WO1991000671A1

WO1991000671A1 - Signal processing system

Info

Publication number: WO1991000671A1
Application number: PCT/EP1990/000989
Authority: WO
Inventors: Carsten Herpel; Dietmar Hepper
Original assignee: Deutsche Thomson-Brandt Gmbh
Priority date: 1989-06-30
Filing date: 1990-06-22
Publication date: 1991-01-10
Also published as: CN1048477A; DE3926154A1; AU5855790A

Abstract

A signal processing system for digital signals comprises a quantizer with a non-linear characteristic curve. When sections of a digital signal are coded in different ways, e.g., by interframe and interframe coding, the quality of the decoded signal may vary from one section to another. The system disclosed makes it possible to obtain a decoded signal of uniform quality as well as search mode. This is achieved by using a quantizer characteristic curve for the signal section consisting of a linear portion which is shifted into the field of the characteristic curve in function of the data rate of a preceding signal section, followed by a steadily increasing portion, and by using a virtual buffer memory. The system is suitable for digital signal processing in data-reducing coding.

Description

Signal processing system

The invention relates to a signal processing system for digital signals.

When using digital storage media, e.g. Moving image on CD (compact disc) or digital video magnetic tape recorder, it is desirable that such modes of operation as "search forward" or "search backward" are also possible, albeit with reduced image quality. Because in this case information from the previous image or the previous images may be missing, information should be stored at certain intervals that can be decoded without the information from directly preceding images, e.g. the information for a complete full screen. You can then e.g. access directly in the search.

In normal operation, the images between the images stored as complete images (this corresponds to an "intra-frame" coding) are only stored in a type of coding that codes the changes in the current image compared to the previous one (this corresponds to a " interframe "coding). In order to achieve the same subjective picture quality in intra- and interframe-coded pictures, more data must be stored for intraframe-coded pictures than for interframe-coded pictures. In order to compensate for a time-varying data rate of a coded digital signal compared to a storage medium with a time-constant write or read data rate, a buffer memory is normally provided in the encoder or decoder. Data reduction in the encoder is achieved, for example, by quantizing the data. The fill level of the buffer memory can be regulated by controlling the corresponding quantization characteristic.

From the documents mentioned below, which call a coding method for video conference transmission and videophone, it is known that the quantization curve depends linearly on the fill level of the buffer memory and that in the case of a scene cut, or if the picture content between picture n changes greatly -1 and n, the coding for the picture n + 1 omitted during recording, the picture n with the double data rate intraframe-coded (otherwise interframe) and output as picture n + 1 during playback, and the previous picture n is replaced by repeating the previous image n-1. This discontinuity is covered by the temporal masking function of the eye. In principle, the data rate in the event of a scene change remains unchanged.

CCITT Study Group XV, Document No. 446, 1988, "Description of Reference Model 7" (RM7) CCITT Study Group XV, Temporary Document 10, March 1989, "Reference Model 8 and Action Points" (RM8)

The RM8 codec (codec means: encoder and decoder) and the previous ones (RM1 - RM7) have the disadvantage that the subjectively perceived image quality changes between interframe and intraframe encoded images, direct access only to images after one Scene changes are restricted and practically no search is possible if there is no scene change. The invention is based on the object of specifying a signal processing system for digital signals, for example a codec, which offers a subjectively uniform image quality when changing scenes and operating modes such as "search run forward" or "search run backward" when using digital storage media.

This object is achieved by the features specified in claim 1. Advantageous developments of the invention are described in the subclaims.

In coding methods in which the coded video signal is stored digitally on a medium, intraframe-coded pictures can be inserted into the sequence of interframe-coded pictures at periodic intervals, not only when there are scene changes. Unless there is a scene change, the temporal masking function of the eye cannot be used here.

One possible solution is to use a very large buffer memory, which can process the increased data rate that arises from the intraframe coding of images.

However, the size of the buffer memory has a direct influence on the minimal decoding delay. The decoder buffer memory must namely have at least the size of the coder buffer memory. When the normal playback mode of operation begins, for example after switching on the device or after a search, the decoder buffer memory is first filled with a certain amount of data before an image becomes visible at the output of the decoder. The larger the buffer memory is, the greater the decoding delay is. A buffer memory which allows a meaningful coding with periodic intra-frame pictures should be in the order of magnitude of 300 kbit with a data rate of 1.15 Mbit / s.

In addition, with a simple increase in the buffer memory, there is still no control over the data rate distribution between interframe and intraframe-coded pictures.

A solution to these problems is provided by a special buffer memory controller that uses the principle of a virtual buffer memory. The dependency of the quantizer step size on the fill level of the physical buffer memory in the encoder is partially decoupled.

The virtual buffer memory represents the encoded but not transmitted amount of data. The fill level of this virtual buffer memory actually controls the quantizer step size.

In the following, "macro block" is understood to mean a block which, e.g. consists of four 8 x 8 DCT luminance blocks arranged in a square and the associated two 8 x 8 DCT chrominance blocks (U and V). The chrominance here has half the horizontal and vertical resolution of the luminance. DCT means: discrete cosine transform.

An exemplary embodiment of the invention is explained below with reference to the drawings. These show in 1 is a block diagram of an encoder according to the invention,

Fig. 2 shows a known function between one

Quantizer step size and the fill level of the buffer memory,

3 shows a function according to the invention between the quantizer step size and the fill level of the virtual buffer memory,

4 shows a function for determining a first of the parameters for the function from FIG. 3,

5 shows a function for determining a second of the parameters for the function from FIG. 3,

Fig. 6 shows a function at low

Buffer tank fill level with parameters changed compared to FIG. 3,

Fig. 7 shows a function with increased

Buffer tank fill level with changed parameters compared to Fig. 6,

Fig. 8 shows a function with further increased

Buffer tank fill level with changed parameters compared to FIG. 7,

9 shows a time comparison of the

Encoder input and encoder output data 1 shows a coder, the input video signal of which is fed to a circuit 10, which switches between interframe and intraframe coding, and a block comparator 172, which checks to what extent an estimated block of pixels with corresponds to the corresponding original block of pixels. A first output signal of circuit 10 is DCT-transformed in circuit 111, scanned in circuit 121, weighted in circuit 131, quantized in circuit 14, first output signals in circuit 132 inversely weighted, scanned in circuit 122, in circuit 112 inversely DCT-transformed and in circuit 19 together with a first output signal of a filter circuit

171 and controlled by a switching signal from circuit 10, again assembled into a block of pixels. The pixel blocks from circuit 19 are buffer-stored in an image memory 18 and fed to the block comparator 172 at the appropriate time. The block comparator

172 calculates motion vectors encoded in variable word length circuit 152.

The image signal provided at the output of the block comparator 172 in accordance with the determined motion vector is fed to the filter circuit 171. The possibly filtered signal is fed to circuit 10 and circuit 19.

The first output signals of the quantizer 14 are also fed to a circuit 151, where they (ie DCT coefficients and their addresses) are coded with a variable word length and are forwarded to the video multiplexer with a physical buffer memory 16. The output signals of the circuit 152 are also fed to this circuit 16. First output signals of the circuit 16 are then transmitted or stored. The circuit 16 receives, as further input signals, the output signal of the circuit 10 with the intra / inter decision, a second output signal of the Circuit 171 with the filter on / off decision and a second output signal of the quantizer 14, which indicates the respective quantizer step size, in order to encode the data from circuit 151 accordingly. The quantizer 14 is in turn controlled by a second output signal from the circuit 16. The circuit 16 and the quantizer 14 with circuit 151 form a control loop.

2 shows a linear function of the quantizer step width 21 as a function of the fill level 22 of the buffer memory in the circuit 16. The quantizer step width 21 can be between QTii.n and 0 * -max. The fill level 22 can be between zero and Fmax.

Level 22, Q ^Ä uantisiererschrittweite 21, Q ax, Qmm. and

F Fmax hhaabbeenn in FIGS. 3 to 8 the corresponding meaning as in FIG. 2

Fig. 3 shows a function of the quantizer step size 31, which consists of a linear part 33 with the equation

0 = 0. + C, x F and consists of a non-linear part 34 with ^x τnιn 1 of the equation Q = C ₂ x F x F + C ₃ F + C ₄ , as a function of the fill level 32 of a virtual buffer memory. The connection point 35 between the linear part 33 and the non-linear part 34 of the function has the coordinates

(F, Q). C. to C. are constants that can be varied by the coder. c c '1 4

The virtual buffer memory can be thought of as being included in circuit 16.

The virtual buffer memory becomes a certain number

B of bits per macroblock or frame are intended to be transferred to the storage medium. B has the value B. for intraframe-coded pictures and B, for interframe-coded pictures. l d

Bι. can e.g. B. correspond to the Ni. Times the average bit rate Ba bits per macroblock or frame. N. is for example chooses to create a subjectively balanced picture quality between intra- and inter-frame coded pictures.

With a net video bit rate of 1.15 Mbit / s to be recorded, a frame rate of 25 Hz and a factor N 1 = 3, e.g. 138 kbit for an intraframe coded

Image needed. As a result of the imaginary transmitted data rate B. bits per macroblock, neither the fill level of the virtual buffer memory will increase significantly during the intraframe coding of an image, nor the quantizer step width.

For the bit rate Bd, only a value smaller than Ba is permitted ^.

Bd, is defined as

where Nf_.r is the period of the intraframe-encoded pictures within the interframe-encoded.

With a net video bit rate to be recorded of 1.15 Mbit / s, a frame rate of 25 Hz, N_ = 30 and a use of 138 kbit for an intraframe-coded picture, e.g. on average 46 kbit for each picture and approx. 42.8 kbit for an interframe-coded picture. The virtual buffer memory has e.g. B. a size of

Fmax = 92 kbit. Then the physical buffer memory must have a minimum size of

Fp.hmax = Fmax + 138 kbit - 46 kbit = 184 kbit.

The control loop between quantizer 14 with circuit 151 and the buffer memory in circuit 16 is advantageously designed such that the fill level 32 of the buffer memory after _f -1 interframe-coded pictures is so low that the data of the next intraframe-coded picture is stored in the virtual buffer memory can be. The linear part 33 of the function of the quantizer step width 31 advantageously enables the buffer memory to reach a certain fill level without a significant change in the quantizer step width and thus in the image quality. The advantageously progressively increasing part 34 still protects the buffer memory from overflowing. Basically, the non-linear part 34 of the characteristic curve forms a quadratic function.

The minimum Q ^Ä uantisiererschrittweite Qmm. has, for example, the value

4/2048 and the maximum quantizer step size Q is 64/2048.

By using a virtual buffer memory, the encoder can advantageously use the same characteristic curve for intra- and interframe-coded pictures, a good picture quality being achievable.

The function of the quantizer step size 33 and 34 can advantageously be adapted to the characteristics of the respective image content by varying the constants C. to C. and thus shifting the connection point 35. In order to determine the values from C ₁ to C. for the current image k, the mean buffer storage level F k _ ^{and d} i ^e mean

Quantizer step size Qa, .Λ -, κ_x for the image k - kx measured in each case.

It can be advantageous to achieve a further increase

Image quality thereby an optimal characteristic for the

Intra- and inter-mode are formed.

In intra mode, kx = N _f , otherwise kx = 1 or kx = 2 if the last picture k-1 was encoded in intra mode.

4 shows the coordinates (F,, Q) of the connection point 35 for image k from the average virtual buffer fill level F _t - kx of ^a previous image k - kx, which was coded in the same mode. The value Q c advantageously remains constant.

5 shows a function 56 of an additional offset Q, for the quantizer step width 31 of the current image k as a function of the average quantizer step width Q, _, of the previous image k - kx.

The aim of an adaptation of the quantizer step size function 33 and 34 is to allow a greater variation of the buffer memory fill level F without a substantial change in the quantizer step size Q if an image is likely to have a small quantizer step size that increases the image quality due to the characteristics of the previous one Q, can be encoded, a, Λ

If the virtual buffer storage level F nevertheless rises significantly, the quantizer step size Q must be increased, advantageously by shortening the linear part 33 of the function.

6 shows a quantizer step width function aimed for by the encoder.

FIG. 7 shows a function changed from FIG. 6 with a shortened linear part 73 with an increased mean buffer storage level F, _,.

If, owing to the characteristics of the image k - kx, for example due to fine image details in the entire image, the buffer level is already significantly increased, an offset Q is additionally added to the quantizer step width Q and to the minimum quantizer step width ζ> ^. If the buffer storage level F has already increased for a picture and so that the quantization is coarser, a very fine quantization is advantageously not permitted in this image.

FIG. 8 shows a function according to FIG. 7 with an additional offset Qo for the quantizer step • width Q ^Ä c or Q ^Ä mm. , Due to the offset Qo, the function advantageously remains non-linear and limits variations in the quantizer step width Q in the image.

Because the permissible fill level F - ^ of the physical buffer memory is greater than the permissible fill level F of the - ^ max virtual buffer memory, it is ensured on the one hand that the physical buffer memory does not overflow in the event of a sudden, unpredictably high increase in data rates, and on the other hand the quantizer step size becomes Q when the virtual buffer fill level rises to near

Fmax early ³ high-erected and thus a clearly visible jump in image quality avoided. In addition, an unpredictable strong data rate increase will advantageously only occur when there are scene changes, where it is tolerated by the viewer (masking function of the eye).

The maximum buffer memory delay Tmax results in

Tmax = Fph. ax / 'R. For Fp.hmax = 184 kbit and R = 1.15 Mbit /' s the maximum buffer memory is therefore ^ eicher-deceleration_ T _maχ

0.16 s. This corresponds to the time for 4 pictures at a frame rate of 25 Hz.

9 shows an example of the approximate temporal relationship between the image data at the encoder input and the coded data at the input of the buffer memory in circuit 16.

Intra-frame coded pictures are labeled i, interframe-coded pictures are labeled d. The distribution of the data rates between i- and d-pictures must be such that after N pictures, two i-pictures face each other again.

Claims

Patent claims

1. Signal processing system for digital signals, in particular video and / or audio signals, which are digitally stored and / or stored on a storage medium, characterized in that the digital signals consist of sequential sequences, each of which start with an intraframe-coded signal section followed by a number of interframe-coded signal sections, the mutual time intervals of the intraframe-coded signal sections from the signal content, for example Scene changes are essentially independent and that a quantization with an adaptive control characteristic is included.

2. System according to claim 1, characterized in that the quantization control characteristic (33 and 34), hereinafter called control characteristic, for each signal section by the fill level (F) of a buffer (in 16) in its step size (Q) controllable quantizer is non-linear.

3. System according to claim 1 or 2, characterized in that the control characteristic (33 and 34) from an im. essential linear part (33) and consists of an adjoining progressively increasing (34) part.

4. System according to claim 1, 2 or 3, characterized gekennzeich¬ net that the buffer (in 16) contains a virtual buffer (in 16), which has a smaller storage capacity than the physical.

5. System according to claim 4, characterized in that select at a low and at a medium data rate After a preceding signal section or several preceding signal sections for the current signal section, the minimum quantizer step width (Qmin) given by the control characteristic curve remains essentially the same and the connection point (35) with the coordinates (Fc, Qc) between the substantially linear (33) and the progressively increasing (34) part of the control characteristic (33 and 34) with an essentially identical quantizer step size (Qc) with increasing data rate from a higher to a lower fill level (Fc) of the virtual buffer (in FIG. 16 ) moves towards (to the left in Fig. 3).

6. System according to claim 4 or 5, characterized in that at a high data rate during a preceding signal section or several previous signal sections for the current signal section the substantially linear part (33) of the control characteristic (33 and 34 ) with the connection point (35) to the progressively increasing (34) part of the control characteristic (33 and 34) with increasing data rate essentially parallel in the direction of higher quantizer increment (Q) (up in Fig. 3).

7. System according to one or more of claims 2 to 6, characterized in that a different control characteristic curve (33 and 34) is formed for interframe or intraframe coded signals.

8. System according to one or more of claims 2 to 7, characterized in that in signal sections which are intraframe-coded, an increased amount of data compared to that in inter-frame-coded data sections which the multiplied by a constant factor corresponds to the amount of data in a signal section from which the virtual buffer (in FIG. 16) is read out.

9. A code circuit for a system according to one or more of the preceding claims 1 to 8, which is provided with a first circuit which interframe-coded first signal sections and intra-frame-coded second signal sections, with a downstream quantizer, which by the during a the previous signal section or several previous signal sections ^' level (F) of a buffer in its step size (Q) can be controlled in accordance with the control characteristic (33 and 34) and with a downstream second circuit which determines the respective quantizer step size ( Q) coded and added to the signal content of the buffer.

10. Decoder circuit for a system according to one or more of claims 1 to 8, which can decode the stored output signals of the code circuit with normal and / or with increased speed and from the stored output signals of the code circuit in a first circuit the quantizer step width (Q) is decoded and, in a second circuit, with the aid of this quantizer step width (Q), the original signal sections are decoded in accordance with the interframe or intraframe before coding, with decoding at an increased speed (search run) only the intraframe-coded signal parts are decoded with the corresponding quantizer step size (Q).