WO2015067741A1

WO2015067741A1 - Apparatus and method for active voice training

Info

Publication number: WO2015067741A1
Application number: PCT/EP2014/074016
Authority: WO
Inventors: Thierry GAUJARENGUES; Klaus LOHMANN
Original assignee: Soundev Holding Sa
Priority date: 2013-11-08
Filing date: 2014-11-07
Publication date: 2015-05-14
Also published as: LU92306B1

Abstract

An apparatus comprises a microphone capable of being placed essentially in front of a user's mouth for capturing the sound of the user's voice and converting said sound into an input electrical signal, a voice processing unit for processing said input electrical signal into an output electrical signal, at least two electromechanical vibration transducers for applying vibrations to the user's skull, which electromechanical vibration transducers are capable of converting said output electrical signal into vibrations. The apparatus is characterized in that the voice processing unit comprises an electronic circuit assembly having a tone control unit to alter the input electrical signal depending on the amplitude of the input electrical signal, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a treble portion (H) if the input electrical signal is above a defined speech intensity threshold (T) for a defined trigger time (t_i).

Description

APPARATUS AND METHOD FOR ACTIVE VOICE TRAINING Technical field

[0001 ] The present invention generally relates to speech and/or voice training and more particularly to an apparatus and a method to actively support the user's learning efforts.

Background Art

[0002] Although a number of solutions exist for allegedly or truly supporting speech and/or voice training, these solutions are generally subject to one or more drawbacks. Some integrate conventional headphones and thus sequester the user from the (sounds of the) surroundings and/or (at least partly) prevent communication with a helping person, such as with a coach or trainer. Others combine aerial sound transmission with bone conduction, which makes these solutions subject to the same disadvantages of acoustic user isolation. Furthermore, it has been noted that the use of bone conduction alone does not always yield the positive training results expected or claimed.

Technical problem

[0003] It is therefore an object of the present invention to provide an (alternative) apparatus and method which effectively allow to actively training speech and voice without isolating the user from the surrounding sounds and noises.

General Description of the Invention

[0004] In order to overcome (at least some of) the above-mentioned problems, the present invention proposes in one of its aspects an apparatus comprising a microphone capable of being placed essentially in front of a user's mouth for capturing the sound of the user's voice and converting said sound into an input electrical signal. It further comprises a voice processing unit for processing said input electrical signal into an output electrical signal. The apparatus also includes at least two electromechanical vibration transducers for applying vibrations to the user's skull, which electromechanical vibration transducers are capable of converting said output electrical signal into vibrations. The voice processing unit of the apparatus according to the invention comprises an electronic circuit assembly having a tone control unit to alter the input electrical signal depending on the amplitude of the input electrical signal, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a treble portion H if the input electrical signal is above a defined speech intensity threshold T for a defined trigger time ti .

[0005] In a further aspect, the invention also provides a voice processing unit for processing an input electrical signal from the voice of a user captured by a microphone into an output electrical signal for producing vibrations by at least two electromechanical vibration transducers. Such a voice processing unit comprises an electronic circuit assembly having a tone control unit to alter the input electrical signal depending on the amplitude of the input electrical signal, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a treble portion H if the input electrical signal is above a defined speech intensity threshold T for a defined trigger time ti .

[0006] In a still further aspect the present invention concerns the use of an apparatus or of a voice processing unit as described herein for dynamic, frequency-selective and active voice-training.

[0007] In yet another aspect, the invention proposes a method for dynamic, frequency-selective and active voice-training using the apparatus or the voice processing unit as described herein. In such a method a microphone is placed essentially in front of a user's mouth for capturing the sound of his voice and at least two electromechanical vibration transducers are placed against the user's skull, preferably against both temporal bones, more preferably the mastoid parts of the temporal bones.

[0008] Preferably, the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a bass portion B if the input electrical signal is below said threshold T for a defined holding time t₂.

[0009] In a preferred embodiment, the apparatus or the voice processing unit, therefore further comprises dynamic voice-detection preferably based on intensity, including specified hold-delays, hysteresis, and signal ramping. [0010] The apparatus or the voice processing unit presented herein preferably also comprises an automatic, preferably voice-coupled, equalizer section and/or high pass-rumble-protection.

[001 1 ] It is important to note that the apparatus of the present invention does not comprise conventional loudspeakers to be placed over the user's ears. Hence, the user is able to normally hear all the sounds of the surroundings. The user is therefore able to use the apparatus in all circumstances and does not need to take the apparatus off for communication with other persons.

[0012] The audio-vocal loop is the natural process by which everyone receives, analyzes, assimilates and continuously adjusts the information transmitted by the voice. To properly function, the audio-vocal loop is based on the capacity of auditory discrimination, phonological awareness and integration of rhythm that each individual must be able to perform effortlessly. These skills are far beyond the word and are necessary in any learning process.

[0013] For many reasons, this loop may be disrupted. By optimizing its operating capacity, the apparatus of the present invention significantly improves attention, concentration, verbal working memory, short-term memory, speech and pronunciation, and/or rhythm and balance.

[0014] The basic principle is founded on being closer to the sensory message through bone conduction. Indeed, when we make a sound, we vibrate our vocal cords. This vibration is transmitted primarily by bone conduction and secondarily by air to the inner ear. Thus in our ears, we hear our own voice without difficulty. This bone transmission is fundamental because it avoids loss and alteration of the message related to air and ambient noise. The apparatus of the invention thus allows to specifically amplifying the transmission through bone conduction.

[0015] The effect is an easier analysis and enhanced assimilation of the information the user emits. The user starting in an optimal self-listening position is simply more aware about what he says. The ear is not blocked by a loudspeaker, it continues to capture the ambient noise in the normal way. The individual can concentrate on his own voice without being isolated from external sounds, so he learns to discriminate relevant sounds from those which are not. In addition, he can normally hear another person and interact with it without removing the headset (parents, teachers, therapists, coaches, ...).

[0016] Even if bone conduction enhances self-listening, an important aspect is the quality of the message emitted by the voice. The present apparatus is equipped with a voice processing unit implementing a dynamic filter that is responsive to the intensity of the voice. The user hears his own voice by bone conduction but filtered with alternating contrast. This alternation will occur mainly on the attack words and long vowels. In fact, it has been found that "higher sounds" are more profoundly involved in the construction of language. By amplifying and thus promote their perception, the user will instantly and unconsciously significantly improve the quality and speed of his voice. His voice will progressively become richer in harmonics, smoother, warmer, more fluent and dynamic.

[0017] Moreover, the present alternative filter allows to raise the treble portion and optionally to correspondingly reduce the bass portion of the voice. It thus enhances the transmission of high frequency harmonics, which play a fundamental role in the cortical stimulation and revitalization.

[0018] In a particularly preferred embodiment, the voice processing unit (of the apparatus) splits the input electrical signal in a first dynamic voice detection signal and a second audio equalizer signal, the first dynamic voice detection signal being converted into a high voltage signal if the input electrical signal is above threshold T for said time ti and into a defined low voltage signal if the input electrical signal is below threshold T, preferably after a defined holding time t₂.

[0019] In the context of the present invention, the speech intensity threshold T is between -40 and -80 db (dbV at microphone-input) at a frequency of 1 kHz, preferably between -54 and -70 db (dbV at microphone-input) at a frequency of 1 kHz. In a further preferred embodiment, threshold T implements a hysteresis function: the threshold level (dbV) for raising the treble portion (upswitch) is preferably different (such as about 8-10 dbV - (higher)) from the threshold for raising the bass portion (backswitch).

[0020] The trigger time ti is chosen in order to allow for a fast capturing of the voice while preventing surrounding noise to trigger the capturing when the user is not speaking. In practice, trigger time ti is generally selected between 1 and 1000 ms, preferably between 5 and 500 ms, particularly preferably between 10 and 50 ms.

[0021 ] The treble portion H to be raised in intensity during speech by the voice processing unit is generally situated between 500 Hz and 20 kHz, preferably between 800 Hz and 15 kHz. The bass portion B in the context of the invention usually is between 5 Hz and 1 kHz, preferably between 20 Hz and 900 Hz, particularly preferably between 100 Hz and 800 Hz. In the context of the present invention there may be a certain overlap between bass and treble portions without departing from the object. However the intersection between both portions only represents a small fraction of the used spectra.

[0022] It is generally advantageous to keep the capturing of the microphone and processing by the voice processing unit even after the end of the last sound of the user in order to allow for appropriate handling of pausing between words or sentences. Hence, it is generally desirable to provide for a certain hysteresis in order to avoid too rapid switching between treble raising and bass raising. Therefore, in a still further preferred variant of the apparatus or of the voice processing unit, holding time t₂ is selected between 1 and 2000 ms, preferably between 2 and 1000 ms, particularly preferably between 5 and 500 ms, and in particular between 20 and 200 ms.

[0023] In practice, it may be useful to pre-amplify the input electrical signal provided by the microphone before its further processing in the voice processing unit.

[0024] In one embodiment of the apparatus presented and described herein, the apparatus takes the form of a headset for being placed on the user's head including a microphone and the at least two transducers, while the voice processing unit itself is integrated in a separate casing and connected to the headset either by wires or wirelessly.

[0025] Alternatively and preferably, the apparatus is in the form of a headset for being placed on the user's head with all the components, such as the microphone, the transducers and the voice processing unit integrated to the headset. Brief Description of the Drawing

[0026] Preferred embodiments of the invention will now be described, by way of example, with reference to the accompanying drawing in which:

Fig. 1 is a schematic diagram of a preferred embodiment of an apparatus of the present invention.

[0027] Further details and advantages of the present invention will be apparent from the following detailed description of several not limiting embodiments with reference to the attached drawing.

Description of Preferred Embodiments

[0028] Unlike usual voice amplifiers (like hearing aid amplifiers and products) a preferred embodiment of the apparatus is a dynamic, frequency-selective, active voice training unit, and therefore advantageously contains several special (psychoacoustic, audiophile) circuits:

• dynamic voice-detection (intensity), including specified hold-delays, hysteresis, and signal ramping

• automatic (voice-coupled) shelve equalizer section (fine-tuned for the training effect), additional high pass rumble protection

• special amplifier-section to drive a customized Lo-Z bone conductor headphone

• a small "one piece" ergonomic design, customized for the use of a special (lightweight) bone conductor headphone/headset microphone.

[0029] In a preferred variant, the apparatus or voice processing unit has the following features and functions, in which the pre-amplified microphone signal is split into two sections: a dynamic voice detection section and an audio-equalizer section.

[0030] Finally, the DC-voltage output of the dynamic voice detection effects the audio-equalizer section, and therefore the frequency content appearance at the apparatus' output. [0031] Unlike usual dynamic audio units such as compressor/I imiter/noise gate, the apparatus effects only the frequency response of the audio-signal itself, but not the average sound-level at 1 kHz.

[0032] The dynamic voice detection causes a defined "low" DC voltage output, if there is no or only a weak signal (below defined threshold) present at the microphone.

[0033] The dynamic voice detection causes a defined "high" DC voltage output, since the user is reaching a defined speech intensity threshold (such as -58 dbV at microphone-input) and can hold this intensity above threshold for a defined time (e.g. 25 ms).

[0034] In case the user's speech intensity drops below the threshold (such as -66 dbV at microphone-input because of hysteresis), and after a defined holding time (such as 70 ms), the output of the dynamic voice detection becomes "low" again.

[0035] The DC output of the dynamic voice detection is thereafter "ramped" by a specified time, both, up and down, such as from 10 to 20 ms, e.g. about 15 ms.

[0036] The "ramped" DC output of the dynamic voice detection is preferably forwarded to voltage controlled resistors, which effect the audio-equalizer section in a specified manner, according to the training effect and psychoacoustic needs.

[0037] The resulting equalized (mono) audio signal is forwarded to a two-channel headphone amplifier with separate Lo-Z outputs (e.g. 10 Ohms) for left and right bone conduction.

[0038] Unlike usual headphones, the present apparatus will preferably be placed at the mastoid bone left and right of the user's head. Therefore, there is no isolation of the ears, and the user can continue to perceive acoustic signals from the surroundings.

[0039] In a preferred embodiment of the invention, the apparatus is thus a small and lightweight bone conduction headset combined to a voice processing unit, either as a separate clip-box or directly affixed to or integrated into the headset, preferably powered by 3.7 V Li-ion battery with USB charging. [0040] The user gets the training effect by listening to his own speech. The audio- path is fully analog. The headset microphone position will most preferably be in front of the user's mouth, to reduce the risk for feedback.

[0041 ] The utilization of the present apparatus is easy: it is sufficient to appropriately place the apparatus such that the microphone is in front of the mouth and the transducers against both temporal bones and to normally speak into the microphone.

[0042] The apparatus is very versatile and may be used

- alone: reading aloud, singing, rehearsing, pronunciation and expression exercises, repetition exercises,

- with parents or any other helping person: reading, homework

- with a professional (therapist, teacher, coach),

- in everyday conversation, oral presentation, discussions.

[0043] To be effective, the method making use of the present apparatus (or voice processing unit) is generally practiced in sessions, each of which should preferably last between about 20 minutes and 1 .5 hours a day. The sessions should be repeated during several weeks, preferably however with so-called "integration" breaks.

[0044] Fig. 1 is a schematic diagram of a preferred embodiment of the apparatus according to the invention comprising a microphone 10 providing an input electrical signal (Mic Input), a voice processing unit 20, as well as two transducers 30 (Bone Conduction Headphone). The reference numbers between brackets in Fig. 1 refer to the paragraph number in which the relevant feature is further described.

[0045] The input electrical signal is generally preamplified by preamplifier 201 . The preamplified signal is then split into a (first) dynamic voice detection signal and a (second) audio equalizer signal.

[0046] The dynamic voice detection signal is processed by converting it into a high voltage signal 234 if the input electrical signal is above threshold 231 (T) for a trigger time 232 (ti) and into a defined low voltage signal 234 if the input electrical signal is below threshold 231 (T), preferably for a defined holding time 233 (t₂). The DC output 234 of the dynamic voice detection is thereafter ramped 235 by a specified time, both, up and down, such as for 15 ms. The ramped DC output of the dynamic voice detection signal is thereafter sent to voltage controlled resistors 236 (VCR), which effect the audio-equalizer section in a predetermined manner, either through treble tone control unit 240 or bass tone control unit 260. Buffer 250 merely is a pull up amplifier. The resulting equalized (mono) audio signal is forwarded to a two-channel headphone amplifier 270 with separate Lo-Z outputs for left and right bone conduction transducers 30.

Legend:

10 Microphone

20 Voice Processing Unit

201 Preamplifier

210 Volume control

220 High pass filter

230 Dynamic voice detection

231 Threshold T

232 Trigger time ti

233 Hold time t₂

234 DC output/Hysteresis

235 Ramping

236 Voltage controlled resistor(s)

240 Tone control: treble

250 Buffer

260 Tone control: bass

270 Amplifier(s)

30 Transducer(s)

Claims

1 . An apparatus comprising a microphone capable of being placed essentially in front of a user's mouth for capturing the sound of the user's voice and converting said sound into an input electrical signal, a voice processing unit for processing said input electrical signal into an output electrical signal, at least two electromechanical vibration transducers for applying vibrations to the user's skull, which electromechanical vibration transducers are capable of converting said output electrical signal into vibrations, characterized in that the voice processing unit comprises an electronic circuit assembly having a tone control unit to alter the input electrical signal depending on the amplitude of the input electrical signal, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a treble portion (H) if the input electrical signal is above a defined speech intensity threshold (T) for a defined trigger time (ti).

2. A voice processing unit for processing an input electrical signal from the voice of a user captured by a microphone into an output electrical signal for producing vibrations by at least two electromechanical vibration transducers, characterized in that the voice processing unit comprises an electronic circuit assembly having a tone control unit to alter the input electrical signal depending on the amplitude of the input electrical signal, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a treble portion (H) if the input electrical signal is above a defined speech intensity threshold (T) for a defined trigger time (ti).

3. The apparatus as claimed in claim 1 or the voice processing unit as claimed in claim 2, wherein the tone control unit alters the input electrical signal into said output electrical signal by applying a differential raising of a bass portion (B) if the input electrical signal is below said threshold (T) for a defined holding time (t₂).

4. The apparatus as claimed in claim 1 or 3 or the voice processing unit as claimed in claim 2 or 3, further comprising dynamic voice-detection preferably based on intensity, including specified hold-delays and signal ramping.

5. The apparatus as claimed in claim 1 or 3 to 4 or the voice processing unit as claimed in claim 2 to 4, further comprising an automatic, preferably voice- coupled, shelve equalizer section and/or high pass-rumble-protection.

6. The apparatus as claimed in claim 1 or 3 to 5 or the voice processing unit as claimed in claim 2 to 5, wherein the voice processing unit splits the input electrical signal in a first dynamic voice detection signal and a second audio equalizer signal, the first dynamic voice detection signal being converted into a high voltage signal if the input electrical signal is above threshold (T) for said time (ti) and into a defined low voltage signal if the input electrical signal is below threshold (T), preferably a defined holding time (t₂).

7. The apparatus as claimed in claim 1 or 3 to 6 or the voice processing unit as claimed in claim 2 to 6, wherein speech intensity threshold (T) is between -40 and -80 dbV at microphone input at a frequency of 1 kHz, preferably between -54 and -70 dbV at microphone input at a frequency of 1 kHz, trigger time (ti) is between 1 and 1000 ms, preferably between 5 and 500, particularly preferably between 10 and 50 ms, and/or treble portion (H) is between 500 Hz and 20 kHz, preferably between 800 Hz and 15 kHz.

8. The apparatus as claimed in claim 1 or 3 to 7 or the voice processing unit as claimed in claim 2 to 7, wherein holding time (t₂) is between 1 and 1000 ms, preferably between 5 and 500 ms, particularly preferably between 20 and 200 ms, and/or bass portion (B) is between 5 Hz and 1 kHz, preferably between 20 Hz and 900 Hz, particularly preferably between 100 and 800 Hz.

9. The apparatus as claimed in claim 1 or 3 to 8 or the voice processing unit as claimed in claim 2 to 8, wherein the input electrical signal is pre-amplified.

10. The apparatus as claimed in any one of claims 1 or 3 to 9, wherein: the apparatus is in the form of a headset for being placed on the user's head with the voice processing unit integrated in a separate casing and connected to the headset by wires or wirelessly.

1 1 . The apparatus as claimed in any one of claims 1 or 3 to 9, wherein: the apparatus is in the form of a headset for being placed on the user's head with the voice processing unit integrated to the headset.

12. A method for dynamic, frequency-selective and active voice training wherein an apparatus as claimed in any of claims 1 or 3 to 1 1 is used by placing the microphone essentially in front of a user's mouth for capturing the sound of his voice and by placing the at least two electromechanical vibration transducers against the user's skull, preferably against both temporal bones, more preferably the mastoid parts of the temporal bones.

13. Use of an apparatus as claimed in claims 1 or 3 to 1 1 or of a voice processing unit as claimed in claims 2 to 9 for dynamic, frequency-selective and active voice training.