, the linear DPCM method, in addition to a maximum of one component outside the coding of the three components of a quaternion. 因此,有必要另外编码在未被编码的该一个分量上的2比特的长信息,并将该2比特的长信息从一编码器传送到一译码器。 Therefore, it is necessary to additionally encode a component on the length information is not encoded by 2 bits, and the 2-bit long information is transmitted from an encoder to a decoder.

为利用通过减小在连续关键字值数据间的冗余的传统线性DPCM方法来解决上述问题,根据本发明的第一实施例,关键字值数据编码器提供一旋转的DPCM操作,其中考虑到实际的旋转差分值。 For use to solve the above problems by reducing redundancy among successive key value data of the conventional linear DPCM method, according to a first embodiment of the present invention, key value data encoder to provide a rotating DPCM operation, taking into account actual rotational differential value.

When

表示一基准矢量,该基准矢量表示在一当前关键帧中的一对象的一基准位置, Represents a reference vector, the reference vector represents a reference position in a current keyframe of an object,

表示当关键字值等于ki-1时的关键字值数据,以及 Indicates that when the key value is equal to ki-1 key value data, and

通过旋转变换在当前关键帧中的对象获得的 By rotating the object in the current keyframe transform obtained

的一位移矢量,在一四元数空间中的旋转变换能用下面的等式表示。 A displacement vector, rotation transformation in a quaternion space can use the following equation.

在等式(18)中,X0,Yi-1,Qi-1以及 In equation (18), X0, Yi-1, Qi-1, and

分别表示 Respectively

以及 As well as

的单位四元数(unit quarternion)。 The unit quaternion (unit quarternion). 另外, In addition,

表示Qi-1的一四元数复共扼,以及表示四元数乘法。 Qi-1 represents a quaternion complex conjugate, and represents quaternion multiplication.

用这种方式,当关键字数据等于ki时,在一四元数空间中的旋转变换能用下面的等式来表示。 In this way, when the key data is equal to ki, the rotation transformation in a quaternion space can be used to represent the following equation.

可用以下等式(20)来计算连续关键字值数据的连续旋转变换值间的一旋转差分值。 Using the following equation (20) calculates a rotation differential value continuously for converting continuous rotation value between the key value data.

因此,能通过由等式(5)和(6)导出的下面的等式来定义一旋转差分值。 Therefore, by the following equation from Equation (5) and (6) to define a rotary derived differential values.

为防止一量化误差扩展到下一旋转差分值,根据本发明的一实施例,关键字值数据编码器300利用在下一关键帧中的一旋转变换值以及前关键帧中的一复原旋转变换值 In order to prevent a quantization error is extended to the next rotational differential value, in accordance with an embodiment of the present invention, key value data encoder 300 utilize a rotating conversion value in the next keyframe and a restored rotational transformation value before keyframe

来重新定义由等式(21)定义的该旋转差分值,如等式(22)中所示。 Redefining shown by equation (21) defined by the rotation difference values, as shown in equation (22).

在下文中,将参考图13B来描述一种用于根据本发明的第一实施例编码关键字值数据的方法。 Hereinafter, with reference to FIG. 13B to describe a method of encoding key value data of a first example embodiment according to the present invention. 图13B是根据本发明的第一实施例的用于编码关键字值数据的方法。 13B is a method for encoding key value data of a first embodiment of the present invention.

根据本发明的第一实施例,关键字值数据编码器在步骤S1300中接收DPCM次数、一熵编码模式以及关键字值数据。 According to a first embodiment of the present invention, key value data encoder DPCM in step S1300 the received frequency, an entropy encoding mode and key value data.

接着,在步骤S13050中,第一四元数乘法器1310接收用一四元数表示的关键字值数据Qi并核对输入的关键字值数据Qi是否是第一关键字值数据Q0。 Next, in step S13050, the first quaternion multiplier 1310 receives key value data Qi with a quaternion key value and check whether the input data Qi is the first key value data Q0. 如果输入的关键字值数据Qi是第一关键字值数据Q0,表示没有用于四元数乘法的累加的四元数变换值。 If the key value data input Qi is the first key value data Q0, indicating no for quaternion multiplication of the accumulated value of the quaternion transformation. 因此,第一四元数乘法器1310将输入关键字值数据Qi输出到量化器1340中。 Thus, the first quaternion multiplier 1310 will enter the keyword value data output to the quantizer Qi 1340. 如果输入关键字值数据Qi不是第一关键字值数据Qo,在步骤S13100中,通过将前关键帧中的复原四元数变换值 If the input key value data Qi is not the first key value data Qo, in step S13100 through the keyframe before restoration quaternion transformation value

的复共扼)四元数乘以由一四元数变换值表示的输入关键字值数据Qi The complex conjugate) quaternion key value multiplied by the input data by a quaternion transformation value represents the Qi

第一四元数乘法器1310计算一四元数差分值 1310 The first quaternion multiplier to calculate the differential value Quaternion

其是当前关键帧的关键字值数据与前关键帧的复原关键字值数据间的一差分值。 Which is a difference value between the current value of the data recovery Keywords Keywords keyframes previous keyframe values between data.

在步骤13300中,量化器1340从第一四元数乘法器1310接收第一关键字值数据Q0或四元数旋转差分值 In step 13300, the quantizer 1340 1310 to receive a first key value data from the first quaternion multiplier Q0 or quaternion rotation difference value

并使用一预定数量的量化位来量化该输入。 And using a predetermined number of quantization bits for quantizing the input.

由于所有旋转四元数差分值均由单位四元数表示,表示一旋转差分值的一四元数的一范数总是为1。 Since all the rotation quaternion difference value by the unit quaternion represents a norm of a difference value between the rotation number of a quaternion is always 1. 因此,在未被编码的一四元数的四个分量中的一分量能使用其它三个分量来译码。 Thus, four components have not been encoded in a quaternion in one component can be decoded using the other three components. 根据本发明的第一实施例,关键字值数据编码器300仅编码一四元数的四个分量中的三个以便将被编码的数据量。 According to a first embodiment of the present invention, key value data encoder 300 encodes the data amount only four components of a quaternion in the three to be encoded. 因此,量化器1340仅量化将被编码的三个分量。 Accordingly, the quantizer 1340 quantizes only the three components to be encoded.

根据本发明,量化器1340执行非线性量化而不是一般的线性量化,以及下面将参考图14A描述该原因。 According to the present invention, quantizer 1340 performing non-linear quantization rather than general linear quantization, and is described below with reference to FIG. 14A that reason.

图14A是描述在一旋转差分值的每个分量中的一概率分布函数(PDF)的一个典型的例子的图。 14A is a description of each component of the probability of a rotational differential value of the distribution function (PDF) of a typical example of FIG. 如图14A所示,旋转差分值的分量值一般集中在0左右,这被称为能源压缩并表示有可能有效地降低旋转信息片中的冗余。 14A, component values of rotation differential values generally concentrate around 0, which is called energy compressed and expressed is possible to efficiently reduce the rotation information sheet redundancy. 因此,为量化旋转差分值,更充分地反映每个旋转差分值的较低的分量值,需要更完善地在较低分量值上执行量化,这就是为什幺根据本发明的量化器1340执行非线性量化的原因。 Accordingly, quantized rotational differential value, more adequately reflect the difference value of each of the rotating lower component values need to better perform quantization on the lower component values, that's why the present invention is performed according to the quantizer 1340 non-linear quantization reasons. 此时,量化器1340使用一反正切曲线以便将一非线性比例因子分配给每个旋转差分值。 In this case, the quantizer 1340 uses a arctangent curve in order to assign a non-linear scale factor to each of the rotational differential value.

图14B根据本发明给出用于量化的一反正切曲线。 According to the present invention is given in FIG. 14B for quantization an arctangent curve. 如图14B所示,该反正切曲线向较高输入值提供一适当的分辨率并向较低输入值提供一更高的分辨率。 14B, the arc tangent curve provides an appropriate resolution to lower input values to provide a higher input values higher resolution. 量化器1340使用如下所示的等式的一非线性比例函数来量化在步骤S13300中的一旋转差分值。 Quantizer 1340 using the equation shown below in a non-linear function to quantify the proportion of a rotational differential value in step S13300 in.

此时, At this time,

表示输入到该量化器1340中的旋转差分值的每个分量,qi′表示 Each component represents the input to the quantizer 1340 in the rotational difference value, qi 'represents

的一比例值,nQBits表示用于量化的预定数量的量化位,以及floor(x)表示用于将一输入值x变换成不大于x的一最大整数的一函数。 A proportional value, nQBits denotes a predetermined number of quantization bits of quantization, and floor (x) indicates for an input value of x is not greater than x is converted into a maximum integer of a function.

量化器1340在输入旋转差分值上执行量化,然后核对量化的旋转差分数据是否与将被编码的最后一个关键字值数据相对应。 On whether the quantizer 1340 performs quantization on the input rotational differential value, and then check the quantized rotational differential data to be encoded with the last key value data, respectively. 如果量化后的旋转差分数据与最后一个关键字值数据相对应,该量化器1340将量化后的旋转差分数据输出到循环DPCM运算符1400中。 If the quantized rotational differential data with the last key value data corresponding to the rotation of the differential quantizer 1340 outputs the quantized data to the loop DPCM operator 1400. 如果量化后的旋转差分数据不与最后的关键字值数据相对应,量化器1340将量化后的旋转差分数据输出到逆量化器1350中。 If the quantized rotational differential data do not correspond to the last key value data, the quantizer 1340 quantized rotational differential data output to the inverse quantizer 1350.

在步骤S13400中,反量化器1350相反地量化从量化器1340输入的量化后的旋转差分数据并在步骤S1350中输出一复原旋转差分数据。 In step S13400, the inverse quantizer 1350 contrary quantized rotational differential data from the quantization unit 1340 inputs and outputs a restored rotational differential data in step S1350.

如上所述,根据本发明的第一实施例,关键字值数据编码器通过量化除第一分量外的三个分量,仅编码表示一旋转差分值的一四元数的三个分量。 As described above, according to the first embodiment of the present invention, key value data encoder other than the first component by quantizing the three components outside, only the coded representation of the three components of the rotational differential value of a number of a quaternion. 一关键字数据译码器的一逆量化器2430,如图21A所示,以及在上面描述过的逆量化器1350被假定使用三个编码过的分量来复原在旋转差分值的四个分量中未被编码的其它的一个。 A key data decoder 2430 an inverse quantizer, as shown, as well as in the above described inverse quantizer 21A 1350 is assumed to use the three encoded components to recover the four components of the rotational differential value Other an uncoded. 由于所有旋转差分值每个均由一单位四元数表示,表示一旋转差分值的一四元数的范数始终为1。 Since all the differential value of each rotation by a unit quaternion representing a rotation differential value norm of a quaternion is always number one. 因此,使用下述等式来复原旋转差分值的四个分量中的第一分量是可能的。 Therefore, using the following equation to recover the four components of the rotational differential value in the first component is possible.

在等式(24)中, In equation (24),

以及 As well as

表示一旋转差分值的三个复原分量,以及 Represents a rotational differential value of the three restored component, and

表示使用三个复原分量 That the use of three rehabilitation component

以及 As well as

而复原的第一分量。 The restoration of the first component.

为利用等式(11)来复原第一分量 To use equation (11) to recover the first component

第一分量 The first component

必须具有一正值,并且该条件可通过利用当将一四元数应用到在一13D空间中的一对象的旋转变换中时所出现的一四元数的特征来满足,并如等式(25)所示。 You must have a positive value, and the conditions when a quaternion to the rotation transformation through the use of a 13D space when an object appearing in a number of features to meet the quad, and as shown in equation ( 25) FIG.

Y=QXQ*=(-Q)X(-Q)* …(25) Y = Q X Q * = (- Q) X (-Q) * ... (25)

等式(25)表示当将旋转变换应用到在一3D空间的一对象时旋转变换值Q和-Q具有相同的物理含义。 Equation (25) indicates that when the rotational transformation applied to rotational transformation value Q when a 3D space of an object and -Q have the same physical meaning. 因此,如果根据本发明的第一实施例,在关键字值数据编码器300中一旋转差分值的第一分量能通过将该旋转差分值的每个分量乘以-1容易被转换成一正值。 Therefore, if according to a first embodiment of the present invention, the key value data encoder 300, a difference between the first rotation component by the value of each component of the rotational differential value is multiplied by -1 to be easily converted into a positive value . 然而,在这种情况下,除第一分量外一复原旋转差分值的每个分量的平方和由于一量化误差可能超过1。 However, in this case, in addition to the first outer component of each component of a restored rotational squared difference value and a quantization error may be due to more than one. 在这种情况下,不能由等式(24)确定 In this case, can not be determined by the equation (24)

并被认为是接近0以及小于能被量化器1340量化的一最小值的一值。 Was considered to be close to 0 and less than the quantizer 1340 can be a quantized value of a minimum value.

具有这样一值的事实表示该对象已经通过几乎等于约180度被旋转变换。 The fact of having such a value indicates that the object has passed almost equal to about 180 degrees is a rotation transformation. 因此,一个用于译码一定向内插器的装置需要用于确定 Therefore, a certain inward interpolator for decoding means for determining the need for

的一种方法同时减小复原第一分量值 A method for restoring a first component value while reducing

在三个复原分量值 In three rehabilitation component values

以及 As well as

上的影响,并且该方法将被应用到用于根据本发明的第一实施例的关键字值数据编码器300的逆量化器1350中。 Impact on, and the method will be applied according to the key value data encoder of the first embodiment of the present invention, an inverse quantizer 300 of 1350. 例如,能由量化器1340量化的最小值乘以一预定整数可被确定为 For example, the minimum value can be multiplied by 1340 a quantizer for quantizing a predetermined integer may be determined as

下面的等式可被用来确定 The following equation can be used to determine

在等式(26)中,'a'表示一任意常数,以及m表示一预定数量的量化位。 In equation (26), 'a' represents an arbitrary constant, and m represents a predetermined number of quantization bits.

当在常规MPEG-4BIFS PMFC方法中要求有关每个关键字值数据的2比特长的信息用于译码时,该2比特长信息在本发明不是必须的,因此当编码N个关键字值数据时,将被编码的位的数量能被减少差不多2N。 When the request data about each key value in the conventional MPEG-4BIFS PMFC method, 2-bit long information is used during decoding, the 2-bit long information is not essential to the present invention, so when the encoding key value data of N , the number of bits to be encoded can be reduced almost 2N.

从该逆量化器350输出的复原旋转差分值被输入到第二四元数乘法器1370中,以及在步骤S13600中该第二四元数乘法器1370核对一输入关键帧是否是将被编码的一第一关键帧。 Recovery from the inverse quantizer 350 outputs the rotation differential value is input to the second quaternion multiplier 1370, and the second quaternion multiplier 1370 in step S13600 in check whether an input key frame to be encoded a first keyframe. 如果该输入关键帧是将被编码的第一关键帧时,第二四元数乘法器1370累加从该逆量化器输入的第一关键帧的复原旋转变换值并在步骤S13800中将累加值输出到该延迟器1390中。 If the input keyframe is the first keyframe to be encoded, the restored rotational transformation value of the first key frame of the second quaternion multiplier 1370 accumulates the input from the inverse quantizer and the accumulated value will be output at step S13800 to the delay in 1390.

如果该输入关键帧不是将被编码的第一关键帧时,在步骤S13700中,通过将当前关键帧的一旋转差分值 If the input keyframe is not the first keyframe to be encoded, in step S13700 by a rotational difference between the current key frame value

与一前关键帧的一复原旋转变换值 And a recovery value of a pre-rotation transformation keyframe

四元数相乘第二四元数乘法器1370复原一当前关键帧的一旋转变换值 Quaternion multiplying the second quaternion multiplier 1370 restored a rotational transformation value of a current keyframe

第二四元数乘法器1370将当前关键帧的复原旋转变换值 The second quaternion multiplier 1370 will transform the current rotation keyframe value recovery

输出给延迟器1370,然后延迟器1390保持该 1370 output to delay, delay and then maintain the 1390

直到下一关键帧的一旋转变换值被输出以便 Until a rotational transformation value of the next key frame is output to

等于 Equal

然后将 Then

输出到第一四元数乘法器1310中。 Output to the first quaternion multiplier 1310.

当一当前关键帧的一旋转变换值Qi被输入到第一四元数乘法器1310中时,第一四元数乘法器1310也从延迟器1390接收一前关键帧的复原旋转变换值 When a current of a rotational transformation value Qi keyframe is input into the first quaternion multiplier 1310, the first quaternion multiplier 1310 also receives 1390 a key frame from the delay before the return rotational transformation value

并在步骤S13100中通过将前关键帧中的复原四元数变换值 And by restoring the former keyframe quaternion transformation in value in step S13100

四元数乘以Qi Quaternion multiplied Qi

来生成当前关键帧的旋转变换值与前关键帧的复原旋转变换值间的一旋转差分值Qi。 A rotational difference rotation transformation value of the current recovery to generate value keyframe rotation transformation between the previous keyframe score Qi.

生成的旋转差分值被输出到量化器1340中,然后在步骤S13300中,按如下所述,由量化器1340量化。 Generating rotational differential value is output to the quantizer 1340, then in step S13300, as described below, the quantization by the quantizer 1340. 在步骤S13400中,如果所输入的旋转差分值与最后一个关键字值数据相对应,量化器1340将通过量化一旋转差分值所生成的旋转差分数据输出到循环DPCM运算符400中。 In step S13400, the rotational differential value if the input to the last key value data corresponding to the quantizer 1340 by the quantized rotational differential value by a rotational difference generated data is output to the loop DPCM operator 400.

循环DPCM运算符1400核对从量化器1340输入的量化的旋转差分数据的DPCM的次数是否为0。 Loop DPCM operator 1400 to check the number of times DPCM quantizer 1340 quantizes a rotational differential data is 0. 在步骤S14000中,如果DPCM的次数为0,循环DPCM运算符1400将量化的旋转差分数据输出到熵编码器1450中而不执行一线性DPCM操作和一循环DPCM操作。 In step S14000, if the number of DPCM is 0, the loop DPCM operator 1400 outputs the quantized rotational differential data to the entropy coder 1450 without performing a linear DPCM operation and a cycle of DPCM operation. 如果DPCM的次数不为0,在步骤S14100中,循环DPCM运算符1400在该量化旋转差分数据上执行一线性DPCM操作以及一循环DPCM操作。 If the number of DPCM is not 0, in step S14100, the loop DPCM operator 1400 perform a linear DPCM operation and a cycle of DPCM operation on the quantized rotational differential data.

图15A是描述从该量化器1340输出的差分数据的例子的图,以及图15B是描述在从量化器1340输出的差分数据上执行一线性DPCM操作的结果的图。 15A is an example of description from the quantizer 1340 outputs the difference data of the map, and Fig. 15B is the result described in the implementation of a linear DPCM operation on the differential data from the output of the quantizer 1340 in FIG.

如图15B所示,作为一线性DPCM操作的结果,将被编码的差分数据的范围可比它过去增加二倍。 15B, as a result of a linear DPCM operation, the range of differential data to be encoded in the past than it increased twice. 循环DPCM操作的性能的目的是将差分数据的范围保持在量化的差分数据的范围内。 The purpose of the performance of the operation cycle is the range of DPCM difference data is maintained in the range of quantized differential data.

循环DPCM操作是在假定在一量化范围内的一最大值和一最小值被循环地彼此连接之下执行的。 DPCM loop operation is assumed in a quantization range of a maximum and a minimum value under each other are connected cyclically executed. 因此,如果在两个连续量化数据上执行线性DPCM的结果的差分数据大于在量化范围内的最大值的一半时,通过从差分数据减去最大值,可用较小值来表示它们。 Therefore, if the continuous quantitative results of performing linear DPCM on two difference data in the data is greater than half the maximum value in the quantization range from the differential data by subtracting the maximum value, the smaller value can be used to represent them.

如果差分数据小于在量化范围内中的最小值的一半时,通过将在量化范围内的最大值加到该差分数据上,可用较小值甚至更大的来表示它们。 If the difference data is smaller than the minimum value in the quantization range to half, by the maximum value in the quantization range is applied to the differential data, the smaller value can be used to represent them even greater.

When

And

分别表示在两个连续时间瞬间ti和ti-1的量化旋转差分数据,在下面的等式(27)的两个连续量化旋转并非分数据 Respectively in two successive time instants ti and quantized rotational differential data ti-1, the two in the following equation (27) for continuous rotation of quantized data is not sub-

And

上执行一线性DPCM操作。 Performed on a linear DPCM operation.

另外,在下面的等式(27)、(28)获得的差分数据上执行一循环DPCM操作。 Further, in the following equation (27), performing a DPCM operation on the differential data loop (28) obtained.

(如果Xi≥0)(相反) (If Xi≥0) (opposite)

在等式(28)中,nQBits表示一预定数量的量化位。 In equation (28), nQBits represents a predetermined number of quantization bits. 图15C是描述在如图15B所示的DPCM差分数据上执行一循环DPCM操作的结果的图。 15C is a description of the implementation of a DPCM loop DPCM operation on the differential data shown in Figure 15B results in FIG. 如图15C所示,循环DPCM差分数据的范围越小于线性DPCM差分数据的范围。 15C, the range of cyclic DPCM difference data is smaller than the linear DPCM difference data.

熵编码器1450根据旋转差分数据的DPCM的次数从循环DPCM运算符1400接收旋转差分数据或循环DPCM旋转差分数据,然后通过移除位冗余来编码输入差分数据。 The entropy encoder 1450 receives the rotational differential data or the rotational differential data from the DPCM loop loop DPCM operator 1400 based on the number of rotational differential data of the DPCM, and then by removing the redundant bits to encode the input differential data.

再参考图13B,熵编码器1450在步骤S14500中核对一熵编码模式。 Referring again to FIG. 13B, an entropy encoder 1450 entropy coding mode checked in step S14500 in.

如果熵编码模式是一二进制熵编码模式,则在步骤S14600中熵编码器1450使用一函数SignedAAC()来编码该输入差分数据。 If the entropy encoding mode is a binary entropy encoding mode in step S14600 in the entropy encoder 1450 uses a function SignedAAC () to encode the input differential data. 如果熵编码模式是一一元熵编码模式,则在步骤S14700中使用一函数UnaryAAC()来编码该输入差分数据。 If the entropy encoding mode is one element entropy coding mode, using a function UnaryAAC in step S14700 in () to encode the input differential data.

函数SignedAAC()被用来利用一自适应的二进制算术编码器来编码差分数据,函数SignedAAC()在每个位平面上编码该差分数据的标记和大小,并且编码方法已经参考图11在上面描述过。 Function SignedAAC () is used to exploit an adaptive binary arithmetic coder to encode differential data, functions SignedAAC () mark and the size of the differential encoding data on each bit plane, and the encoding method has been described above with reference to FIG. 11 too.

另一方面,函数UnaryAAC()被用来通过将被编码的一值转换成一系列0、表示该0系列结束的一特征位位1以及表示该值的标记的另一位来编码符号,0系列的数量与该值的大小相对应。 On the other hand, the function UnaryAAC () is used by a value to be encoded into a series of 0, a flag bit indicates the end of a series of 0 and represents another symbol mark of the encoded value, 0 series The number and size of the value, respectively.

在下文中,函数UnaryAAC()将参考图16B在下面更全面地描述。 Hereinafter, the function UnaryAAC () with reference to FIG. 16B described more fully below. 例如,2156被编码成一系列位,这些位由256个Os,表示Os系列结束的一特征位位1,以及表示256的标记的0,即函数UnaryAAC()的一加号标记组成。 For example, 2156 is encoded as a series of bits, these bits by 256 Os, a flag bit indicates the end of a series of Os and represents 256 labeled 0, i.e., the function UnaryAAC () is composed of a plus sign. 由于UnaryAAC(),表示将被编码的符号的位间的冗余增加,这提高了编码该符号的有效性。 Since UnaryAAC (), indicates increasing redundancy symbols to be encoded between the bits, which improves the effectiveness of the encoding of the symbol.

在下文中,将描述根据本发明的第二实施例的一关键字值数据编码器300。 Hereinafter, description will be based on a key value data encoder to a second embodiment of the present invention 300. 根据本发明的第二实施例的关键字值数据编码器300包括一装置,用于校正在量化期间产生的一旋转方向误差以及用于根据本发明的第一实施例编码的关键字值数据编码器300的所有组件。 Key value data encoder according to a second embodiment of the present invention 300 comprises a means for correcting a rotation direction error generated during quantization and means for encoding key value data encoder according to a first embodiment of the present invention All components 300.

图17是描述在使用旋转差分值编码四元数旋转变换值期间生成的一旋转方向误差的图。 FIG 17 is a description of the direction of rotation during the rotation difference value encoded using quaternion rotational transformation value of the generated error FIG. 由于在根据本发明的关键字值数据编码器300中执行的四元数编码方法是有损编码方法(loss encoding method),从而导致旋转方向误差。 Since the lossy encoding method (loss encoding method), leading to the rotation direction error from the quaternion encoding method key value data encoder 300 of the present invention is performed.

在图17中,当Qi,Qi-1, In Fig. 17, when Qi, Qi-1,

And

分别表示当前正被应用到在一第I关键帧中的一对象的一旋转变换值、应用到在一前关键帧中的该对象的一旋转变换值、从QI复原一值以及从Qi-1复原一值时,通过几乎等于分别由Qi,Qi-1, Respectively, is currently being applied to a rotation transformation values in the first one I keyframe object rotation transformation is applied to a key value in a front frame of the object, as well as recover from a value from the QI Qi-1 restoring a value, through almost equal respectively, by Qi, Qi-1,

And

表示的旋转变换旋转的该对象的位置可由四个不同区域表示,如图17所示。 The position of the object represented by the rotational transformation rotates by four different regions that shown in Figure 17.

换句话说,如果该对象沿一最短弧线从Qi-1旋转到Qi因此在由Qi和Qi-1旋转后,该对象位于区1和区3中,该对象被认为按逆时钟方向从Qi-1旋转到Qi。 In other words, if the object along a shortest arc of rotation from Qi-1 to Qi Qi and therefore by Qi-1 after the rotation of the object located in the zone 3 and zone 1, the object is considered to press the counter-clockwise from Qi -1 rotation to Qi. 此时,如果该对象沿一最短路线从Qi-1旋转到Qi因此在按Qi-1和Qi旋转后,该对象位于区2和区4中,该对象被认为按顺时针方向从Qi-1旋转到Qi。 At this time, if the object along a shortest path from Qi-1 to the rotation so by Qi and Qi Qi-1 after the rotation, the object is located in area 2 and area 4, the object is considered clockwise from Qi-1 rotate to Qi.

另一方面,如果已经被编码,然后被译码的旋转信息旋转该对象,用于译码一定向内插器的一种装置利用分别与Qi,Qi-1相对应的 On the other hand, if it is already encoded, and then rotate the object rotation information decoding, for decoding a certain inward interpolator utilizing an apparatus, respectively, Qi, Qi-1 corresponding to the

And

来旋转该对象。 To rotate the object. 因此,如果相对于由 Therefore, if with respect to the

所旋转的该对象的位置而言,由 Position of the object in terms of rotation, by

所旋转的对象的相对位置是区2或3,该对象逆时针旋转。 The relative position of the object is rotated by the zone 2 or 3, the object rotates counterclockwise. 如果相对于由 If with respect to the

所旋转的该对象的位置而言,由 Position of the object in terms of rotation, by

所旋转的对象的相对位置是区1或4,该对象顺时针旋转。 The relative position of the object is rotated by the area 1 or 4, the object rotates clockwise. 在区1和2中,当利用一原始旋转变换值时,该对象的旋转方向可能与当利用一译码的旋转变换值的该对象的旋转方向相反,因为Qi和 In zones 1 and 2, when the use of an original rotational transformation value, the rotation direction of the object may be related to the direction of rotation when utilizing a decoded rotational transformation value of the object contrast, because Qi and

与被执行以编码四元数旋转变换值的损失编码是不同的。 And is performed to encode quaternion rotational transformation values of lossless coding is different. 为解决该问题,需要减小该对象在与所需要的方向相反的一错误方向中所旋转的程度或当该对象在一错误方向中旋转时校正该对象的旋转方向以便该对象在一需要的方向中旋转。 To solve this problem, the need to reduce the extent of the object in a direction opposite to a desired direction in the error correction of the rotation direction or rotation of the object when the object rotates in a wrong direction so that the object in a desired direction of rotation. 在本发明中,采用用于校正该对象的旋转方向的方法以便使该对象在一需要的方向中旋转。 In the present invention, a method for correcting the direction of rotation of the object so that the object rotates in a direction in need thereof.

再参考图17,在下文中将简单地描述根据本发明的校正一旋转方向误差的原理。 Referring again to Figure 17, will hereinafter be briefly described in accordance with the present invention, the correction of a rotation direction error in principle. 如果检测到一旋转方向误差,如在区1和2中产生的现象,将被编码的四元数旋转差分值被控制以便在一正确方向中旋转,在这种情况下在区2中仍然会发生在旋转方向中的不一致性。 If a rotation direction error is detected, resulting in phenomena such as region 1 and 2, will be encoded quaternion rotation difference value is controlled to rotate in a right direction, in this case in the zone 2 will still occurs in the direction of rotation of the inconsistency. 然而,在区2中,不象在区1中,原始四元数值以及复原四元数值间的差值相比较小。 However, in the area 2, unlike in area 1, the original quaternion value and the restored quaternion value of the difference between small compared. 因此,根据本发明第二实施例的旋转方向校正仅在区1上执行。 Therefore, the correction is performed only on the area a according to the rotation direction of the second embodiment of the present invention.

在下文中,将参考图18A至19B来描述根据本发明第二实施例的执行旋转方向校正的关键字值数据编码器300。 Hereinafter, with reference to FIGS. 18A to 19B will be described the correction of the direction of rotation according to the implementation of the second embodiment of the present invention, key value data encoder 300. 根据本发明第二实施例的关键字值数据编码器300与根据本发明第一实施例的关键字值数据编码器300的结构几乎相同。 Key value data encoder according to a second embodiment of the present invention 300 and the key value data encoder according to a first embodiment of the invention the structure 300 is almost the same. 它们间的唯一区别是一旋转DPCM运算符的结构,因此在下面将仅描述根据本发明第二实施例的关键字值数据编码器300中的一旋转DPCM运算符的结构。 The only difference between them is the rotational DPCM operator of a structure, only the structure will be described below in a second embodiment of the present invention, key value data encoder 300 in one rotational DPCM operator according to.

图18A根据本发明第二实施例给出了旋转DPCM运算符1800的框图,并且图18B是如图18A所示的一旋转方向误差计算器1820的一框图。 Figure 18A shows a block diagram of the rotational DPCM operator 1800 according to a second embodiment of the present invention, and FIG. 18B is a rotation direction error calculator shown in FIG. 18A is a block 1820.

参考图18A,该旋转DPCM运算符1800包括一旋转方向误差计算器1820,一旋转方向误差检测器1830,一旋转方向校正器1815,以及一旋转方向选择器1835。 With reference to FIG. 18A, the rotational DPCM operator 1800 includes a rotation direction error calculator 1820, a rotation direction error detector 1830, 1815, and a rotational direction of a rotation direction corrector selector 1835. 旋转方向误差计算器1820接收在一当前关键帧中的一对象的一旋转变换值以及在一前关键帧中该对象的一复原旋转变换值并计算旋转方向误差。 Rotation direction error calculator 1820 receives a current value of a restored rotational transformation value of a rotation transformation keyframe and an object in one keyframe before the object and calculate the direction of rotation errors. 该旋转方向误差检测器1830基于从旋转方向误差计算器1820输入的旋转方向误差检测在译码过程中是否足以改变该对象的旋转方向的一误差已经发生。 The rotation direction error detector 1830 from the rotational direction based on the rotation direction error calculator 1820 inputs the error detection decoding process is sufficient to change the direction of rotation of an error of the object has occurred. 该旋转方向校正器1815校正和输出从第一四元数乘法器1810输入的一旋转差分值以便通过采用前关键帧的一译码旋转变换值的一变换来定位的该对象能进一步在一原始旋转方向中旋转约180度。 The direction of rotation corrector 1815 correction and output from the first rotation quaternion differential value input to the multiplier 1810 through a transformation using a decoding key frame before rotation transformation values to locate the object further in an original rotational direction rotated approximately 180 degrees. 该旋转方向选择器1835根据从旋转方向误差检测器1830输入的一值来选择从旋转方向校正器1815输入的旋转差分值或从第一四元数乘法器1810输入的旋转差分值并将所选择的值输出到该量化器1840中。 The direction of rotation selector 1835 from a value based on the direction of rotation error detector 1830 to select the input from the rotation direction of the rotational difference corrector 1815 score or the selected input from the first quaternion multiplier 1810 and entered the rotation differential value The value of the output to the quantizer 1840.

参考图18B,如图18A所示的旋转方向误差计算器1820包括一延迟器1822,一第三四元数乘法器1824,一第四四元数乘法器1826,以及一第五四元数乘法器1828。 Referring to FIG. 18B, as shown in the rotation direction error calculator 1820 shown in 18A includes a delay 1822, a third quaternion multiplier 1824, a fourth quaternion multiplier 1826, and a fifth quaternion multiplication 1828. 该延迟器1822存储一输入旋转变换值直到下一关键帧的一旋转变换值被输入为止。 The delay device 1822 storing an input rotational transformation value until a rotational transformation value of a next keyframe is input so far. 该第三四元数乘法器1824接收该输入旋转变换值以及从该延迟器1822输出的一前关键帧的一旋转变换值并通过四元数乘法计算该输入旋转变换值以及前关键帧的旋转变换值间的一旋转差分值。 The third quaternion multiplier 1824 receives the input rotation transformation value and the value of the rotation of the input rotation transformation and former key frame from a rotational transformation value of a keyframe before 1822 the output of the delay and is calculated by quaternion multiplication a rotational difference between the value converted value. 该第四四元数乘法器1826计算从延迟器1822输出的前关键帧的旋转变换值和该前关键帧的复原旋转变换值间的一旋转差分值。 The 1826 fourth quaternion multiplier to calculate a difference value between the rotation value from the rotation transformation rotation transformation value of the previous key frame delay unit 1822 output and the restoration of the former between keyframes. 该第五四元数乘法器1828在输入旋转变换值与前关键帧的复原旋转变换值间计算。 The fifth quaternion multiplier 1828 in restoring the input rotational transformation value of the previous keyframe between the rotation transform value calculations.

在下文中,参考图19A来更详细地描述根据本发明的第二实施例的一旋转DPCM操作。 Hereinafter, is described with reference to FIG rotational DPCM operation according to a second embodiment of the present invention in more detail to 19A.

在步骤S19000中,打算被编码的一定向内插器节点的关键字值数据Qi被输入到第一四元数乘法器1810以及该旋转DPCM运算符1800的旋转方向误差计算器1820中。 In step S19000, the intended key value data are encoded in a certain inward interpolator node Qi are inputted into the first quaternion multiplier 1810 and the rotation direction of the rotational DPCM operator 1800 in the error calculator 1820.

第一四元数乘法器1810,象根据本发明第一实施例的关键字值数据编码器300中的第一四元数乘法器一样,通过将一当前关键帧的一输入旋转变换值四元数乘以从第二四元数乘法器1870输入的一前关键帧的一复原旋转变换值 The first quaternion multiplier 1810, like the key value data encoder according to a first embodiment of the invention 300 in the first quaternion multiplier as an input by a rotational transformation value of the current keyframe quaternion A restored rotational transformation value of a pre-multiplied by the number of key frames from the 1870 second quaternion multiplier input

来生成一旋转差分值 To generate a rotation differential value

并在步骤S19100中将所生成的旋转差分值 S19100 and rotational difference in the scores generated in step

输出到该旋转方向校正器1815和旋转方向选择器1835中。 Output to the correct direction of rotation of 1815 and 1835 in the direction of rotation selector.

旋转方向校正器1815按等式(29)校正输入到其中的一旋转差分值并将校正的旋转差分值Qs输出到校正方向选择器1835中。 Rotation direction corrector 1815 according to the equation (29) into which a corrective input rotational differential value and the corrected rotational differential value Qs to the direction of correction of the output selector 1835. 旋转方向误差计算器1820接收当前关键帧的旋转变换值Qi以及从第二四元数乘法器1870中输入的前关键帧的复原旋转变换值 Restored rotational transformation value of the rotation direction error calculator 1820 receives the current keyframe and the rotational transformation value Qi before keyframe input from the second quaternion multiplier 1870 in

并计算旋转变换值QA,QB以及QC,这将在以后描述。 And calculates rotational transformation values QA, QB and QC, which will be described later. 旋转方向误差检测器1830利用从旋转方向误差计算器1820所输入的旋转差分值来检测是否已经发生一旋转方向误差并在步骤S19200中将检测结果输出到旋转方向选择器1835中。 Rotation direction error detector 1830 from the rotational differential value utilizing the rotation direction error calculator 1820, the input to detect whether an error has occurred and a rotational direction in the detection result in step S19200 is output to the rotation direction selector 1835.

在等式(29)中,δr表示非常接近0的一常数,并且(qR,0,qR,1,qR,2,qR,3)T表示从第一四元数乘法器1810输出的旋转差分值 In equation (29), δr represents a constant very close to 0, and (qR, 0, qR, 1, qR, 2, qR, 3) T represents the rotational differential from the first quaternion multiplier 1810 outputs value

旋转方向误差计算器1820、旋转方向误差检测器1830以及旋转方向校正器1815将参考图19B描述如下。 Rotation direction error calculator 1820, the rotation direction error detector 1830 and the rotation direction corrector 1815 will be described below with reference to FIG 19B.

旋转方向选择器1835核对是否已经发生一旋转方向误差以便当使用从旋转方向误差检测器1830输入的一逻辑值来译码一编码关键字值数据时该对象在相反方向中旋转到一需要的方向。 Rotation direction selector 1835 to check whether a rotation direction error has occurred so that when using the logic values from one rotational direction error detector 1830 for decoding a coded input key value data of the object in the opposite direction of rotation to a desired direction . 如果一旋转方向误差没有发生,在步骤S19300中,旋转方向选择器1835将从第一四元数乘法器1810输入的旋转差分值输出到量化器1840中。 If an error does not occur the direction of rotation, in step S19300, the rotation direction selector 1835 from the first quaternion rotation difference between the input of the multiplier 1810 to quantify the output value in 1840. 如果一旋转方向误差已经发生,旋转方向选择器1835在步骤S19300中将从旋转方向校正器1815输入的校正旋转并分值输出。 If an error has occurred direction of rotation direction of rotation selector in step S19300 in 1835 from the rotation direction of the rotation correction corrector 1815 input and output value.

量化器1840在步骤S19400和S19500中用与本发明的第一实施例中相同的量化方法量化原始旋转差分数据 A quantizer quantizing the original rotational differential data 1840 of the first embodiment of the present invention the same quantization method used in steps S19400 and S19500

或校正的旋转差分数据Qs。 Or corrected differential rotation data Qs.

量化器1840在步骤S19600中核对量化的旋转差分数据 Quantizer 1840 collation quantized rotational differential data in step S19600 in

是否属于最后一个关键字值数据。 Whether it is the last key value data. 如果量化的旋转差分数据 If the differential rotation of quantitative data

与最后一个关键字值数据相对应,量化器1840在步骤S19700中将量化的旋转差分数据 And the last key value data corresponding to the quantizer 1840 quantization step S19700 in the rotation differential data

输出到一循环DPCM运算符1400中。 Outputs to a loop DPCM operator 1400. 如果量化的旋转差分数据 If the differential rotation of quantitative data

不与最后一个关键字值数据相对应,量化器1840将量化的旋转差分数据 Not with the last key value data corresponding to the quantizer 1840 will be quantized rotational differential data

输出到逆量化器1850中。 Output to the inverse quantizer 1850.

逆量化器1850在步骤S19800中用与本发明的第一实施例相同的逆量化方法逆量化该量化的旋转差分数据 An inverse quantizer 1850 in step S19800 using the first embodiment of the present invention the same inverse quantization method quantizes the quantized inverse differential rotation data

并将复原旋转差分数据 Differential rotation and data recovery

输出到第二四元数乘法器1870。 Output to the second quaternion multiplier 1870.

第二四元数乘法器1870通过将复原旋转差分数据 The second quaternion multiplier 1870 by rotating the differential data recovery

四元数乘以一前关键帧的一旋转变换值 Quaternion multiplied by a keyframe before a rotation transformation value

生成一当前关键帧的一复原旋转变换值 Generating a current key frame of a restored rotational transformation value

第二四元数乘法器1870在步骤S19900中经延迟器1890将一当前关键帧的一旋转变换值输出到第一四元数乘法器1810和旋转方向误差计算器1820中。 The second quaternion multiplier 1870 via the delay circuit 1890 to a rotational transformation value of a current keyframe in step S19900 is output to the first quaternion multiplier 1810 and the rotation direction error calculator 1820.

在下文中,将参考图19B来描述旋转方向误差计算器1820、旋转方向误差检测器1830以及旋转方向校正器1815的操作。 Hereinafter, the operation will be described with reference to FIG. 19B rotation direction error calculator 1820, the rotation direction error detector 1830 and the rotation direction corrector 1815.

旋转方向误差计算器1820计算与区2至4相对应的旋转误差值,这已经参考图17描述过。 Calculating rotation direction error calculator 1820 and zone 2-4 corresponds to the rotation error value, which has been described with reference to FIG. 17.

旋转方向误差计算器1820接收与将被编码的一关键帧的关键字值数据相对应的一旋转变换值,接收在当前关键帧前的一前关键帧的一复原旋转变换值并在步骤S19220中计算一旋转方向误差。 Rotation direction error calculator 1820 receives key value data to be encoded in a key frame corresponding to a rotation transformation value, receives a restored rotational transformation value of a key frame before the current key frame at the front and at the step S19220 calculating a rotation direction error.

与当前关键帧的一定向内插器的关键字值数据相对应的旋转变换值被直接输入到第三四元数乘法器1824和第五四元数乘法器1828中并且当一下一关键帧的一旋转变换值被输入时经延迟器1822被输入到第四四元数乘法器1826中。 Inward current must insert the key frame key value data corresponding to the rotation transformation values are entered directly into the third quaternion multiplier 1824 and the fifth quaternion multiplier in 1828 and when at a keyframe After a rotation transformation value is input delay 1822 is input to the fourth quaternion multiplier 1826. 另外,从第二四元数乘法器1870输出的复原旋转变换值被输入到第四四元数乘法器1826和第五四元数乘法器1828中。 Further, from the rotational transformation value of the second restoration quaternion multiplier 1870 output is input to the fourth quaternion multiplier 1826 and the fifth quaternion multiplier 1828.

第三四元数乘法器1824通过将当前关键帧的旋转变换值四元数乘以前关键帧的旋转变换值生成旋转差分值QA并将QA输出到旋转方向误差检测器1830中。 The third quaternion multiplier 1824 by the rotational transformation value of the current keyframe quaternion multiplication previous keyframe value generation rotary rotation transformation QA and QA output difference value to the rotation direction error detector 1830. 旋转差分值QA表示在一时间间隔[ti-1,ti]间一对象的一原始旋转方向。 Rotational differential value QA represents an interval [ti-1, ti] between an original rotation direction of an object.

第四四元数乘法器1826通过将前关键帧的旋转变换值四元数乘以前关键帧的复原旋转变换值生成旋转差分值QB并将QB输出到旋转方向误差检测器1830中。 The fourth quaternion multiplier 1826 by the rotational transformation value of the previous keyframe quaternion multiplication rotation transformation keyframe values recover before generating QB and QB rotation differential value output to the rotation direction error detector 1830. 通过第四四元数乘法器1826计算的旋转差分值QB表示由一量化误差在一预定时间瞬间ti-1引起的一对象的旋转方向误差和旋转方向并与如图17所示的区2相对应。 Rotational difference through the fourth quaternion multiplier 1826 represents the calculated value and QB as shown in the direction of rotation and the rotation direction error by a quantization error in a predetermined time instant ti-1 induced an object 17 with the two-phase region correspondence.

第五四元数乘法器1828通过将当前关键帧的旋转变换值四元数乘以前关键帧的复原旋转变换值生成旋转差分值QC并将QC输出到旋转方向误差检测器1830。 The fifth quaternion multiplier 1828 by the rotational transformation value of the current keyframe quaternion rotation transformation value multiplied recover before generating rotation keyframes QC and QC output differential value to the rotation direction error detector 1830. 由第五四元数乘法器1828计算的旋转差分值QC表示在预定时间瞬间ti被编码的一旋转差分值且与如图17所示的区4相对应。 Rotational difference by the fifth quaternion multiplier 1828 represents the calculated value QC region and shown in Figure 17 in a predetermined time instant ti be a coded rotational differential value and 4, respectively.

旋转方向误差检测器1830在步骤S19240中使用从旋转方向误差计算器1820输入的旋转差分值QA、QB和QC来核对在译码编码的关键字值数据的过程中该对象是否在与一原始方向相反的方向中旋转。 Rotation direction error detector 1830 using the rotation direction error calculator 1820 from a rotary input differential value QA, QB and QC to check whether the object is in the original direction with a key value in the process of decoding the encoded data in step S19240 in in the opposite direction of rotation. 为完成该功能,旋转方向误差检测器1830核对旋转差分值QA、QB和QC是否满足不等式(30)至(34)。 To accomplish this function, the rotation direction error detector 1830 to check the rotational differential values QA, QB and QC satisfies inequality (30) to (34). 首先,旋转方向误差检测器1830核对旋转差分值QA和QC是否满足不等式(30)。 First, the rotation direction of the rotation error detector 1830 to check whether the differential value QA and QC satisfy the inequality (30).

在不等式(30)中,当从旋转方向误差计算器1820输入的旋转差分值QA由QA=(qA,0,qA,1,qA,2,qA,3)T时, In inequality (30), when viewed from the direction of rotation of the rotary input error calculator 1820 by the differential value QA QA = (qA, 0, qA, 1, qA, 2, qA, 3) T, the

表示QA的四个分量qA,0,qA,1,qA,2,qA,3除第一分量qA,0所组成的一三维矢理(qA,1,qA,2,qA,3)T。 QA represent the four components qA, 0, qA, 1, qA, 2, qA, 3 except the first component qA, a three-dimensional vector consisting of Li 0 (qA, 1, qA, 2, qA, 3) T.

表示由QC的四个分量qC,0,qC,1,qC,2,qC,3除第一分量qC,0外组成的一三维矢量(qC,1,qC,2,qC,3)T。 QC is represented by four components qC, 0, qC, 1, qC, 2, qC, 3 in addition to the first component qC, 0 outside the composition of a three-dimensional vector (qC, 1, qC, 2, qC, 3) T. 不等式(30)表示两个3D矢量(qA,1,qA,2,qA,3)T和(qC,1,qC,2,qC,3)T的一内积小于0的一条件。 Inequality (30) shows two 3D vectors (qA, 1, qA, 2, qA, 3) T and (qC, 1, qC, 2, qC, 3) an inner product of less than one condition T 0.

如果不等式(30)中的内积小于0,一对象的旋转方向当使用QA时与当使用QC时的旋转方向相反。 If the inequality (30) in the product is less than 0, the rotation direction of an object when using QA and QC when using the opposite direction of rotation. 当内积小于0时,在不等式(30)中所示的一逻辑表达式的一值被设置为“true”。 When the inner product is less than 0, the inequality (30) a value of a logical expression shown is set to "true". 当内积大于0时,该逻辑表达式的值被设置为“false”。 When the inner product is greater than 0, the value of the logical expression is set to "false".

在不等式(31)中,当从旋转方向误差计算器1820输入的旋转差分值QB由QB=(qB,0,qB,1,qB,2,qB,3)T时, Inequality (31), when QB when the QB = (qB, 0, qB, 1, qB, 2, qB, 3) T from the rotational differential value input rotation direction error calculator 1820, and

表示QB的四个分量qB,0,qB,1,qB,2,qB,3除第一分量qB,0组成的一三维矢理(qB,1,qB,2,qB,3)T。 Said the four components of QB qB, 0, qB, 1, qB, 2, qB, 3 except the first component qB, a three-dimensional vector processor 0's (qB, 1, qB, 2, qB, 3) T.

表示由QC的四个分量qC,0,qC,1,qC,2,qC,3除第一分量qC,0外组成的一三维矢量(qC,1,qC,2,qC,3)T。 QC is represented by four components qC, 0, qC, 1, qC, 2, qC, 3 in addition to the first component qC, 0 outside the composition of a three-dimensional vector (qC, 1, qC, 2, qC, 3) T. 不等式(31)表示两个3D矢量(qB,1,qB,2,qB,3)T和(qC,1,qC,2,qC,3)T的一内积小于0的一条件。 Inequality (31) shows two 3D vectors (qB, 1, qB, 2, qB, 3) T and (qC, 1, qC, 2, qC, 3) an inner product of less than one condition T 0.

如果不等式(31)中的内积小于0,一对象的旋转方向当使用QB时与当使用QC时的旋转方向相反。 If the inequality (31) in the product is less than 0, the rotation direction of an object when using QB and QC when using the opposite direction of rotation. 当内积小于0时,在不等式(31)中所示的一逻辑表示达的一值被设置为“true”。 When the inner product is less than zero, a logic in inequality (31) represents a value shown up is set to "true". 当内积大于0时,该逻辑表达式的值被设置为“false”。 When the inner product is greater than 0, the value of the logical expression is set to "false".

ATH<2cos-1|qA,0| ...(32) ATH <2cos-1 | qA, 0 | ... (32)

在不等式(32)中,qA,0表示旋转差分值QA的第一分量,以及ATH被设置为接近0的一预定常数。 Inequality (32), qA, 0 represents the rotational differential value QA of the first component, and ATH is set to a predetermined constant close to zero. 当在不等式(32)中,2cos-1|qA,0|大于ATH时,在不等式(32)中所示的一逻辑表达式的一值被设定为“true”。 When the inequality (32), 2cos-1 | qA, 0 | is greater than ATH, a value of a logical expression of the inequality (32) is set to be shown in "true". 当2cos-1|qA,0|不大于ATH时,逻辑表达式的值被设定为“false”。 When 2cos-1 | qA, 0 | is not greater than the ATH, logical expression value is set to "false". 当不等式(32)被设置为“false”时,这表示当根据本发明实现关键字值数据编码器300时小于预定常数ATH的一旋转角θ(=2cos-1|qA,0|)能被忽略。 When inequality (32) is set to "false", this indicates that when implementing key value data encoder 300 according to the present invention is less than a predetermined constant ATH rotation angle θ (= 2cos-1 | qA, 0 |) can be ignored. 即使存在一旋转方向误差几乎等于旋转角θ,旋转方向误差不能导致对人眼来说严重失真的图像,特别是当使用由本发明建议的一误差测量方法。 Even if there is an error of rotational direction is almost equal to the rotation angle θ, the rotation direction error does not cause a serious distortion of the human eye image, especially when using an error measurement method proposed by the present invention.

ATH<2cos-1|qB,0| ...(33) ATH <2cos-1 | qB, 0 | ... (33)

在不等式(33)中,qB,0表示旋转差分值QB的第一分量,并且ATH是与不等式(32)中相应的一个相同。 In inequality (33), qB, 0 represents the rotation differential value of the first component QB, and ATH is (32) in a respective one of the same inequality. 当在不等式(33)中,2cos-1|qB,0|大于ATH时,在不等式(33)中所示的一逻辑表达式的一值被设定为“true”。 When the inequality (33), 2cos-1 | qB, 0 | is greater than ATH, a value of a logical expression of the inequality (33) is set to be shown in "true". 当2cos-1|qB,0|不大于ATH时,该逻辑表达式的值被设定为“false”。 When 2cos-1 | qB, 0 | is not greater than the ATH, the value of the logical expression is set to "false".

ATH<2cos-1|qC,0| ...(34) ATH <2cos-1 | qC, 0 | ... (34)

在不等式(34)中,qC,0表示旋转差分值QC的第一分量,以及ATH是与不等式(32)中相应的一个相同。 Inequality (34), qC, 0 represents the rotational differential value QC of the first component, and ATH is the inequality (32) in a corresponding one of the same. 当在不等式(34)中,2cos-1|qC,0|大于ATH时,在不等式(34)中所示的一逻辑表达式的一值被设定为“true”。 When the inequality (34), 2cos-1 | qC, 0 | is greater than ATH, a value of a logical expression of the inequality (34) is set to be shown in "true". 当2cos-1|qC,0|不大于ATH时,该逻辑表达式的值被设定为“false”。 When 2cos-1 | qC, 0 | is not greater than the ATH, the value of the logical expression is set to "false".

旋转方向误差检测器1830在不等式(30)至(34)的逻辑值上执行一AND操作并将该AND操作的结果输出到旋转方向选择器1835中。 And the result of performing an AND operation on the AND operation to (34) of the logical value of the rotation direction error detector 1830 Inequality (30) is outputted to the rotation direction selector 1835.

在一预定时间瞬间t,该第一四元数乘法器1810具有两个输入值Qi和 At a predetermined time instant t, the first quaternion multiplier 1810 has two input values Qi and

第一四元数乘法器1810使用两个输入值Qi和 The first quaternion multiplier 1810 uses two input values Qi and

将一旋转差分值输出。 Will be a rotation differential value output. 如上所述,在区1中,接收旋转差分数据的用于译码已编码关键字值数据的装置顺时针旋转一对象。 As described above, in area 1, receives the rotational differential data of apparatus for decoding encoded key value data of the clockwise rotation of an object. 然而,该对象必须从在按由Qi-1表示的预定度数旋转后当前所处的一位置旋转到假定在按由Qi表示的预定度数旋转后所处的一位置,从而该对象的原始旋转方向必须是一逆时针方向。 However, the object must be from the rear by a predetermined number of degrees of rotation, said Qi-1 is currently located in a position to assume a position after a predetermined number of degrees of rotation indicated by Qi which, thus the original direction of rotation of the object must be a counterclockwise direction.

因此,旋转方向校正器1815校正该对象的旋转方向以便该对象能在基于 Thus, the rotation direction corrector 1815 correcting a rotational direction of the object so that the object can be based on

所旋转的相同方向旋转等于由 The same direction of rotation by a rotation equal to

所表示的一旋转角,如图17所示,即,因此该对象能从基于 Represented by a rotation angle, as shown in Figure 17, i.e., so the object based from

旋转后的当前所处的位置逆时针旋转到假定按等于由 Rotated counter-clockwise current position to assume equal by pressing

表示的一旋转角旋转后所处的位置。 Location indicated after a rotation angle of rotation.

为此,该旋转方向校正器1815建立新的旋转信息以使该对象按差不多1180度逆时针从在按 For this purpose, the rotation direction corrector 1815 to establish a new rotation information to make the object almost 1180 counterclockwise from the press in the press

的预定角度旋转后当前所处的位置旋转。 The current location of the rotation after a predetermined angle of rotation. 因此,该对象的旋转方向被校正成与最初方向相同,以及能最小化一旋转方向误差。 Therefore, the direction of rotation of the object is corrected to the same as the original direction, and a rotational direction that minimizes the error. 该旋转方向校正器1815在步骤S19260中接收来自该第一四元数乘法器1810的当前关键帧的旋转差分数据 The rotation direction corrector 1815 receives the rotational differential data of the current keyframe from the first quaternion multiplier 1810 in step S19260 in

生成具有一校正的旋转方向的旋转差分数据,并将该校正的旋转差分数据输出到旋转方向选择器1835中。 Generating rotational differential data having a corrected rotation direction of, and the corrected data is output to the differential rotation of the rotation direction selector 1835.

参考图19A,在步骤S19300中,旋转方向选择器1835核对从旋转方向误差检测器1830输入的逻辑值是否为真。 Referring to FIG. 19A, in step S19300, the rotation direction selection logic value from 1835 to check the direction of rotation error detector 1830 is true whether the input device. 如果输入的逻辑值为真,在步骤S19400中,该旋转方向选择器1835确定如图17所示的区1发生的相同的现象已经发生并将由等式(29)定义的校正的旋转差分值Qs输出到量化器1840中。 If the logic input is true, in step S19400, the rotation direction selector 1835 identified the same phenomenon in a region as shown in Figure 17 has occurred and will occur by equation (29) defined by the rotation correction differential value Qs output to the quantizer 1840.

另一方面,如果输入的逻辑值是假,在步骤S19500中,旋转方向选择器1835确定没有发生在区1中发生的相同的现象并将从第一四元数乘法器1810输入的旋转差分值 On the other hand, if the value of the input logic is false, in step S19500 in 1835 to determine the direction of rotation of the selector does not occur in zone 1 and the same phenomenon occurring from the rotation quaternion difference value between the first input of the multiplier 1810

输出到量化器1840中。 Output to the quantizer 1840.

在下文中,将描述一种根据本发明第三实施例的关键字值数据编码器300。 Hereinafter, will be described a third embodiment according to the present invention, key value data encoder 300.

由于根据本发明的第一和第二实施例的关键字值数据编码器300仅编码一旋转差分值的四个分量中的三个分量,因为一量化误差它们不能使用等式(24)来复原该四元数的一第一分量值,例如,一复原旋转差分值的一第一分量可能是一虚数。 Since only four component encoder 300 according to a rotation differential value of key value data encoder of the first and second embodiments of the present invention the three components, because they can not use a quantization error of equation (24) to recover The quaternion component values of a first, e.g., a return rotational differential value of a first component may be an imaginary number.

为防止由于仅编码一四元数的三个分量所带来的问题,根据本发明第三实施例的关键字值数据编码器300包括一量化器,其能适当地调节三个量化的分量值以便当译码编码的关键字值数据时,另一分量能被复原成一正实数并减小图像的矢真。 To prevent the encoding only three components of a quaternion problems brought about, key value data encoder according to a third embodiment of the present invention 300 includes a quantizer, which can appropriately adjust three quantized component values so that when decoding the encoded key value data, and the other components can be restored to a positive real number and reduce real vector images.

根据本发明第三实施例的关键字值数据编码器300,除一量化器2000的结构外,与根据本发明的第一或第二实施例的关键字值数据编码器相同,因此下面将仅描述该量化器2000的结构。 Key value data encoder according to a third embodiment of the present invention 300, in addition to the structure of a quantizer 2000, and according to the first or second embodiment of the present invention the same key value data encoder, so the following will only Description of the structure quantizer 2000.

图20A是包含在根据本发明第三实施例的关键字值数据编码器300中的一量化器2000的框图。 20A is a block diagram comprising a third embodiment of the present invention, key value data encoder 300 in 2000 in a quantizer. 参考图20A,一量化器2000包括一量化单元2010,一量化数据调节器2020,一逆量化器2030,以及一误差测量单元2050。 With reference to FIG. 20A, a quantizer 2000 includes a quantization unit 2010, a quantized data adjuster 2020, an inverse quantizer 2030, and an error measurement unit 2050. 该量化单元2010量化输入到其中的一旋转差分值。 The quantization unit 2010 quantizes the input to which a rotational differential value. 该量化数据调节器2020逆调节一量化旋转差分数据。 The quantitative data controller 2020 inverse quantization adjust a rotation differential data. 该逆量化器2030逆量化已量化的旋转差分数据。 The inverse quantizer 2030 inverse-quantizing the quantized rotational differential data. 该旋转差分值复原器2040通过使用逆量化分量来复原量化的旋转差分数据的一第一分量来复原所有量化的旋转差分数据的分量值。 The rotational differential value restorer 2040 by using an inverse quantization component for restoring a first component of the quantized rotational differential data to recover all of the quantized component values of rotational differential data. 该误差测量单元2050测量一复原差分值与一原始输入的旋转差分值间的一误差并更新量化的旋转差分值。 2050 measures the error measuring unit a rotation error recovery a differential value and a difference value between the original input and update the rotational difference quantitative score.

图20B是量化器2010的操作流程图。 Figure 20B is a flowchart of the operation 2010 of the quantizer. 参考图20B,当从第一四元数乘法器输入一旋转差分值 Referring to FIG. 20B, when the input from the first quaternion rotation difference value multiplier

时,在步骤S20050中,该量化器2010使用等式(23)量化输入的旋转差分值 When, in step S20050, the quantizer 2010 using equation (23) quantizes the input rotational differential value

并将量化的旋转差分数据 And to quantify the differential rotation data

输出到量化数据调节器2020中。 Output data to quantify the regulator 2020.

量化数据调节器2020在步骤S20100中用下面的等式(35)来调节量化旋转差分数据 Quantized data using the regulator 2020 in step S20100 the following equation (35) to adjust the quantized rotational differential data

的三个分量。 The three components.

i,j,k,和d是整数) …(35)在等式(35)中,i,j,k均是变量,其将能被添加到旋转差分数据上以便调节旋转差分数据,并被用于定义i,j,k的范围。 i, j, k, and d is an integer) ... (35) in equation (35), i, j, k are variables, which will be added to rotational differential data so as to adjust the rotational differential data, and is is used to define i j, k range. 量化旋转差分数据 Quantify the differential rotation data

的三个调节分量被输出到逆量化器2030中。 The three adjusted components are output to the inverse quantizer 2030.

在步骤S00200中,逆量化器2030逆量化调节量化旋转差分数据Iijk并将逆量化的旋转差分值 In step S00200, the inverse quantizer 2030 inverse quantizing the quantized rotational differential rotational difference adjustment data and the inverse quantized value Iijk

(或逆量化的结果)输出到旋转差分值复原器2040中。 (Or the inverse quantized result) is output to the rotational differential value restorer 2040.

在步骤S20300中,接收逆量化的旋转差分值 In step S20300, the receiving inverse quantized rotational differential value

的三个分量的旋转差分值复原器2040用等式(24)复原旋转差分数据的一第一分量并将一复原旋转差分值输出到误差测量单元2050中。 Rotational differential value restorer 2040 with the three components of the equation (24) restored rotational differential data of a first component and a restored rotational differential value is output to the error measurement unit 2050.

误差测量单元2050核对输入到其中的一旋转差分值的一第一分量值。 Error measuring unit 2050 to check the input to which a first rotational differential value of a component value. 在步骤S20400中,如果该第一分量值是一实数,则误差测量单元2050执行误差测量。 In step S20400, if the first component value is a real number, then the error measurement execution error measurement unit 2050. 另一方面,如果该第一分量值是一虚数,则该方法转入步骤S20600。 On the other hand, if the first component value is an imaginary number, the method proceeds to step S20600.

当该输入旋转差分值的第一分量是一实数时,误差测量单元2050在步骤S20450中测量在一原始旋转差分值和一复原旋转差分值间的一误差eijk并在步骤S20500中核对eijk是否小于一调节误差 When the input rotational difference scores first component is a real number, measurement error measurement unit 2050 in step S20450 in an original rotation differential value and a recovery error eijk a difference value between the rotation and check in step S20500 is less than the eijk an adjustment error

下面将参考图27来描述测量原始旋转差分值与复原旋转差分值间的eijk的方法。 27 below with reference to the method described in the original measured value and the restored rotational differential eijk between the rotational differential value of FIG.

如果测量误差eijk小于调节误差 If the measurement error is less than the adjustment error eijk

在步骤S20550中eijk替换 In step S20550 replace the eijk

然后在步骤S0060中核对变量i,j以及k是否属于一调节范围[-d,+d]。 Then check in step S0060 the variable i, j and k whether an adjustment range [-d, + d]. 如果变量i,j以及k属于调节范围[-d,+d],则误差测量单元2040重复执行步骤S20100至S20550在步骤S20100中,在步骤S20100至S20550的每个循环间,在一嵌套的循环方式中,将量化的旋转差分数据的二至4个分量加1。 If the variable i, j, and k belong to the adjustment range [-d, + d], then repeat step 2040 error measuring units S20100 to S20550 in step S20100, and in between each cycle of steps S20100 to S20550, in a nested cycle mode, the 2-4 components quantized rotational differential data plus 1.

例如,一量化的数据调节器2020保持第二和第三分量,试图通过增加k值到第四分量值上,同时在k值超过+d之前通过连续将k值加1来逐步增加k值使复原第一分量成为一实数,然后找出第四个调节分量值,这能最小化一输入旋转差分值与一复原旋转差分值间的一误差。 For example, a quantized data adjuster 2020 holding the second and third components, tries to continuously adding 1 to the value of k increases gradually so that by increasing the k value k value to the fourth component value while the value of k exceeds + d before by The first component of the recovery into a real number, and then find a fourth component value adjustment, which can minimize the difference between an input value and a rotation error recovery a difference value between the rotation.

如果k值到达+d,该量化数据调节器2020试图通过用-d来初始化k值并将一j值加到第三分量值上同时通过将j(-d≤j≤+d)值加1来增加j值并将一k值加到第四分量值上同时在k值超过+d前通过连续地将k(-d≤k≤+d)值加1来逐步地增加k值来使该复原第一分量变成一实数值,然后找到第四调节分量值,这能减少一输入的旋转差分值与一复原旋转差分值间的一误差。 If the k value reaches + d, the quantitative data regulator 2020 attempt to initialize by -d j k value and a value added to the third component values simultaneously by j (-d≤j≤ + d) value plus 1 to increase the value of j and a k value is added to the fourth component value while the value of k exceeds + d k value before the increase stepwise by successively k (-d≤k≤ + d) adding 1 to a value to make the The first component of the recovery into a real value, and then find the fourth adjustment component values, which can reduce the rotational differential value of an input error with a rotation differential value between a recovery.

如果j值到达+d,该量化数据调节器2020试图通过用-d来初始化j和k值并将一i值加到第二分量值上同时通过将i(-d≤i≤+d)值加1来增加i值并将一j值加到第三分量值和将一k值加到第四分量上同时在k值超过+d前通过连续地将k(-d≤k≤+d)值加1来逐步地增加k值来使该复原第一分量变成一实数值,然后找到第四调节分量值,这能减少一输入的旋转差分值与一复原旋转差分值间的一误差。 If j value reaches + d, the quantitative data regulator 2020 attempt by -d to initialize j and k i values and a value added to the second component values simultaneously by i (-d≤i≤ + d) increase the value of i by adding 1 to the value j and a value is added to the third component value and a k value is added to the fourth component value while k exceeds + d ago by continuously k (-d≤k≤ + d ) to increase the value added to a gradual recovery k values to make the first component into a real value, and then find the fourth adjustment component values, which can reduce the rotational differential value of an input error with a recovery of a difference value between the rotation .

重复上述处理直到i,j,k达到+d为止然后找出能减少一输入的旋转差分值与一复原旋转差分值间的一误差的四个调节分量值。 Above process is repeated until i, j, k reach + d and then find out until to reduce the rotational differential value and a restored four adjusted component values of an error between a rotational differential value input.

误差测量单元2050在步骤S20700中核对调节误差 Error measurement unit 2050 reconciliation of the error in step S20700

是否小于一最终误差 Is less than one final error

同时改变第二至第四分量值。 While changing the second through fourth component values. 如果 In case

小于 Less than

代替 Instead

同时在步骤S20750中用下面的等式(36)来校正量化的旋转差分数据。 Meanwhile step S20750 using the following equation (36) to correct the quantized rotational differential data.

此后,误差测量单元2050将校正的旋转差分数据输出到量化的数据调节器2020中。 Thereafter, the error measurement unit 2050 outputs the corrected rotational differential data to the quantized data adjuster 2020.

量化的数据调节器2020将变量i,j,k值设置为-d并在输入到其中的旋转差分数据上再次执行步骤S20100至S20600。 Quantized data adjuster 2020 variable i, j, k value is set to -d and wherein the rotational input to the differential data in step S20100 through S20600 again performed. 然后,量化数据调节器2020用输入旋转差分数据来核对是否存在一具有比在前存储的最终误差更小的误差的旋转差分数据。 Then, quantized data input rotation regulator 2020 using a differential rotation data to check whether there is a difference data having a smaller error than the preceding stored final error.

如果调节误差 If the adjustment error

大于最终误差 Larger than the final error

误差测量单元2050在步骤S20800中将与当前存储的最终端误差 Error measurement unit 2050 in step S20800 will be stored in the terminal with the current error

相对应的量化的旋转差分数据输出到一循环DPCM运算符中。 Corresponding to the quantized rotational differential data output to an operator in the DPCM loop.

在下文中,将参考图21A至25来更全面地描述一种根据本发明一优选实施例对一位流进行译码的装置和方法,将一定向内插器进行译码成上述位流。 Hereinafter, with reference to FIG. 21A to 25 to be more fully described embodiment of an apparatus and method for decoding a bit stream of a preferred embodiment according to the present invention, a certain inward interpolator into the bit stream for decoding.

图21A是根据本发明一优选实施例的用于译码一定向内插器的装置的框图,以及图21B根据本发明一优选实施例给出了一种用于编码一定向内插器的方法流程图。 21A is a block diagram of apparatus in accordance with a preferred embodiment of the present invention for decoding a certain inward interpolator, and Figure 21B according to a preferred embodiment of the present invention gives a method for encoding a certain inward interpolator flowchart.

参考图21A,用于译码一定向内插器的装置包括一关键字数据译码器2120,一关键字值数据译码器2150,一标题译码器2110,以及一定向内插器合成器2180。 Referring to FIG. 21A, apparatus for decoding a certain inward interposer includes a key data decoder 2120, a key value data decoder 2150, a caption decoder 2110, as well as certain inward plug Synthesizer 2180. 该关键字数据译码器2120译码来自一输入位流的关键字数据。 The key data decoder 2120 decodes the input bit stream from a keyword data. 该关键字值数据译码器2150译码来自该输入位流的关键字值数据。 The key value data decoder 2150 decodes key value data from the input bit stream. 该标题译码器2110译码来自该输入位流的标题信息并将该译码标题信息输出到关键字数据译码器2120和关键字值数据译码器2150。 The subtitle decoder 2110 decodes header information from the input bit stream and outputs the decoded information to the title key data decoder 2120 and the key value data decoder 2150. 该一定向内插器合成器2180合成译码关键字数据和译码关键字值数据并输出一译码定向内插器。 The interposer certain inward synthesizer 2180 synthesizes the decoded key data and key value data and outputting the decoded within a decoder directional interpolator.

参考图21B,由用于编码如图3所示的一定位内插器的装置所编码的一位流在步骤S21000中被输入到关键字数据译码器2120、关键字值数据译码器2150以及标题译码器2110中。 With reference to FIG. 21B, interpolator means by a positioning as shown in Figure 3 for encoding of the encoded bit stream is input to the key data decoder 2120 in step S21000, the key value data decoder 2150 and a subtitle decoder 2110.

在步骤S21100中,标题译码器2110译码来自该输入位流的标题信息并将译码标题信息提供给关键字数据译码器、关键字值数据译码器2150以及定向内插器合成器2180中。 In step S21100, the header decoder 2110 decodes header information from the title information and decodes the input bit stream is supplied to the key data decoder, key value data decoder 2150, and the orientation interpolator synthesizer in 2180.

关键字数据译码器2120熵译码来自输入位流的关键字数据,通过执行一预定逆DND操作、一逆折叠操作以及一逆移位操作来生成译码关键字数据以及将译码关键字数据输出到定向内插器合成器2180中。 Key data decoder 2120 entropy decodes key data from an input bit stream to generate the decoding key and the decoding key data by performing a predetermined inverse DND operation, an inverse folding operation, and an inverse shifting operation Data output to the interior orientation interpolator synthesizer in 2180. 关键字值数据译码器2150熵译码来自输入位流的关键字值数据,生成一旋转差分值,该旋转差分值用来通过几乎等于旋转变换间的一差值来旋转一对象,通过逆量化译码旋转差分数据的每个关键帧的四元数关键字值数据将该旋转变换应用到该对象,通过将一当前关键帧的一旋转差分值四元数乘以一前关键帧的一旋转变换值来复原该当前关键帧的一旋转变换值并在步骤S21200中将该当前关键帧的旋转变换值输出到该定向内插器合成器2180。 Key value data key value data decoder 2150 entropy coding from the input bit stream to generate a rotational difference value, the difference value is used by rotating the rotational almost equal to a difference between the rotation transformation of an object, by the inverse quaternion key value data for each keyframe data to quantify the differential rotation of the decoded rotation transformation is applied to the object by a rotational difference between the current value of a key frame is multiplied by a quaternion before a key frame rotation transformation to recover the value of the conversion value of a rotating current keyframe and step S21200 in the current rotation keyframes transform values into the output of the directional interpolator synthesizer 2180.

定向内插器合成器2180通过将输入到其中的关键字数据和关键字值数据用一浮点生成模式和从标题译码器2110输入的一关键字选择特征位来复原一定向内插器,并在步骤S21400中将复原的定向内插器输出。 Interior orientation interpolator synthesizer 2180 via the input to which key data and key value data generated by a floating-point mode and 2110 decoder input from the title of a keyword selection flag to recover certain inward interpolator and inserted output will recover in step S21400 orientation.

在下文中,将参考图22至23B来描述根据本发明一优选实施例的一关键字数据译码器2120以及根据本发明一优选实施例的一种译码关键字数据的方法。 Hereinafter, 2120 as well as a method for decoding key data according to a preferred embodiment of the present invention, 22 to 23B will be described in accordance with the present invention, a preferred embodiment of a key data decoder with reference to FIG.

图22是关键字数据译码器2120的一框图。 Figure 22 is a block diagram of a key data decoder 2120. 关键字数据译码器2120接收一编码位流以及通过译码将其重建成关键字数据。 Key data decoder 2120 receives an encoded bit stream and its reconstruction by the decoder as the keyword data.

关键字数据译码器2120包括一熵译码器2260、一逆DND处理器2250、一逆折叠处理器2240、一逆移位器2230、一逆DPCM处理器2220、一逆量化器2210、一线性关键字译码器2200、以及一浮点数逆转换器2205。 Key data decoder 2120 includes an entropy decoder 2260, an inverse DND processor 2250, an inverse folding processor 2240, an inverse shifter 2230, an inverse DPCM processor 2220, an inverse quantizer 2210, a Keywords linear decoder 2200, and a floating point inverse converter 2205.

图23A是根据本发明一优选实施例的一种用于译码关键字数据的方法流程图。 A method for decoding key data according to the present invention, FIG. 23A is a preferred embodiment of a flow chart. 参考图22和23A,将关键字数据压缩成其的一位流被输入到标题译码器2110以及熵译码器2260中。 With reference to Figures 22 and 23A, into which the key data of the compressed bit stream is input to the header decoder 2110 and the entropy decoder 2260.

在步骤S23000中,标题译码器2110译码用于译码所需的每个步骤的信息片并将它们提供给译码它们的相应步骤。 In step S23000, the header decoder 2110 decodes pieces of information required for decoding each step of decoding and provides them to their corresponding steps. 用每个译码步骤来描述由标题译码器2110译码的信息。 With each decoding step will be described by the header decoder 2110 decodes the information.

熵译码器2260从标题译码器2110接收将被译码的差分数据的数量以及已经用于编码的位的数量即将被用于译码的的位的数量,以及在步骤S23100中译码输入位流。 2260 2110 number of bits of the received differential data to be decoded from the header decoder entropy decoder number and the number of bits already used for encoding is about to be used for decoding, and the decoding input in step S23100 bitstream. 差分数据的数量等于从关键字数据的数量减去通过执行DPCM获得的内部关键字数据的数量的结果。 Minus the number of differential data is equal to the internal key data by performing a DPCM number of results obtained from the data of the number of keywords.

熵译码器2260识别基于包括在该位流中的预定信息,如在本实施例中的bSignedAACFlag来识别将被译码的差分数据是否具有负值或正值。 Entropy decoder 2260 based on predetermined identification information included in the bit stream, as in this example bSignedAACFlag to identify the differential data to be decoded whether negative or positive value embodiment. 如果编码的差分数据具有负值,熵译码器2260使用函数decodeSignedAAC()来译码它们。 If the encoded differential data have negative values, the entropy decoder 2260 using the function decodeSignedAAC () to decode them. 另一方面,如果编码差分数据仅具有正值,熵译码器2260使用函数decodeUnsignedAAC()来译码它们。 On the other hand, if the encoded differential data have only positive values, the entropy decoder 2260 using the function decodeUnsignedAAC () to decode them. 然后,将译码差分数据传送到逆DND处理器2250。 Then, the decoded difference data is transferred to the inverse DND processor 2250.

逆DND处理器2250从标题译码器2110接收在每个DND周期中的DND的次数以及一最大值nKeyMax。 Frequency inverse DND processor 2250 receives 2110 DND in each cycle of DND from the header decoder and a maximum value nKeyMax.

如果DND的次数为-1,这表示正被译码的编码差分数据已经经过一DPCM操作以及一移位操作而不是经过DND被熵译码,以及该方法直接进入到执行一逆移位操作的步骤上。 If the number of DND is -1, which means that the differential data being decoded have been encoded through a DPCM operation and a shifting operation instead of passing through DND is entropy decoded, and the method proceeds to perform an inverse direct shift operation Step on. 如果DND的次数为0,这表示正被译码的编码差分数据已经经过一折叠操作而不是经过DND被熵译码,因此该方法直接进入到执行一逆折叠操作的步骤上。 If the number of DND is 0, this indicates the encoded difference data being decoded have been through a folding operation instead of passing through DND being entropy coding, so that the method directly proceeds to the step of performing an inverse folding operation. 如果DND次数大于0,在步骤S23200中执行一逆DND操作。 If DND number greater than 0, perform a reverse operation in step S23200 in DND.

逆DND处理器2250在步骤S23300中确定正被译码的编码差分数据是否已经经过一上移操作被编码。 Inverse DND processor 2250 determines the data being encoded difference decoding has elapsed on a shifting operation is encoded in the step S23300. 在本发明的一优选实施例中,通过核对包括在一位流中的nKeyInvertDown是否大于0来确定正被译码的编码差分数据经过一上移操作被编码。 In a preferred embodiment of the present invention, by checking including nKeyInvertDown in a stream is greater than 0 to determine the encoded differential data being decoded after a Move operation is encoded.

如果正被译码的编码差分数据没有经过一上移操作,该方法进入到执行一逆DND的步骤上。 If the encoded differential data being decoded have not been a move operation, the method proceeds to the step of performing an inverse DND. 另一方面,如果正被译码的编码差分数据已经经过一上移操作,通过执行一上移操作已经从一正数区域转移到一负数区域的差分数据在步骤S23400中被移回到负数区域。 On the other hand, if the encoded differential data being decoded have been through a shift operation on a move operation has been transferred from a positive number region to a negative region of the differential data by performing are moved back in the negative region in step S23400 . 在本发明一优选实施例中,已经经过一上移操作的差分数据通过执行用下面的等式来表示的一下移操作(一反转操作)被复原。 In a preferred embodiment of the present invention, the differential data have been through a shift operation by executing the following equation to represent the bit shifting operation (a reverse operation) is restored.

invert-down(v) …(37) invert-down (v) ... (37)

=v (v≤nKeyInvertDown) = V (v≤nKeyInvertDown)

=nKeyInvertDown-v (v>nKeyInvertDown) = NKeyInvertDown-v (v> nKeyInvertDown)

在这里,nKeyInvertDown与用在上移操作中的最大值nKeyMax具有相同值。 Here, nKeyInvertDown and used in the shift operation of the maximum value nKeyMax have the same value. 作为下移操作的结果,具有超过nKeyInvertDown的一值的差分数据被转换成低于-1的负值。 As a result of down operation, with more than nKeyInvertDown difference data is converted into a value lower than the negative value of -1.

根据在每个DND周期中的最大值nKeyMax在经过下移操作的差分数据上有选择地执行一逆下划操作或一逆上划操作。 The maximum value nKeyMax in each cycle of DND selectively execute an inverse operation or an inverse underline the designated operation on the differential data after the operation down.

参考图23B,逆DND处理器2250执行与该差分数据在编码期间经过的一DND操作一样的多次逆DND操作。 With reference to FIG. 23B, the inverse DND processor 2250 performs the difference data during encoding after the DND operation as many times an inverse DND operation. 换句话说,逆DND处理器2250设置等于DND的次数的逆DND的次数的一初始值。 In other words, the inverse DND processor 2250 DND setting an initial value equal to the number of times of inverse DND. 接着,当逆DND处理器2250每次执行一逆DND操作时,从逆DND次数的初值值减1并继续执行逆DND操作直到逆DND的次数变为1为止。 Next, when the inverse DND processor 2250 executing an inverse DND operation each time, from the initial value minus the value of the inverse DND number 1 and continues performing inverse DND operation until the number of inverse DND becomes 1 so far. 逆DND处理器2250在每个DND周期中查找nKeyMax以及在步骤S23510中核对每个nKeyMax是否小于0。 Inverse DND processor 2250 to find nKeyMax in each cycle of DND and check each nKeyMax is smaller than 0 in step S23510 in.

如果nKeyMax小于0,它表示在编码过程中已经执行一上划操作,因此,在步骤S23530中,通过执行一逆上划操作,逆DND处理器2250将正被编码的差分数据的范围扩展到一负数区域。 If nKeyMax is smaller than 0, it indicates in the encoding process has been performed on a designated operation, and therefore, in step S23530 by performing an inverse operation on a draw, an inverse DND processor 2250 will extend the range of being encoded difference data into a negative zone. 在本发明的一优选实施例中,可使用由下述等式(38)定义的一逆上划操作。 In a preferred embodiment of the present invention may be used by the following equation (38) defined by an inverse operation on the draw.

inverse-divide-up(v) …(38) inverse-divide-up (v) ... (38)

=v (v≥0) = V (v≥0)

然而,如果nKeyMax不小于0,逆DND处理器2250核对逆DND的次数是否为1。 However, if nKeyMax is not smaller than 0, the inverse DND 2250 to check the number of times the inverse DND processor is 1. 如果逆DND的次数不为1,它表示对编码过程中所译码的差分数据执行一下划操作,因此逆DND处理器2250通过执行一逆下划操作将差分数据的范围扩展到一正数区域。 If the number of inverse DND is not 1, it indicates the encoding process performed on the decoded differential data about the operation plan, so the inverse DND processor 2250 extends the range of operation designated by the differential data to a positive number region perform an inverse under .

在本发明的一优选实施例中,可使用由下述等式定义的一逆下划操作。 In a preferred embodiment of the present invention, may be used an inverse is defined by the following equation under the designated operation.

inverse-divide-down(v) …(39) inverse-divide-down (v) ... (39)

=v (if v≥0) = V (if v≥0)

如果nKeyMax不小于0且逆DND的次数为1,在步骤S23590中,逆DND处理器2250在执行一逆分割操作后结束整个逆DND操作。 If the number is not less than 0 and inverse nKeyMax DND is 1 in step S23590, the inverse DND processor 2250 to perform a reverse split after the end of the whole operation inverse DND operation. 在本发明的一优选实施例中,可使用由下述等式(40)定义的一逆分割操作。 In a preferred embodiment of the present invention, may be used an inverse split operation by the following equation (40) is defined.

inverse-divide(v) …(40) inverse-divide (v) ... (40)

=v (v≥0) = V (v≥0)

=v+(nKeyMax0+1) (v<0) = V + (nKeyMax0 + 1) (v <0)

已经经过逆DND操作的关键字数据的差分数据被输入到逆折叠处理器2240中,以及逆折叠处理器2240在该差分数据上执行一逆折叠操作以便过去仅在一正数区域中的差分数据在步骤S23600中被划分成正值和负值。 After difference data key data has the inverse DND operation are input to the inverse folding processor 2240, and an inverse folding processor 2240 executing an inverse folding operation on the differential data so that only the difference data in the past a positive number region is divided into positive and negative values in the step S23600. 在本发明的一优选实施例中,可使用由等式(41)定义的一逆折叠操作。 In a preferred embodiment of the present invention may be used by the equation (41) defined by an inverse folding operation.

已经经过逆折叠操作的差分数据被输出到逆移位器2230,以及在步骤S23700中,逆移位器2230将用在编码过程中并从标题译码器2110输入的一模式nKeyShift添加到从逆折叠处理器2240输入的差分数据上。 Has passed through the inverse folding operation of the differential data is output to the inverse shifter 2230, and in step S23700, the inverse shifter 2230 will be used in the encoding process and the input from the header decoder 2110 to add a mode nKeyShift from the inverse differential data input from the folding processor 2240. 该操作由下述等式来表示。 This operation is represented by the following equation.

inverse-shift(v)=v+nKeyshift …(42) inverse-shift (v) = v + nKeyshift ... (42)

逆DPCM处理器在步骤S23800中使用从标题编码器2110输入的DPCM的次数将从逆移位器2230输入的差分数据复原成量化关键字数据。 An inverse DPCM processor number used in step S23800 from the header encoder 2110 inputs of DPCM input from the inverse shifter 2230 into quantized differential data restored key data. 逆移位器2230用下述等式(43)执行与DPCM的次数同样多次数的一逆DPCM操作。 Inverse shifter 2230 by the following equation (43) times as many DPCM performs an inverse DPCM operation times.

v(i+1)=v(i)+delta(i) …(43) v (i + 1) = v (i) + delta (i) ... (43)

在这里,i表示差分数据和关键字数据的一指数,v表示一整数数组,以及delta(i)表示差分数据。 Here, i indicates differential data and key data of an index, v indicates an array of integers, and delta (i) indicates differential data.

已经经过逆DPCM操作的量化关键字数据被输入到逆量化器2210中。 Has an inverse DPCM quantized key data is input to the inverse operation of quantizer 2210. 然后,逆量化器2210从标题译码器2110接收有关量化位的大小nKeyQbit以及用于逆量化的最大和最小值是否已经被浮点数转换器905编码的信息并在步骤S23900k使用下述等式将量化关键字数据转换成逆量化关键字数据。 Then, the inverse quantizer 2210 receives information about quantizing bits 2110 from the header decoder and the maximum and minimum size nKeyQbit for inverse quantization of floating-point number has already been coded information converter 905 and using the following equation in the step S23900k inverse quantizing the quantized key data into the key data.

如果在编码关键字数据的过程中,用于量化的最大和最小值没有被浮点数转换器905转换,将在等式(44)中所示的fKeyMin和fKeyMax分别设置为0和1。 If in the process of encoding key data, the maximum and minimum values for quantization have not been converted floating-point number converter 905, the equation (44) fKeyMin and fKeyMax shown in are set to 0 and 1. 然而,如果浮点数转换器905已经转换用于量化的最大和最小值,被浮点数逆转换器2205逆转换的最大和最小值被分别用作用于逆量化的最大和最小值。 However, if the floating-point number converter 905 has converted maximum and minimum values for quantization, the maximum and minimum values are floating-point number inverse converter 2205 are used as the reverse conversion of the maximum and minimum values used for inverse quantization.

下面将描述实现逆量化逆DND操作的程序代码的例子。 Will be described below to achieve the inverse quantization inverse DND operation program code examples.

从逆量化器2210输出的译码关键字数据被添加到在线性关键字译码器2200中译码的关键字数据中,从而构成译码关键字数据。 Was added decoding key data output from the inverse quantizer 2210 to linear keyword from the decoder 2200 decodes key data, thereby constituting a decoding key data.

在下文中,下面将描述一线性关键字译码过程。 In the following, will be described in a linear key decoding process.

标题译码器2110译码来自一位流的关键字标题信息。 The title key decoder 2110 decodes header information from the bit stream. 如果有关一线性关键字数据的信息存在于该位流中,标题译码器2110将要求用于译码该线性关键字数据区域的开始和结尾关键字输出到浮点数逆转换器2205中并将被编码为线性关键字的关键字的数量输出到线性关键字译码器2200中。 If the information about the linear key data is present in the bit stream, the header decoder 2110 will ask for start and end of the linear key data region is output to a keyword float inverse converter 2205 and decodes is encoded as the number of linear keyword is output to the linear key decoder 2200.

浮点数逆转换器2205将用十进制数表示的线性关键字数据区域的开始和结尾关键字逆转换成二进制数并将该二进制数输出到线性关键字译码器2200中。 Floating-point number inverse converter 2205 will begin and end keywords linear key data region represented by a decimal number and the reversed binary number into a binary number is output to the linear key decoder 2200.

假定将被译码的两个浮点数被称为fKeyMin和fKeyMax,译码fKeyMin的过程如下。 Assumed that the two floating-point numbers to be decoded is referred fKeyMin and fKeyMax, the process of decoding fKeyMin is as follows.

标题译码器2110从一位流读取fKeyMin的位数。 2110 from a stream of bits read fKeyMin title decoder. 如果fKeyMin的位数为0,则fKeyMin被设置为0,以及从该位流读取fKeyMax的位数以便译码fKeyMax。 If the number of fKeyMin is 0, fKeyMin is set to 0, and reads the number of bits to decode fKeyMax fKeyMax from the bitstream. 如果fKeyMax的位数不小于8,这表示fKeyMax已经用IEEE标准754编码过。 If the number is not less than 8 fKeyMax, which means fKeyMax already encoded with IEEE Standard 754. 因此,浮点数fKeyMax在读取其32位后被编码。 Therefore, float fKeyMax at 32 after reading its encoding.

然而,如果fKeyMax的位数在1和7之间,标题译码器2110从该位流读取一特征位。 However, if the number of bits fKeyMax is between 1 and 7, the subtitle decoder 2110 reads from the bit stream a flag. 在本发明的一优选实施例中,如果特征位为1,MinKeyMantissaSign被设置为-1。 In a preferred embodiment of the present invention, if the flag is 1, MinKeyMantissaSign is set to -1. 另一方面,如果特征位为0,MinKeyMantissaSign被设置为1。 On the other hand, if the flag is 0, MinKeyMantissaSign is set to 1. 此后,参考表1获得要求用于编码的位的数量,该表示出了一尾数的位数与编码所需的位的数量间的关系。 Thereafter, with reference to Table 1 to obtain the required number of bits for encoding, the table shows the relationship between the number of bits of a mantissa bits required for encoding. 接着,读取与编码所需的位的数量一样多的该位流的数量并存储在MinKeyMantissa。 Then, as many number of bits required to code reading and the number of the bit stream and stored in MinKeyMantissa. 然后读取该位流的下一位被用与存储在MinKeyMantissa中的尾数的特征位的相同的方式来存储在MinKeyExponentSign中。 Then reads the next bit is used and stored in the bit stream in the same manner as MinKeyMantissa characteristic mantissa bits to store in MinKeyExponentSign in. 与一指数值相对应的该位流的接下来六位被读取并存储在MinKeyExponent中。 And an index value corresponding to the next six bit stream is read and stored in MinKeyExponent in.

通过将从标题译码器2110输入的值代入等式(45)中来复原fKeyMin。 By substituting the value from the subtitle decoder 2110 entered into equation (45) to recover fKeyMin.

复原fKeyMax的过程与复原fKeyMin的过程相同。 Recovery and rehabilitation process fKeyMax fKeyMin same process. 具体来说,在从该位流读取fKeyMax的指数前,确定与fKeyMin的指数相同的值是否被用作fKeyMax的指数。 Specifically, the index fKeyMax before reading from the bit stream, the same value as the index to determine whether or not to be used as fKeyMax fKeyMin index. 如果与fKeyMin的指数相同的值不被用作fKeyMax的指数,与从该位流读取fKeyMin的指数相同的方式从该位流读取fKeyMax的指数。 If the same value fKeyMin index fKeyMax not be used as an index, with the index in the same way fKeyMin read from the bit stream read fKeyMax index from the bit stream.

线性关键字译码器2200从浮点数逆转换器2205接收线性关键字数据区域的开始和结尾关键字并用下述等式(46)译码线性关键字数据区域。 Keywords linear decoder receives the beginning and end 2200 2205 Keywords linear key data from the float area inverse converter with the following equation (46) decoding linear key data area.

(i=0,…nNumberOfLinearKey-1) (I = 0, ... nNumberOfLinearKey-1)

在这里,fKeyMin和fKeyMax分别表示该线性关键字数据区域的开始和结尾关键字数据。 Here, fKeyMin and fKeyMax denote the beginning and end of the key data of the linear key data region.

在线性关键字数据区域中利用上述方法所译码的关键字数据被添加到从逆量化器2210输出的关键字数据中,然后将添加结果输出作为最终关键字数据。 Linear key data region decoded using the above method is added to the key data from the key data outputted inverse quantizer 2210, and then outputs the result of addition as the final key data.

在下文中,将参考图24A至25来描述根据本发明一优选实施例的用于译码一定向内插器的装置以及根据本发明一优选实施例的用于编码关键字值数据的方法。 Hereinafter, 24A to 25 will be described certain inward interpolator apparatus and method for encoding key value data according to a preferred embodiment of the present invention according to a preferred embodiment of the present invention for decoding with reference to FIG.

图24A是根据本发明一优选实施例的一种用于译码一定向内插器的装置的框图。 24A is in accordance with the present invention, a preferred embodiment of an apparatus for decoding a block diagram of certain inwardly interpolator. 参考图24A,用于译码一定向内插器的装置包括一标题译码器2110,一熵译码器2410,一逆循环DPCM运算符2420,一逆量化器2430,以及一四元数乘法器2440。 Referring to FIG. 24A, apparatus for decoding a certain inward interposer includes a caption decoder 2110, an entropy decoder 2410, a reverse cycle DPCM operator 2420, an inverse quantizer 2430, and a quaternion multiplication is 2440. 该标题译码器2110对为译码由来自一输入位流的一四元数所表示的关键字值数据所需的标题信息进行译码并将该译码的标题信息提供给一关键字值数据译码器2150。 The subtitle decoder 2110 pairs of key value data decoded by the number of a quaternion from an input bit stream represented by the header information required for decoding and the decoded header information to a key value data decoder 2150. 该熵译码器2410通过熵译码来自该输入位流的熵编码关键字值数据来生成一循环的DPCM旋转差分数据或量化的旋转差分数据。 The entropy decoder 2410 generates a DPCM loop rotational differential data or quantized rotational differential data by entropy decoding key value data from the entropy coding of the input bit stream. 该逆循环DPCM运算符2420通过在输入到其中的循环DPCM旋转差分数据上执行一逆循环DPCM操作来生成量化的旋转差分数据。 The reverse cycle DPCM operator 2420 by performing a reverse-cycle input to the DPCM operation generates quantized rotational differential data on which the rotation cycle of DPCM difference data. 该逆量化器2430通过逆量化已量化的旋转以及差分数据来生成一旋转差分值。 The inverse quantizer 2430 generates a rotational differential value by inverse quantizing the rotational differential data and quantized. 该一四元数乘法器2440通过将当前关键帧的一旋转差分值四元数乘以一前关键帧的一旋转变换值来生成一当前关键帧的一旋转变换值。 The one quaternion multiplier 2440 by multiplying the difference between the current keyframe a rotation quaternion score a rotation transformation of a former keyframe values to generate a current value of a rotational transform keyframes.

图24B是根据本发明一优选实施例的用于译码关键字值数据的方法流程图。 A method for decoding key value data according to the present invention, FIG. 24B is a preferred embodiment of a flow chart. 参考图24B,一位流被输入到标题译码器2110以及关键字值数据译码器2150的熵译码器2140中,利用根据本发明的用于编码一定向内插器的装置将关键字值数据编码成上述位流。 With reference to FIG. 24B, a bit stream is inputted to the subtitle decoder 2110 and the key value data decoder 2150 entropy decoder 2140, the use of certain inward interpolator according to the present invention, the apparatus for encoding key value data is encoded into the bit stream.

在步骤S24100中,标题译码器2110译码用于译码来自该输入位流的关键字值数据所要求的标题信息并将该译码的标题信息提供给关键字值数据译码器2150中。 In step S24100, the header decoder 2110 decodes header information for decoding key value data from the input bitstream required and supplies the decoded header information to the key value data decoder 2150 in .

在标题信息中,对基于DPCM次数的已被量化成四元数中的第一和第二内部关键字值数据以及多个用来译码其它关键字值数据的特征位进行编码。 In the header information, based on the number of DPCM is quantized into quaternions first and second plurality of internal data, and key value used to decode other key value data are encoded flag.

如果DPCM的次数为1(例如,如果多个特征位之一的nKVDPCMOrder被设置为0),将第一量化的关键字值数据包括在标题信息中作为内部关键字值数据。 If the number of DPCM is 1 (e.g., nKVDPCMOrder if one of the plurality of flag is set to 0), first quantized key value data including key values as internal data in the header information. 如果第一逆量化的关键字值数据的一四元数值满足第一逆量化关键字值数据用下述等式(48)来计算。 If a four-element numerical key value data of the first inverse quantization inverse quantization meet first key value data by the following equation (48) is calculated.

在等式(48)中,仅当类OrilDPCMKeyValueHeader中的nFirstXSign为1或在其它条件下为-1时,xSign为1。 In equation (48), only when the class OrilDPCMKeyValueHeader in nFirstXSign is 1 or under other conditions is -1, xSign 1. Ysign和zSign分别与nFirstYSign和nFirstZSign具有如xSign与nFirstXSign间的关系相同的关系。 Ysign and zSign respectively nFirstYSign and nFirstZSign xSign have such relations with the same relationship between nFirstXSign.

由等式(48)定义的复原四元数分量值被转换成将被用作一定向内插器的一角位移。 By Equation (48) defines restoration quaternion component values are converted to be used as a certain corner of the interpolator inward displacement. 从关键字值数据复原每个角位移能用一四维矢量表示,其中i表示当前关键字数据, Each key can be used to recover the value of the angular displacement data from a four-dimensional vector, where i represents the current key data,

表示一旋转轴的矢量,以及 Represents a vector rotating shaft, and

表示一逆时针旋转角。 Represents a counterclockwise rotation angle. 因此,复原四元数分量值用下面的等式(49)被转换成角位移。 Thus, recovery quaternion component values by the following equation (49) is converted into angular displacement.

如果DPCM的次数为2,例如,如果nKVDPCMOrder被设置为1,第一和第二量化的关键字值数据被包括在标题信息中。 If the number of DPCM is 2, for example, if nKVDPCMOrder is set to 1, the first and second quantized key value data are included in the header information. 用与上面描述的相同的方式复原第一量化的关键字值数据。 In the same manner described above restoring first quantized key value data. 然后,用一种不同的方法复原第二关键字值数据。 Then, using a different method for restoring a second key value data. 换句话说,仅将第二量化关键字值数据的第三分量与一编码的位流一起被传送以及它们的值不是内部关键字值数据而是具有第一关键字值数据的不同的数据假定表示逆量化的关键字值数据的第二关键字值数据的一四元数满足 In other words, only the third component of the second quantized key value data and a coded bit stream is transmitted along with their values are not the internal key value data but having a first key value data different from the data assumed A quaternion represents the inverse quantization meet key value data of the second key value data

用下面的等式(50)来计算。 (50) calculated by the following equation.

在等式(50)中,仅当类OrilDPCMKeyValueHeader中的nSecondXSign为1或在其它条件下为-1时,secondXSign为1。 In equation (50), only when the class OrilDPCMKeyValueHeader in nSecondXSign is 1 or under other conditions is -1, secondXSign 1. secondYsign和secondZSign分别与nSecondYSign和nSecondZSign具有如xSign与nSecondXSign间的关系相同的关系。 secondYsign and secondZSign respectively nSecondYSign and nSecondZSign xSign have such relations with the same relationship between nSecondXSign. 如果表示第二逆量化的关键字值数据的一四元数 If you represent a quaternion inverse quantization of the key values of the second data

满足 Meet

是通过 Through

乘以 Multiplied by

计算得出的。 Calculated. 换句话说, In other words,

标题译码器2110将译码的关键字值数据和译码的标题信息输出给关键字值译码器2150。 Caption decoder 2110 title information output decoded key value data and decoding key value to the decoder 2150.

熵译码器2410接收一位流并利用在步骤S24120至S24128中由标题译码器2110所编码的译码信息对输入的位流进行熵编码,其中关键字值数据的差分数据被编码成上述位流。 Entropy decoder 2410 receives the bit stream and using in step S24120 to S24128 by the subtitle decoder 2110 decodes the encoded bit stream for the input information entropy encoding, wherein the differential data of key value data are encoded into the above-mentioned bitstream.

图25是描述输入到熵译码器2140中的一位流的结构图。 FIG 25 is a configuration diagram describing input to the entropy decoder 2140 of the bit stream. 在图25中,假定N(nNumberOfKeyValue)表示编码的关键字值数据的数量,当DPCM的次数为0时,包括在一位流中的每个分量的旋转差分数据的数量为N-1(0,1,...,nNumberOfKeyValue-2)。 In Figure 25, it is assumed N (nNumberOfKeyValue) indicates the number of encoded key value data, the number of times when the DPCM is 0, the number of each component in a bit stream including the rotational differential data to N-1 (0 , 1, ..., nNumberOfKeyValue-2). 当DPCM的次数为1时,包括在位流中的每个分量的旋转差分数据的数量为N-2(0,1,...,nNumberOfKeyValue-3)。 When the number of DPCM is 1, the bit stream includes a rotation differential data of each component in an amount of N-2 (0,1, ..., nNumberOfKeyValue-3).

熵译码器2410接收来自标题译码器2110的x_keyvalue_flag、y_keyvalue_flag以及z_keyvalue_flag并在步骤S24120中核对x_keyvalue_flag、y_keyvalue_flag以及z_keyvalue_flag每个是否被设定为1。 Entropy decoder 2410 receives x_keyvalue_flag from the subtitle decoder 2110, y_keyvalue_flag and z_keyvalue_flag and check x_keyvalue_flag in step S24120 in, y_keyvalue_flag and z_keyvalue_flag whether each is set to 1.

在步骤S24l22中,当x_keyvalue_flag、y_keyvalue_flag以及z_keyvalue_flag每个均被设定为0时,被认为所有量化的关键字值数据或每个分量的所有差分数据同在类OrilKeyValueCodingBit中的nAllKeyValues相同。 In step S24l22, when x_keyvalue_flag, y_keyvalue_flag and z_keyvalue_flag each are set to zero, the difference is considered that all quantized key value data of all the data with the or each component in the same class OrilKeyValueCodingBit in nAllKeyValues. 因此,熵译码器2410将每个分量的关键字值数据译码成与从标题译码器2110输入的nAllKeyValues的相同值并将译码的关键字值数据输出到逆循环DPCM运算符2420中。 Therefore, the entropy decoder 2410 outputs key value data for each component of the key value data decoded into 2110 with the title of the decoder input nAllKeyValues the same value and decoded into reverse cycle DPCM operator 2420 .

如果x_keyvalue_fiag、y_keyvalue_flag以及z_keyvalue_flag不是设定为0,例如,如果x_keyvalue_flag、y_keyvalue_flag以及z_keyvalue_flag被设定为1,熵译码器2410核对从关键字值标题译码器2110输入的一熵译码模式以便在步骤S24124中译码输入的关键字值数据的每个分量 If x_keyvalue_fiag, y_keyvalue_flag and z_keyvalue_flag not set to 0, for example, if x_keyvalue_flag, y_keyvalue_flag and z_keyvalue_flag is set to 1, the entropy decoder 2410 check from the title key value decoder 2110 an entropy coding mode in order to enter Each component in the decoding step S24124 input key value data

当熵译码模式是一二进制译码模式时,在步骤S24126中,熵译码器2420使用如表3所示的一函数decodeSignedAAC()译码一自适应的算术上的编码的位流并将量化的旋转差分数据输出到逆循环DPCM运算符2420中。 When the entropy coding mode is a binary decoding mode, a function decodeSignedAAC step S24126, the entropy decoder 2420 used as shown in Table 3 () encoded by an adaptive arithmetic decoding on the bit stream and outputs quantized rotational differential data to the inverse DPCM operator 2420 cycles.

表3 Table 3

另一方面,当熵译码模式不是一二进制译码模式时,在步骤S24128中,熵译码器2410使用一函数decodeUnaryAAC()译码输入位流。 On the other hand, when the entropy coding mode is not a binary decoding mode, in step S24128, the entropy decoder 2410 uses a function decodeUnaryAAC () decodes the input bit stream. 函数decodeUnaryAAC()通过连续地读取0直到从该位流中读出1为止、将连续Os的数量转换成它的绝对值,读取与位“1”邻近的一位,并将该位转换成一特征位来译码该输入位流,然后将量化的旋转差分值输出到逆循环DPCM运算符2420中。 Function decodeUnaryAAC () by continuously reads 0 until read out from the bit stream a date, the number of consecutive Os converted to its absolute value, read "1" adjacent a bit, and the bit conversion into a flag for decoding the input bit stream, and then the quantized rotational differential value is output to the reverse cycle DPCM operator 2420. 函数decodeUnaryAAC()的例子如表4所示。 Function decodeUnaryAAC () shown in the example in Table 4.

表4 Table 4

在上述函数decodeSignedAAC()和decodeUnaryAAC()中所采用的qf_decode()被用来从一自适应的算术上的编码的位流中读取位1并且由国际标准组织授权的文献ISO/IEC14496-2:1999Coding of Audio-VisualObject:Visual定义。 In the above-described function decodeSignedAAC () and decodeUnaryAAC () adopted in the qf_decode () is used to read from an adaptive arithmetic coding on the bit stream in bit 1 and authorized by the International Standards Organization document ISO / IEC14496-2 : 1999Coding of Audio-VisualObject: Visual definition.

逆循环DPCM运算符2420从熵译码器2410接收熵译码的关键字值数据并核对从关键字值标题译码器2110输入的DPCM的次数。 Reverse cycle DPCM operator 2420 2410 to receive from the entropy entropy decoding key value data decoder and check the number of key values from the title DPCM decoder 2110 input. 如果DPCM的次数为0,在步骤S24130中,逆循环DPCM运算符2420将译码的旋转差分数据 If the number of DPCM is 0, in step S24130, the inverse DPCM operator 2420 will loop decoded rotational differential data

输出到逆量化器2430中,因为从熵译码器2410输入的熵译码的关键字值数据是量化的旋转差分数据。 Output to the inverse quantizer 2430, as key value data input from the entropy decoder 2410 entropy coding is to quantify the differential rotation data.

另一方面,如果DPCM的次数为1,逆循环DPCM运算符2420在步骤S24135执行一逆循环DPCM操作,因为从熵译码器2410输入的熵译码关键字值数据是循环的DPCM旋转差分数据。 On the other hand, if the number of DPCM is 1, the reverse cycle DPCM operator 2420 performing a reverse-cycle DPCM operation in step S24135, as the entropy decoding key value data input from the entropy decoder 2410 is DPCM loop rotational differential data .

假定nKVQBit表示逆量化位,逆循环DPCM运算符2420在旋转差分数据 Assume nKVQBit represents the inverse quantization bit, reverse cycle DPCM operator 2420 data in the differential rotation

上用等式(51)执行一逆循环DPCM操作并在步骤S24135中生成量化的循环差分数据 The equation (51) to perform a reverse cycle operation and generates quantized DPCM loop differential data in step S24135 in

(i=2,…nNumberOfKeyValue-1) (I = 2, ... nNumberOfKeyValue-1)

此后,逆循环DPCM运算符2420分别使用 Since then, reverse cycle using DPCM operator 2420 respectively

And

获得一逆DPCM值A以及一逆DPCM值B,如等式(52)所示。 Obtaining an inverse DPCM value A and an inverse DPCM value B, as shown in equation (52) below.

(i=2,…,nNumberOfKeyValue-1) (I = 2, ..., nNumberOfKeyValue-1)

如果B+(2nKVQBit-1-1)位于0和在一量化范围内的一最大值之间,逆循环DPCM运算符2420将B输出作为逆循环DPCM数据 If B + (2nKVQBit-1-1) is located between 0 and quantified within a range of a maximum reverse cycle DPCM operator B 2420 will output data as a reverse cycle DPCM

另一方面,如果B+(2nKVQBit-1-1)小于0或大于在量化范围内的最大值时,逆循环DPCM运算符2420输出A作为 On the other hand, if the B + (2nKVQBit-1-1) is less than 0 or greater than the maximum value in the quantization range, and the reverse cycle DPCM operator 2420 outputs A as

C十十程序代码的一个例子,其中写出了逆循环DPCM运算符2420的上述操作,如图3所示。 An example C program code size ten, which wrote the reverse cycle operation above DPCM operator 2420, shown in Figure 3.

表5逆量化器2430接收由执45-一逆循环DPCM操作的逆循环DPCM运算符2420生成的量化的旋转差分数据通过在 Table 5 inverse quantizer 2430 receives the quantized rotational differential data by executing a reverse-cycle 45- DPCM operation inverse DPCM operator 2420 cycles generated by

上用等式(53)执行一逆量化操作来复原一旋转差分值将在步骤S24140中旋转差分值 The equation (53) to perform an inverse quantization operation to recover a rotation differential value will rotate the differential value in step S24140

输出到四元数乘法器2440中。 Quaternion output to the multiplier 2440.

(f=2,…,nNumberOfKeyValue-1,j=1,2,3) (F = 2, ..., nNumberOfKeyValue-1, j = 1,2,3)

此后,四元数乘法器2440接收旋转差分值 Since then, the quaternion rotation difference value multiplier receives 2440

四元数乘法器2440在步骤S24154中通过将输入旋转差分值 Quaternion multiplier 2440 by the rotation of the input of the differential value in step S24154

按下面的等式(54)四元数乘以一前关键帧的一旋转变换值 According to the following equation (54) multiplied by a quaternion key frame before a rotational transformation value

复原一当前关键帧的旋转变换值 Restoration of a current rotation transformation keyframe

(i=2,…nNumberOfKeyValue-1) (I = 2, ... nNumberOfKeyValue-1)

在复原一旋转变换值后,关键字值数据译码器2400在步骤S24156中核对复原旋转变换值是否与最后一个关键字值数据相对应。 During the recovery of a rotation transformation value, key value data decoder 2400 in step S24156 check whether the recovery value of the final rotational transform data corresponding to a key value. 如果复原旋转变换值与最后一个关键字值数据不相对应,关键字值数据译码器2400重复执行步骤S24140至S24154。 If the recovery rotational transformation value of the last key value data does not correspond, keyword value data decoder 2400 Repeat steps S24140 to S24154. 另一方面,如果复原旋转变换值是最后一个关键字值数据,则在步骤S24158中,关键字值数据译码器2400将复原旋转变换值输出。 On the other hand, if the restored rotational transformation value is the last key value data in step S24158, the key value data decoder 2400 outputs the restored rotational transformation value.

如果DPCM的次数为0,逆循环DPCM运算符2420将译码的量化的旋转差分数据 If the number of DPCM is 0, the reverse cycle DPCM operator 2420 will be decoded quantized rotational differential data

(其中 (Among them

)输出到逆量化器2430中。 ) Is output to the inverse quantizer 2430. 然后,通过逆量化从逆循环DPCM运算符2420输入的旋转差分数据 Then, from the reverse cycle by the inverse quantization DPCM operator 2420 rotational differential data input

来生成一旋转差分值 To generate a rotation differential value

(其中并在步骤S24140中将Qi输出到四元数乘法器2440中。 (Which will be in step S24140 Qi quaternion output to the multiplier 2440.

(i=1,…,nNumberOfKeyValue-1,j=1,2,3) (I = 1, ..., nNumberOfKeyValue-1, j = 1,2,3)

即使当DPCM的次数为0,四元数乘法器2440将与当DPCM的次数不为0时的几乎相同的方式(步骤S24150至S24158)来复原旋转变换值。 Even when the number of DPCM is 0, the quaternion multiplier 2440 to almost the same way (steps S24150 through S24158) and the number of times when the DPCM is not 0 when restores the rotational transformation value. 当DPCM的次数不为0时,四元数乘法器2440用等式(56)执行四元数乘法。 When the number is not DPCM 0:00, quaternion multiplier 2440 by equation (56) to perform quaternion multiplication.

(i=1,…nNumberOfKeyValue-1) (I = 1, ... nNumberOfKeyValue-1)

在下文中,将描述定向内插器合成器2180的操作。 Hereinafter, the orientation interpolator synthesizer 2180 will be described.

图26是定向内插器合成器2180的操作流程图。 Figure 26 is a directional interpolator synthesizer 2180 of an operational flowchart. 参考图26,在步骤S26000中,定向内插器合成器2180接收译码关键字数据和译码关键字值数据并从标题译码器2110接收关键字数据的数量以及一断点生成模式以及关键字选择特征位。 With reference to FIG. 26, in step S26000, the interior orientation interpolator synthesizer 2180 receives decoded key data and decoded key value data and receives the number of key data and a break from the title key generating mode and the decoder 2110 word selection flag.

在步骤S26100中,当断点生成模式是一动画路径保存模式时,定向内插器合成器2180输出译码关键字数据和译码关键字值数据。 In step S26100, when the breakpoint generation mode is a path to save the animation mode, the internal directional interpolator synthesizer 2180 output decoding key data and decoding key value data. 在一动画路径保存模式中已经抽取断点的情况下,关键字数据与关键字值数据相对应,因此,没有必要使用内插生成新的关键字值数据。 In the case of an animation path-saving mode has been extracted breakpoint, key data and key value data corresponding to, and therefore, there is no need to generate new key value is inserted within the data used. 然而,如果在一动画关键字保存模式中已经抽取断点,这表示仅与断点相对应的关键字值数据被编码同时所有关键字数据已经被编码和译码。 However, if a movie has been extracted keywords saving mode breakpoint, which means that only the breakpoint corresponding to the key value data are encoded while all key data have been encoded and decoded. 因此,在这种情况下,关键字数据与关键字值数据不相对应,因此有必要内插将在下述段路中描述的关键字值数据。 Therefore, in this case, the key data and key value data does not correspond, so the key value data have the necessary interpolation will be described in the following section of path.

当断点生成模式是一动画关键字保存模式时,在步骤S26200中,表示关键字选择特征位的数据的一指数的一计数器被设置从而与关键字数据相对应,以便计算与该关键字数据相对应的关键字值数据是否存在。 When a breakpoint generation mode is an animation key save mode, in step S26200, indicates keyword selection characteristics of a data bit of an index counter is set so as to correspond with the key data in order to calculate the key data whether the corresponding key value data exists.

定向内插器合成器2180在步骤S26300中核对如果与由该计数器表示的关键字数据相对应的关键字值数据存在的话,则转入步骤S26500。 Interior orientation interpolator synthesizer 2180 in step S26300 check if the key data from the counter indicates a value corresponding to the key data exists, then proceeds to step S26500.

如果没有与由计数器表示的关键字数据相对应的关键字值数据,在步骤S26400中,通过使用与由两个连续断点表示的关键字数据相对应的关键字值数据,线性内插当前关键字值数据,定向内插器合成器2180生成关键字值数据,这两个连续断点包括在由该计数器所表示的关键字数据前的一个断点以及在由该计数器表示的关键字数据后的另一个断点。 If there is no key data indicated by the counter value corresponding to the key data in step S26400 by using the key data indicated by two successive break key value corresponding to the data, a linear interpolation of the current key word value data, within the directional interpolator synthesizer 2180 generates key value data, which included two consecutive breakpoints in a breakpoint key data represented by the counter in front of and in the keyword data represented by the counter after Another breakpoint.

此后,定向内插器合成器2180核对所有关键字值数据是否具有它们相应的关键字值数据并且在步骤S26500中对所有不具有与其相应的关键字值数据的关键字数据执行线性操作。 Since then, the interior orientation interpolator synthesizer 2180 for all key value data to check whether their corresponding key value data and S26500 perform linear operation on all the key data that does not have its corresponding key value data in step. 如果仍然存在还没有被核对的关键字数据,定向内插器合成器2180更新计数器并再次执行步骤S26300至S26500。 If you still have not been checked for key data, within the orientation interpolator synthesizer 2180 update and perform the procedure again counter S26300 to S26500. 如果已经核对所有关键字数据,定向内插器合成器2180在步骤S26600中输出合成关键字数据和合成关键字值数据作为一定向内插器。 If you have checked all the key data within the directional interpolator synthesizer 2180 in step S26600 output key data synthesis and synthetic data as a key value certain inward interposer.

在下文中,将描述SDL语言程序代码的例子,通过该SDL语言程序代码,可实现用于根据本发明的译码一定向内插器的装置,该定向内插器译码关键字数据和关键字值数据。 Hereinafter, will be described examples of SDL-language program code, through the SDL language program code, can be achieved according to a certain inwardly interpolator means for decoding the present invention, within the orientation interpolator decoding key data and key value data.

图29是描述一类CompressedOrientationInterpolator的图。 Figure 29 is a diagram describing a class CompressedOrientationInterpolator. CompressedOrientationInterpolator是一个用于读取一定向内插器的一编码的位流的一最高类。 CompressedOrientationInterpolator is a top class for reading a certain inward interpolator an encoded bit stream. KeyHeader、KeySelectionFlag和Key均是用于从一位流读取与在一常规内插器的关键字现场数据相对应的关键字数据信息的类。 KeyHeader, KeySelectionFlag and Key are for a stream reader and plug in a regular spot in the keywords corresponding to the data type from the keyword data information. OrilKeyValueHeader和OrilDPCMKeyValue是用于读取有关与在一常规的定向内插器中关键字值现场数据相对应的关键字值数据的信息的类。 OrilKeyValueHeader and OrilDPCMKeyValue class is used to read information about the inserted within the vessel of a conventional directional key value field data corresponding to the key value data. 函数qf_start()是使用自适应算术编码(AAC),用来在读取一位流前初始化算术译码器,其中编码关键字值数据。 Function qf_start () is the use of adaptive arithmetic coding (AAC), is used to initialize the stream before reading an arithmetic decoder, wherein the encoded key value data.

图29B给出了对译码所需的一关键字标题进行译码的一关键字标题类的图。 Figure 29B shows the desired one of the decoding key for decoding a title keyword title class diagram.

关键字标题类包括关键字数据的数量、量化位、内部关键字数据、一DND标题以及实际用于译码译码的位的数量。 Keyword Title category includes the number of the number of key data, the quantized bits, internal key data, a DND and the actual title for decoding coded bits. nKeyQBit表示用在逆量化中来复原浮点关键字值的量化位。 nKeyQBit expressed in inverse quantization to recover the quantization bit floating-point value of the keyword. nNumKeyCodingBit表示nNumberOfKey的一位长,nNumberOfKey表示关键字数据的数量。 nNumKeyCodingBit represents a long nNumberOfKey of, nNumberOfKey represents the number of key data. nkeyDigit表示原始关键字数据的多个最高有效位并用来舍入译码值。 nkeyDigit representing a plurality of most significant bits of the original key data and used to decode the value of rounding.

当有关线性关键字子区域的信息被包括在一关键字标题中时,一特征位blsLinearKeySubRegion被设置为1。 When information on linear key sub-regions is included in a key header, a flag is set to 1 blsLinearKeySubRegion. 在这种情况下,使用其后的一特征位blsLinearKeySubRegion的译码标题信息能计算包括在一整个关键字区域的特定线性关键字子区域中的关键字。 In this case, the use of a subsequent flag blsLinearKeySubRegion can calculate a decoding header information included in a particular sub-region of a linear key region in the entire keyword keywords. bRangeFlag表示关键字数据范围是否从0到1。 bRangeFlag indicate whether key data range from 0-1. 如果关键字数据范围不是从0到1,在关键字数据范围中的最小和最大值被从一类KeyMinMax类被译码。 If the keyword is not a data range from 0 to 1, the minimum and maximum values in the keyword data range is from a class of KeyMinMax class is decoded. KeyMinMax重新获得要求用于逆量化的最小和最大值。 KeyMinMax regain requirements for the minimum and maximum values of the inverse quantization. 上述每个值能被划分成它们各自的尾数和指数。 Each of these values can be divided into their respective mantissa and exponent.

nBitSize是nQIntraKey的位的大小,以及nKeyShift是nKeyMax的一初始位的大小。 nBitSize is nQIntraKey size bits, and nKeyShift is an initial position of the size nKeyMax. nQIntraKey表示第一量化内部数据的大小并与表示nQIntraKey的一特征位的nQintraKeySign结合。 nQIntraKey represents the size of the first quantization and the internal data indicating a flag nQintraKeySign nQIntraKey binding. nQintraKey被用作为复原其它量化关键字数据所需的一基数。 nQintraKey is used as the quantized key data restored other required a base. 在用在内插器压缩中的所有特征位中,一值“0”表示一正号,以及一值“1”表示一负号。 Interpolator compression using all features bits, a value "0" indicates a positive number, and a value of "1" indicates a negative sign. nKDPCMOrder表示一值,该值是DPCM的次数减1。 nKDPCMOrder represents a value that is the number of DPCM minus one. DPCM的次数可是1、2或3。 DPCM is 1, 2 or 3, but the number of times. 量化内部数据的数量与DPCM的次数相同。 Internal data quantized DPCM number of times with the same.

nKeyShift以及一符号位nKeyShiftSign是表示在一关键字数据译码器中的位移量的一整数。 nKeyShift and a sign bit nKeyShiftSign is showing a key data decoder in the amount of displacement of an integer. 如果bShiftFlag被设置为“true”,译码nKeyShift以及nKeyShiftSign。 If bShiftFlag is set to "true", as well as decoding nKeyShift nKeyShiftSign. nDNDOrder是二次划分(DND)操作的次数。 nDNDOrder secondary divide (DND) number of operations. DND操作在上面已经结合一关键字数据译码器描述过。 DND has been combined with an operation in the above key data decoder described. 如果nDNDOrder的值为7,则译码bNoDND。 If nDNDOrder value of 7, the decoding bNoDND. bNoDND表示是否将执行一逆DND操作。 bNoDND indicates whether DND will perform an inverse operation. nKeyMax是在一逆DND操作的连续周期期间所使用的一最大值或一最小值。 nKeyMax is a maximum or a minimum value of an inverse DND operation during successive cycles are used. nKeyCodingBit表示用于编码关键字数据的位。 nKeyCodingBit denotes a bit coded key data.

bSignedAACFlag表示在AAC译码期间将执行的一译码过程。 bSignedAACFlag indicates a decoding process to be performed during AAC decoding. 如果bSignedAACFlag被设置为0,将执行一无符号的AAC译码过程。 If bSignedAACFlag is set to 0, AAC decoding process will be executed one unsigned. 否则,将执行一带符号的AAC译码过程。 Otherwise, will perform a signed AAC decoding process. bKeyInvertDownFlag是表示是否将使用nKeyInvertDown的一布尔值。 bKeyInvertDownFlag nKeyInvertDown indicates whether the use of a Boolean value. nKeyInvertDown是用来将大于其自身的所有量化关键字数据转换成不小于-1的负数值的一整数。 nKeyInvertDown is used to an integer greater than all its quantized key data into non-negative value less than -1. 如果nKeyInvertDown被设置为-1,那幺将不执行一下移操作。 If nKeyInvertDown is set to -1, NA me about the shifting operation will not be performed.

图29C是表示一类LinearKey的图。 29C is a diagram represents a class LinearKey. 在图29C中,nNumLinearKeyCodingBit是表示对多个预定线性预测关键字进行编码所必须的位的数量。 In Figure 29C, nNumLinearKeyCodingBit shows a plurality of predetermined number of bits of linear predictive coding keyword necessary. nNumberOfLinearKey是表示线性预测关键字的一值。 nNumberOfLinearKey is a linear predictive value keywords.

图29D是描述一类KeyMinMax的图。 FIG. 29D is a diagram describing a class KeyMinMax. 在图29D中,bMinKeyDigitSame是表示所有关键字的最高有效位的数量(nKeyDigit)和在关键字中的一最小值的最高有效位的数量是否相同的一特征位。 In the FIG. 29D, bMinKeyDigitSame is the most significant bit of the number (nKeyDigit) all keywords and the most significant bit of a minimum number of keywords is the same in a flag. nMinKeyDigit是表示在关键字中最小值的最高有效位的数量的一值。 nMinKeyDigit is the most significant bit indicates the number of a minimum value in keywords. nMinKeyMantissaSign是表示nMinKeyMantissa的一特征位的一值。 nMinKeyMantissaSign is nMinKeyMantissa of a flag of a value. nMinKeyMantissa是表示在关键字中最小值的尾数的一值。 nMinKeyMantissa is a minimum value in keywords mantissa. nMinKeyExponentSign是表示nMinKeyExponent的一特征位的一值。 nMinKeyExponentSign is nMinKeyExponent of a flag of a value.

nMinKeyExponent是表示在关键字中最小值的尾数的一值。 nMinKeyExponent is a minimum value of the keyword in the mantissa.

fKeyMin是表示在关键字中最小值的一值。 fKeyMin is a minimum value in keywords. bMaxKeyDigitSame是表示所有关键字的最高有效位的数量nKeyDigit和在关键字中的一最大值的最高有效位的数量是否相同的一特征位。 bMaxKeyDigitSame all keywords is the most significant bit of the number nKeyDigit and the most significant bit of a maximum number of keywords is the same in a flag. nMaxKeyDigit是表示在关键字中最大值的最高有效位的数量的一值。 nMaxKeyDigit is a number of the most significant bit of a value in the maximum value of the keyword. nMaxnKeyMantissaSign是表示nMaxnKeyMantissa的一特征位的一值。 nMaxnKeyMantissaSign is nMaxnKeyMantissa of a flag of a value. nMaxnKeyMantissa是表示在关键字中最大值的尾数的一值。 nMaxnKeyMantissa is a value in the maximum value of the mantissa keywords.

bSameExponent是表示在关键字中最大值的尾数是否与nMinKeyExponent相同的一特征位。 bSameExponent indicates whether the maximum value of the mantissa keyword identical with nMinKeyExponent a flag. nMaxKeyExponentSign是表示nMaxKeyExponent的一特征位的一值。 nMaxKeyExponentSign is nMaxKeyExponent of a flag of a value. nMaxKeyExponent是表示在关键字中最大值的尾数的一值。 nMaxKeyExponent shows a maximum value of the keyword in the mantissa. FkeyMax是表示在关键字中最大值的一值。 FkeyMax is a maximum value in the keywords.

图29E是描述一类OrilKeyValueHeader的图。 Figure 29E is a diagram describing a class OrilKeyValueHeader. 用在该类中的每个变量的含义如下。 Meaning with each variable in the class is as follows. bPreserverKey表示一当前译码模式是否是一动画关键字保存模式或一动画路径保存模式。 bPreserverKey represents a current decoding mode is an animation key-saving mode or a path to save the animation mode. 当bPreserverKey被设置为“true”时,当前译码模式是一动画关键字保存模式。 When bPreserverKey is set to "true", the current decoding mode is an animation key save mode.

nKVQBit表示关键字值数据的一逆量化位的大小。 nKVQBit represents the size of an inverse quantization bit key value data. nKVDPCMOrder表示用于译码关键字值数据的逆DPCM操作的次数。 nKVDPCMOrder represents the number of inverse DPCM decoding key value data for the operation. 当nKVDPCMOrder=0时,不需要执行一逆循环DPCM操作。 When nKVDPCMOrder = 0, the need to perform a reverse-cycle DPCM operation. 另一方面,当nKVDPCMOrder=1时,应该执行一具有2次方的逆循环DPCM操作。 On the other hand, when nKVDPCMOrder = 1, it should perform a power of 2 having an inverse DPCM operation cycle.

图29F是描述一类OriDPCMKeyValueHeader的图。 Figure 29F is a diagram describing a class OriDPCMKeyValueHeader. 用在该类中的每个变量的含义如下。 Meaning with each variable in the class is as follows. firstQKV_S、firstQKV_X、firstQKV_Y以及firstQKV_Z分别表示四个分量s,x,y以及z的第一值,构成一四元数(s,x,y,z)表示量化的关键字值数据。 firstQKV_S, firstQKV_X, firstQKV_Y and firstQKV_Z represent four components s, x, y and z values of the first, constituting a quaternion (s, x, y, z) represents the quantized key value data. nfirstXSign,nfirstYSign以及nfirstZSign分别表示firstQKV_X、firstQKV_Y以及firstQKV_Z的特征位。 nfirstXSign, nfirstYSign and nfirstZSign represent firstQKV_X, firstQKV_Y and firstQKV_Z the flag. secondQKV_X、secondQKV_Y以及secondQKV_Z分别表示三个分量x,y以及z的第二值,以及nsecondXSign,nsecondYSign以及nsecondZSign分别表示secondQKV_X、secondQKV_Y以及secondQKV_Z的特征位。 secondQKV_X, secondQKV_Y and secondQKV_Z denote the three components x, y and z of the second value, and nsecondXSign, nsecondYSign and nsecondZSign represent secondQKV_X, secondQKV_Y and secondQKV_Z the flag. BblsMoreTwoKVs表示当必须执行一逆循环DPCM时是否有将被译码的两个以上的关键字值数据。 BblsMoreTwoKVs representation must be performed when there will be decoded if two or more key value data is a reverse cycle DPCM. x_keyvalue_flag、y_keyValue_flag以及z_keyvalue_flag表示每个分量x,y以及z的所有量化值是否相同。 x_keyvalue_flag, y_keyValue_flag and z_keyvalue_flag represent each component x, y and z all quantized values are the same.

图29G是描述一类OrilKeyValueCodingBit的图。 Figure 29G is a description of a class OrilKeyValueCodingBit Fig. 在该类中每个变量的含义如下。 The meaning of each variable in the class below. NnKVCodingBit表示用来存储量化后除内部关键字值数据(在类OrilDPCMKeyValueHeader中的firstQKV_S、firstQKV_X、firstQKV_Y、firstQKV_Z、secondQKV_X、secondQKV_Y以及secondQKV_Z外的所有量化关键字值数据的位的数量。 After NnKVCodingBit represents the number of quantization bits used to store the key value in addition to the internal data (in class OrilDPCMKeyValueHeader in firstQKV_S, firstQKV_X, firstQKV_Y, firstQKV_Z, secondQKV_X, secondQKV_Y and secondQKV_Z foreign key values for all the quantized data.

nAllKeyValue表示当用于每个分量的keyvalue_flag设置为0时所有关键字值数据的每个分量的一量化的值。 nAllKeyValue expressed keyvalue_flag settings for each component when the value of a quantized value of each component of the data of all the keywords 0:00. nSign表示nAllKeyValue的特征位。 nSign nAllKeyValue the flag represents. BlsUnaryAAC表示用于译码每个分量x,y和z的量化值的一种自适应算术量化方法。 BlsUnaryAAC represents for decoding each component x, a quantized value of y and z adaptive arithmetic quantization method. 如果BlsUnaryAAC被设置为“true”,将使用一unaryAAC译码函数。 If BlsUnaryAAC is set to "true", will use a unaryAAC decoding functions. 另一方面,如果BlsUnaryAAC被设置为“false”,则将使用一binaryAAC译码函数。 On the other hand, if BlsUnaryAAC is set to "false", then use a binaryAAC decoding functions.

图29H是描述一类KeySelectionFlag的图。 Figure 29H is a description of a class KeySelectionFlag Fig. 在该类KeySelectionFlag中,keyFlag表示是否已经编码第i个关键字数据。 In such KeySelectionFlag in, keyFlag indicate whether the i-th key data coding. nNumOfKeyValue是表示将被译码的关键字值数据的数量的一整数。 nNumOfKeyValue is a number to be decoded key value data of an integer.

图29I是描述一类Key的图。 Figure 29I is a description of a class of Key Figure. 在该类Key中,nQKey是从一位流译码的量化关键数据的一数组。 Key in the class, nQKey quantify critical data bit stream is decoded from one array. KeyContext是用于读取nQKey的一大小的一环境(context)。 KeyContext is used to read a size of a nQKey environment (context). KeySignContext是用于读取nQKey的一符号的一环境。 KeySignContext nQKey is for reading a symbol of an environment.

DecodeUnsignedAAc是用来执行具有一给定环境的自适应算术编码的一不带符号的译码过程的一函数,这将在下面描述。 DecodeUnsignedAAc is used to execute a function having a given environment adaptive arithmetic coding a symbol without the decoding process, which will be described below. DecodeSignedAAC是用于执行具有一给定环境的自适应算术编码的一带符号的译码过程的一函数,这将在下面描述。 DecodeSignedAAC is a function for executing a signed with a given environment adaptive arithmetic coding decoding process, which will be described below.

图29J是描述一类OrilDPCMKeyValue的图。 Figure 29J is a description of a class OrilDPCMKeyValue Fig. 在该类中每个变量的含义如下。 The meaning of each variable in the class below. DeltaKeyValue被用来以四元数的形式存储包括三个分量x,y和z的量化的关键字值数据。 DeltaKeyValue is used in the form of quaternions stored quantized key value data including the three components x, y and z. 存储在DeltaKeyValue中的量化的关键字值数据使用一函数decodeUnaryAAC或decodeSignedAAC从一位流中被译码。 Quantized key value data stored in the use of a function DeltaKeyValue decodeUnaryAAC or decodeSignedAAC be decoded from a stream.

kVXSignContext,kVYSignContext,kVZSignContext是用于使用函数decodeUnaryAAC或decodeSignedAAC来译码DeltaKeyValue三个分量x,y和z的环境。 kVXSignContext, kVYSignContext, kVZSignContext for using functions or decodeSignedAAC decodeUnaryAAC decodes DeltaKeyValue three components x, y and z environment.

kVXUnaryContext,kVYUnaryContext,kVZUnaryContext是用于使用函数decodeUnaryAAC来译码DeltaKeyValue三个分量x,y和z的环境。 kVXUnaryContext, kVYUnaryContext, kVZUnaryContext decodeUnaryAAC to use the function for decoding DeltaKeyValue three components x, y and z environment.

kVXContext,kVYContext,kVZContext是用于使用函数decodeedAAC来译码DeltaKeyValue三个分量x,y和z的环境。 kVXContext, kVYContext, kVZContext decodeedAAC to use the function for decoding DeltaKeyValue three components x, y and z environment.

在下文中,将参考本发明第三实施例来描述在包含在根据本发明第三实施例的量化器2000中的误差测量单元2050中所执行的用于测量一原始旋转差分数据值与一旋转差分数据值间的一误差的方法,以及用于测量一原始旋转差分值与一旋转差分值间的一误差的方法,上述旋转差分数据值是通过对分别包含在如图4A至4C所示的断点抽取器42、46、48中的误差计算器42b,46b以及48b中的已量化旋转差分值进行逆量化而复原的,上述旋转差分值是通过对已量化旋转差分值进行逆量化而复原的。 Hereinafter, with reference to the third embodiment of the present invention will be described in included in an original rotational differential value and a rotational differential data used to measure the quantizer according to a third embodiment of the present invention, the error measurement unit 2000 in 2050 performed The method of data values between an error, an error and a method for measuring an original rotational differential value and a difference value between the rotation, the rotation difference data values are shown by broken respectively included in FIG. 4A to 4C 42,46,48 point extractor in the error calculator 42b, 46b and 48b of the quantized rotational differential value inverse quantization and restored, the rotation differential value is quantized by inverse-quantizing the rotational differential value and the restored .

根据本发明,测量一原始旋转差分值和一复原旋转差分值间的一误差的方法也能被用来测量编码前的一原始定向内插器与通过译码一编码定向内插器生成的一定向内插器间的误差。 According to the present invention, a measurement method of the original rotational differential value and a restored rotational differential value between an error of the intra-coding can also be used to measure before an original orientation interpolator and a decoded by the encoder generates a certain directional interpolator interpolation error is between inward. 因此,一定向内插器的一动画路径,将由断点抽取器42、46或48从其抽取断点,以及用在误差测量单元250中的原始旋转差分值将被描述成与编码前的一定向内插器相对应。 Therefore, a certain inward interpolator animation path, 42, 46 or by the breakpoint extractor 48 extracts from the breakpoint, and used in the error measuring unit 250 in the original rotational differential value will be described as some of the previous coding interposer corresponding inward. 同样,由抽取断点所构成的一定向内插器的一动画路径以及用在误差测量单元2050中的复原旋转差分值将被描述成与译码定向内插器的关键字值数据相对应。 Similarly, certain inward interpolator constituted by the extracted break animation path and an error measuring unit 2050 is used in the return rotational differential value will be described as the inner decoder orientation interpolator keyword value corresponding to the data.

在编码一定向内插器的过程中,在量化期间在原始定向内插器和一复原定向内插器间产生一误差。 In certain inward interpolator encoding process, during quantization in the original orientation interpolator and a restored generating an error between the orientation interpolator. 在这里,原始定向内插器和复原定向内插器间的误差是由使用在原始旋转变换和复原旋转变换间的角度中的一差值定义。 Here, in the original directional interpolation and error recovery within the directional interpolator is defined by the use of inter-angle rotation transformation and restoration of the original rotation transformation of a difference between.

换句话说,假定包含在一定向内插器节点中的一关键字值数据与在一译码器中的其复原关键字值数据分别被称为 In other words, assuming that a value of data contained in a certain keyword inwardly interpolator node and its restored key value data in a decoder is referred to as, respectively,

And

其中 Among them

表示一旋转轴,以及θ表示一旋转角并满足θ∈[-π,π],通过旋转变换,在一三维空间中的一对象基于 Represents a rotation axis, and θ represents a rotation angle and satisfies θ∈ [-π, π], by rotation transformation, in a three-dimensional space of an object based on

And

被分别从一任意位置 Are respectively, from an arbitrary position

移动到一任意位置 Moved to an arbitrary position

以及从 And from

移动到一任意位置 Moved to an arbitrary position

一量化误差是 A quantization error is

And

间的差值并满足以四元数形式表示的 And to meet the difference between quaternion form represented by

And

如等式(57)中所示。 As shown in equation (57) in Fig.

当表示 When expressed

And

的四元数被称为Q和 Quaternion is called Q and

时,可导出下述等式。 When, the following equation can be derived.

Y=QXQ* …(58) Y = Q X Q * ... (58)

X=Q*YQ X = Q * Y Q

此时,AB表示四元数乘法,以及A*表示A的一复共轭。 At this point, A B represents quaternion multiplication and A * denotes a complex conjugate of A. 基于等式(57)和(58),可导出下述等式。 Based on Equation (57) and (58), the following equation can be derived.

此时,Qerror是表示在旋转变换方面 At this time, Qerror shows in terms of a rotational transformation

And

间的关系以及由下述等式来定义。 Relations among and defined by the following equation.

When

And

间的一差分旋转角被称为θerror时,使用一四元数变换公式以及等式(61)可获得θerror。 One difference between the angle of rotation is called when θerror, using a quaternion transformation formula and equation (61) can be obtained θerror.

此时,表示一内积操作。 At this point, * represents an inner product operation. 等式(61)定义在一预定时间瞬间在所有动画关键帧中发生的瞬时量化误差,因此,根据本发明的第三实施例,包括在量化器2000中的误差测量单元2050用等式(61)计算一原始定向内插器和一复原定向内插器间的一误差。 Instantaneous quantization error equation (61) is defined at a predetermined time instant occurring in all animation key frames, and therefore, according to the third embodiment of the present invention, comprises the quantizer 2000 error measuring unit 2050 by the equation (61 ) in the calculation of an original and a directional interpolation error recovery within a directional interpolator between.

另外,可用下述等式定义在预定时间瞬间(t)的一瞬时量化误差以便从等式(62)导出用于获得在一整个动画间隔中的一量化误差的一公式。 Further, by the following equation is defined at a predetermined time instant (t) of an instantaneous quantization error to derive from Equation (62) is used to obtain an entire animation interval of a quantization error of a formula.

通过大量地将公式(62)应用到整个关键帧间隔来导出一平均误差Em以及一最大误差Ep,在这期间,可使用一定向内插器来执行一动画。 By the large amount of the formula (62) to the entire keyframe interval to derive an average error Em and a maximum error Ep, during which some can be used to perform an inwardly interpolator animation.

此时,为获得平均误差Em,必须首先获得在间隔[ti-1,ti]中误差的部分和 At this point, in order to obtain the average error Em, you must first obtain in some interval [ti-1, ti] error and

如图27所示。 As shown in Figure 27. 在下文中,当在预定时间ti-1和ti与原始关键字数据相对应的译码关键字数据存在时,预定时间将分别被称为 Hereinafter, when the predetermined time ti-1 and ti original key data corresponding to the decoding key data is present, will be referred to, respectively, a predetermined time

And

与原始关键字值数据Qi-1和Qi相对应的译码的关键字值数据将被分别称为 The original key value data Qi Qi-1 and the corresponding decoding key value data will be referred to

And

由于在编码关键字数据期间生成的噪音,不可能直接计算一原始定向内插器的原始动画路径Qi′与一译码定向内插器的一动画路径 Because during the encoding key data generated noise, can not be calculated directly within an original directional insert the original animation path Qi 'and within a decoder directional interpolator a motion path

间的一误差,如图27所示。 Between an error, shown in Figure 27. 因此,间隔[ti-1,ti]必须被划分成三个子间隔 Accordingly, the interval [ti-1, ti] must be divided into three sub-interval

And

接着,使用等式(63)来获得在 Next, using equation (63) to obtain

And

的四元数值。 The four yuan value.

在等式(63)中,函数SLERP()被用来执行球性线性内插。 In equation (63), the function SLERP () is used to perform linear interpolation of the ball.

由于在一三维空间中一动画路径必然矢真的事实,如图27所示,间隔 Since a three-dimensional space a vector animation path necessarily true fact, shown in Figure 27, the spacer

必须被划分成两个子间隔 Must be divided into two sub-intervals

And

然后这两个子间隔必须彼此分开计算。 Then the two sub-interval must be calculated separately for each other. 此时,假定两个动画路径间,即一组Qi值与一组 At this point, it is assumed between the two animation path, that is, a set of values and a set Qi

值间的一距离在间隔 A distance value in the interval between

中的ti”被减小。使用下述等式来计算在 In ti "is reduced. Use the following equation to calculate the

And

间的瞬时误差。 Instantaneous error between.

另外,在间隔 Further, in the interval

中的ti“与 The ti "and

成比例,如等式(65)所示。 Proportional, as shown in equation (65) below.

用下述等式(66)来计算中间四元数值以及在ti”处的一瞬时误差。然而,用下述等式(67)获得在任意时间瞬间(t)的一瞬时误差。 By the following equation (66) to calculate intermediate quaternion values and an instantaneous error at ti "at. However, by the following equation (67) obtained at an arbitrary time instant (t) is an instantaneous error.

此时,以及然而,计算在任意时间瞬间(t)的一瞬时误差e(t)并不容易。 In this case, as well, however, calculated at an arbitrary time instant (t) of an instantaneous error e (t) is not easy. 因此,使用如近似法(68)所示的线性近似确定e(t)。 Thus, using such approximation formula (68) is a linear approximation to determine the e (t).

在间隔[ti-1,ti]中的误差的部分和 In portions and space [ti-1, ti] in error

以及在间隔 And in the interval

中的最大误差的一部分和 A portion of the maximum error

也能使用近似法(69)和(70)来获得。 Approximation method can also be used (69) and (70) to obtain.

近似法(69)能被重新整理如下。 Approximation (69) can be rearranged as follows.

最后,在一时间间隔[tmin,tmax]中的一平均Em以及一最大误差Ep用以下的近似法(71)计算。 Finally, at a time interval [tmin, tmax] in an average Em and a maximum error Ep by the following approximation (71) calculations.

因此,包括在如图4A至4C所示的断点抽取器42、46和48中的误差计算器42b、46b和48b分别使用等式(71)计算一原始定向内插器和一复原定向内插器中的一误差并能更精确地测量由于由于一量化误差所带来的在一四元数空间中图像矢真的程序。 Thus, including breakpoints decimator in FIG. 4A to 4C, 42, 46 and 48 in the error calculator 42b, 46b and 48b, respectively, using equation (71) calculating an original orientation interpolator and a restored orientation within interpolator in an error and can more accurately measure due to a quantization error due caused by a quaternion space is really an image vector program.

图28是与用于编码关键字值数据的一常规方法相比的根据本发明的用于编码关键字值数据的方法的性能的图形。 Figure 28 is compared with a conventional method for encoding key value data according to the graphics performance of the method for encoding key value data according to the present invention. 如图28所示,根据用于编码本发明的一定向内插器方法,当给定要求用于编码的一预定数量位时,与在常规的MPEG-4BIFS PMFC方法相比,图像矢真的程序能被相当大地降低。 28, according to a certain method for encoding inwardly interpolator of the present invention, when a given requirement for encoding a predetermined number of bits, than in the conventional MPEG-4BIFS PMFC method, the image really vector program can be considerably reduced.

本发明能按写在一计算机可读记录介质上的计算机可读代码来实现。 According to the present invention can be written on a computer-readable recording medium on a computer-readable code to implement. 在这里,计算机可读记录介质包括能由一计算机系统读取的任何种类的记录介质。 Here, a computer-readable recording medium includes any kind of readable recording medium by a computer system. 例如,该计算机可读记录介质可能包括一ROM、一RAM、一CD-ROM、一磁带、一软盘、一光数据存储器、载波(通过Internet传输)等等。 For example, the computer-readable recording medium may include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage, carrier etc. (Internet transmission through). 计算机可读记录介质能被分散到连接到网络上的计算机系统,以及一计算机能用一分散的方式读取该记录介质。 The computer-readable recording medium can be dispersed to be connected to the computer system on the network, and a computer can read a distributed manner the recording medium.

根据本发明,用于编码一定向内插器的方法和装置能以高效率编码动画关键字数据和关键字值数据同时保持高质量动画。 According to the present invention, a method for encoding a certain inwardly and apparatus interpolator at high efficiency encoding animation key data and key value data while maintaining high quality animation. 另外,根据本发明,通过利用测量动画数据路径间的一误差的方法来保持高质量动画是可能的,而且通过利用重新采样方法和本发明的断点抽取,也可能相当大地减小将被编码的关键字数据和关键字值数据量。 Further, according to the present invention, by using the measurement method of the animation data path between an error to maintain high quality animation is possible, and by using the resampling method and the break point extraction present invention may also be reduced considerably to be coded The key data and key value data volume.

另外,通过计算一旋转差分值,有可能提供具有一高压缩比的高质量动画,该旋转差分值能充分反映在与一定向内插器的关键字数据相对应的旋转变换中的冗余,并因此编码该定向内插器的关键字值数据。 Further, by calculating a rotational differential value, it is possible to provide high-quality animation with a high compression ratio, the rotation can be fully reflected in the difference value with a certain inward interpolator keyword data corresponding to the redundancy in rotational transformation, encoding the interposer and therefore the value of the data within the targeted keywords.

同时本发明已经明确表示和参考其优选实施例描述过,对本领域的普通技术人员来说在不脱离由附加权利要求限定的本发明的精神和范围的情况下可对形式和细节做出改变。 While the present invention has been specifically shown and with reference to the preferred embodiments described, one of ordinary skill in the art that without departing from the spirit and scope as defined by the appended claims the present invention changes may be made in form and detail.

Classifications
International ClassificationH03M7/40, H03M7/36, H04N7/26, H03M7/30, G06T13/20, G06F7/38, H03M7/48, G06F17/22, H04N7/36, H04N7/32, H04N7/24, G06T9/00
Cooperative ClassificationH04N19/70, H04N19/25, H04N19/13, H04N19/136, H04N19/14, H04N19/103, H04N19/126, H04N19/46, H04N19/12, H03M7/40, H03M7/30
European ClassificationH04N7/26A4C, H04N7/26J8, H04N7/26A4K, H04N7/26Y, H04N7/26A4V, H04N7/26A4Q2, H04N7/26A10S, H04N7/26A6C, H04N7/26A6C2
Legal Events
DateCodeEventDescription
10 Sep 2003C06Publication
26 Nov 2003C10Entry into substantive examination
27 Sep 2006C14Grant of patent or utility model
11 Jan 2017CF01