WO2015144715A1

WO2015144715A1 - Electronic circuit for osteophonic headset

Info

Publication number: WO2015144715A1
Application number: PCT/EP2015/056276
Authority: WO
Inventors: Patrick Jean François ROBUCHON; Eric Bernard Jacques CLOWEZ; Julie Marie Anne ROSIER
Original assignee: Elno
Priority date: 2014-03-25
Filing date: 2015-03-24
Publication date: 2015-10-01
Also published as: FR3019423B1; FR3019423A1

Abstract

The invention relates to an electronic circuit (10) for an osteophonic headset comprising: an osteophonic microphone (20) able to detect a bone-conducted vibratory signal emitted by the user; a sound emitting module (18) for transmitting a signal to the user; and a module (16) for detecting vocal activity able to operate in a first mode and a second mode, the module (16) for detecting vocal activity being able to generate an emission signal corresponding to the voice of the user in the first mode and the module (16) for detecting vocal activity being able to generate a reception signal in order to supply the sound emitting module (18) in the second mode. The module (16) for detecting vocal activity is able to switch between the first mode and the second mode depending on the vibratory signal detected by the osteophonic microphone (20).

Description

Electronic circuit for osteophonic headband

The present invention relates to an electronic circuit for osteophonic headband and an osteophonic headband comprising such a circuit.

In the field of telecommunications, the problem of increasing the communication rate is recurrent. During a conversation between two mobile phones, it is desirable to detect the moments of silence where one or both interlocutors do not speak and avoid coding these moments of silence to maximize the flow.

For this, it is known from the article by Ramirez, J. et al, whose title is "Voice Activity Detection, Fundamentals and Speech Recognition System Robustness" and published in 2007 in the book by M. Grimm and K. Kroschel entitled " Robust Speech Recognition and Understanding "to detect the presence or absence of human voices in an audio signal using a technique called voice activity detection. Voice activity detection is usually referred to by the acronym VAD for "voice activity detection".

However, the speech activity detection modules capable of implementing the speech activity detection are generally not robust to the ambient noise, ie the speech activity detection modules are triggered inadvertently with ambient noise while the audio signal does not include a human voice.

There is therefore a need for an osteophonic band circuit for a better triggering of voice activity detection.

For this purpose, the subject of the invention is an electronic circuit for osteophonic headband comprising an osteophonic microphone capable of being positioned on a lateral flank of the head of a user and suitable for detecting a vibratory signal of bone conduction emitted by the user. , a sound transmission module to the user of a signal, and a voice activity detection module adapted to operate in a first mode and in a second mode, the voice activity detection module being able to generate a transmission signal corresponding to the voice of the user in the first mode and the voice activity detection module being adapted to generate a reception signal for supplying the sound emission module in the second mode. The voice activity detection module is adapted to switch between the first mode and the second mode depending on the vibration signal detected by the osteophonic microphone.

According to particular embodiments, the electronic circuit comprises one or more of the following characteristics, taken individually or in any technically possible combination:

the voice activity detection module comprises a first input adapted to receive an input signal comprising a voice signal from the user and a first amplifier of the input signal, the transmission signal generated in the first mode corresponding to the signal amplified by the first amplifier.

the voice activity detection module comprises a second input capable of receiving a communication signal and a second amplifier of the communication signal, the reception signal generated in the second mode corresponding to the signal amplified by the second amplifier.

the voice activity detection module also comprises a first input signal processing unit able to perform a processing on the input signal to obtain a processed input signal, the second amplifier having a gain varying according to the the processed input signal, and a second communication signal processing unit adapted to perform a processing on the communication signal to obtain a processed communication signal, the first amplifier having a gain varying according to the processed communication signal.

the first signal processing unit comprises a first sub-unit for converting an audio signal into a continuous component capable of generating a first signal converted by filtering with a first time constant, a second conversion sub-unit of an audio signal in a continuous component adapted to generate a second converted signal by filtering the audio signal with a second time constant, the first time constant being different from the second time constant, and a comparison subunit suitable to compare the two converted signals, the processed input signal being equal to the first converted signal when the comparison subunit detects a difference between the two converted signals.

the second signal processing unit comprises a first sub-unit for converting an audio signal into a continuous component capable of generating a first signal converted by filtering with a first time constant, a second conversion sub-unit of an audio signal in a continuous component adapted to generate a second converted signal by filtering the audio signal with a second time constant, the first time constant being different from the second time constant, and a comparison subunit suitable to compare the two converted signals, the processed communication signal being equal to the first converted signal when the comparison subunit detects a difference between the two converted signals,

the electronic circuit further comprising an aerial microphone capable of detecting an air vibration signal, an environment microphone capable of detecting an air vibration signal corresponding to the noise of the environment, and an signal generation module capable of generating the input signal on the basis of the vibration signal from the overhead microphone, the vibration signal from the environment microphone and the vibratory signal detected by the osteophonic microphone.

the generation module comprises two subtraction units, the first subtraction unit being able to generate a first output signal equal to the difference between the vibration signal coming from the overhead microphone and the vibration signal coming from the environment microphone, and the second subtraction unit being able to generate a second output signal equal to the difference between the first signal and the vibratory signal detected by the osteophonic microphone.

the generation module comprises three subtraction units, the first subtraction unit being able to generate a first output signal equal to the difference between the vibration signal coming from the overhead microphone and the vibration signal coming from the environment microphone, the third subtraction unit being able to generate a third output signal equal to the difference between the vibrating signal detected by the osteophonic microphone and the reception signal and the second subtraction unit being able to generate a second output signal equal to the difference between the first output signal and the third output signal.

In addition, the invention also relates to an osteophonic strip comprising an electronic circuit as defined above.

Other features and advantages of the invention will appear on reading the following description of embodiments of the invention, given by way of example only and with reference to the drawings which are:

- Figure 1, a schematic view of an electronic circuit for osteophonic band according to a first embodiment;

FIG. 2 is a diagrammatic view of an exemplary positioning of microphones belonging to the electrical circuit of FIG. 1;

FIG. 3 is a schematic view of the elements of a unit of the electronic circuit of FIG. 1 comprising a comparison sub-unit with a threshold;

FIGS. 4 and 5, two curves illustrating the variation of a signal respectively at the input and at the output of the comparison sub-unit at a threshold;

FIGS. 6 and 7, two curves illustrating the variation of a signal respectively at the input and at the output of the unit of FIG. 2 when the signal comprises the human voice;

FIGS. 8 and 9, two curves illustrating the variation of a signal respectively at the input and at the output of the unit of FIG. 2 when the signal comprises only ambient noise, and

- Figure 10, a schematic view of an electronic circuit for osteophonic band according to a second embodiment. An electronic circuit 10 for osteophonic headband is illustrated in FIG. The osteophonic headband is intended to be worn by a user.

The electronic circuit 10 comprises an acquisition module 12, a signal generation module 14, a voice activity detection module 16 and a sound transmission module 18 towards the user.

The acquisition module 12 comprises an osteophonic microphone 20, an overhead microphone 22 and an environment microphone 24.

The osteophonic microphone 20 is adapted to be positioned on a lateral flank of the user's head and adapted to detect a vibratory signal of bone conduction emitted by the user.

The overhead microphone 22 is capable of detecting a signal of vibration of the air. Such a signal comes in particular from the ambient noise and / or a voice broadcast by the user.

The environment microphone 24 is able to detect an air vibration signal corresponding to the noise of the environment. In particular, even if the user emits voice sounds, the environment microphone 24 is not able to detect the vibration signal unlike the air microphone 22. This difference is obtained in practice with an appropriate positioning of the microphone. 24 to a position distinct from the position chosen for the aerial microphone 22. Typically, as shown in FIG. 2, the overhead microphone 22 is oriented towards a voice diffusion zone delimited in FIG. 2 by dashed lines. while the environment microphone 24 is oriented towards a noise diffusion zone delimited in FIG. 2 by solid lines. Thus, the overhead microphone 22 and the environment microphone 24 are located in opposition, one capturing the voice and the other the noise. More specifically, in the case of Figure 2, the two air microphones 22 and environment 24 are 180 ⁰ from each other.

Preferably, the overhead microphone 22 and the environment microphone 24 are identical microphones.

The generation module 14 is able to generate a signal based on the vibration signal from the overhead microphone 22, the vibration signal from the environment microphone 24 and the vibratory signal detected by the osteophonic microphone 20.

The generation module 14 comprises three inputs 14E1, 14E2, 14E3, an output 14S, two subtraction units 26, 28, two filters 30, 32 and a mixer 34.

The first input 14E1 of the generation module 14 is connected to the overhead microphone 22 to receive the vibration signal from the overhead microphone 22.

The second input 14E2 of the generation module 14 is connected to the environment microphone 24 to receive the vibration signal from the environment 24. The third input 14E3 of the generation module 14 is connected to the osteophonic microphone 20 to receive the vibratory signal detected by the osteophonic microphone 20.

The first subtraction unit 26 comprises two inputs 26E1, 26E2 and an output 26S. The first subtraction unit 26 is able to generate a first output signal equal to the difference between the signal of the first input 26E1 of the first subtraction unit 26 and the signal of the second input 26E2 of the first subtraction unit 26.

In the case of FIG. 1, the first input 26E1 of the first subtraction unit 26 is connected to the first input 14E1 of the generation module 14 to receive the vibration signal coming from the overhead microphone 22. The second input 26E2 of the first subtraction unit 26 is connected to the second input 14E2 of the generation module 14 to receive the vibration signal from the environment microphone 24. The first output signal that the first subtraction unit 26 is able to generate, is equal to the difference between the vibration signal from the overhead microphone 22 and the vibration signal from the environment microphone 24.

The second subtraction unit 28 comprises two inputs 28E1, 28E2 and a 28S output. The second subtraction unit 28 is able to generate a second output signal equal to the difference between the signal of the first input 28E1 of the second subtraction unit 28 and the signal of the second input 28E2 of the second subtraction unit 28.

In the case of FIG. 1, the first input 28E1 of the second subtraction unit 28 is connected to the output 26S of the first subtraction unit 26 to receive the first output signal that the first subtraction unit 26 is capable of generating . The second input 28E2 of the second subtraction unit 28 is connected to the third input 14E3 of the generation module 14 to receive the vibrating signal detected by the osteophonic microphone 20. The second output signal that the second subtraction unit 28 is suitable for generate, is equal to the difference between the first output signal of the first subtraction unit 26 and the vibratory signal detected by the osteophonic microphone 20. The second output signal is equal to the difference between the vibration signal from the air microphone 22 on the one hand and the sum of the vibration signal from the environment 24 and the vibratory signal detected by the osteophonic microphone 20.

The first filter 30 has an input 30E and a 30S output. The first filter 30 is adapted to generate a signal at the output 30S equal to the signal at the input 30E filtered by a high-pass filter. According to the example described, the first filter 30 is a high-pass filter, between 800 hertz (Hz) and 1600 Hz. In this case, the input 30E is connected to the output 28S of the second subtraction unit 28 to receive the second output signal that the second subtraction unit 28 is able to generate.

The second filter 32 has an input 32E and an output 32S. The second filter 32 is able to generate a signal at the output 32S equal to the signal at the input 32E filtered by a low-pass filter. According to the example described, the second filter 32 is a low-pass filter, between 400 Hz and 1200 Hz. In this case, the input 32E is connected to the third input 14E3 of the generation module 14 to receive the vibratory signal. detected by the osteophonic microphone 20.

The mixer 34 has two inputs 34E1, 34E2 and an output 34S. The mixer 34 is able to generate a signal at the output 34S equal to the mixing of the two signals at the two inputs 34E1, 34E2. In the case of Figure 1, the first input 34E1 is connected to the output 30S of the first filter 30 to receive the second filtered output signal. The second input 34E2 is connected to the output 32S of the second filter 32 for the vibratory signal detected by the filtered osteophonic microphone. The signal at the output 34S that the mixer 34 is able to generate is equal to the mixture of the filtered second output signal and the vibratory signal detected by the filtered osteophonic microphone.

The output 14S is connected to the output 34S of the mixer 34S to receive the mixture of the second filtered output signal and the vibratory signal detected by the filtered osteophonic microphone. The generation module 14 is able to generate a signal equal to the mixture of the filtered second output signal and the vibratory signal detected by the filtered osteophonic microphone.

The voice activity detection module 16 comprises two inputs 16E1, 16E2, two outputs 16S1, 16S2, a switching unit 36, two processing units 38, 40 and two amplifiers 42, 44.

The voice activity detection module 16 is adapted to operate in two modes. In the first mode, called transmission mode, the voice detection module 16 is able to generate a transmission signal corresponding to the voice of the user of the circuit 10. In the second mode, called the reception mode, the module voice detection device 16 is adapted to generate a reception signal for supplying the sound transmission module 18 to the user.

The first input 16E1 is connected to the output 14S of the generation module 14 to receive the signal that the generation module 14 is able to generate, in this case a signal equal to the mixture of the second filtered output signal and the detected vibration signal. by the filtered osteophonic microphone. The second input 16E2 is adapted to receive a communication signal coming from the outside, for example from a relay station, as indicated schematically in FIG.

The first output 16S1 is able to emit the transmission signal that the voice activity detection module 16 is able to generate in the first mode of operation as is schematically indicated in FIG.

The second output 16S2 is connected to the sound transmission module 18 to the user so that the sound transmission module 18 receives the reception signal that the voice activity detection module 16 is able to generate in the second mode of transmission. as is schematically indicated in FIG. 1.

The tilt unit 36 has an input 36E and two control outputs 36S1 and 36S2. The failover unit 36 is adapted to control the switchover of the speech activity detection module 16 between the two modes of operation of the speech activity detection module 16 as a function of the signal at the input 36E of the speech unit. 36. In the case of FIG. 1, the input 36E of the tilting unit 36 is connected to the osteophonic microphone 20 to receive the vibratory signal detected by the osteophonic microphone 20.

As can be seen in FIG. 3, the first processing unit 38 comprises an input 38E, an output 38S, a filter 46, a first conversion subunit 48 of an audio signal in a continuous component, a second sub-unit 48 conversion unit 50 of an audio signal into a DC component, a rectifier subunit 52 and a comparison subunit 54.

The input 38E of the first processing unit 38 is connected to the first input 16E1 which is connected to the output 14S of the generation module 14 to receive the signal that the generation module 14 is able to generate, in this case a signal equal to the mixture of the filtered second output signal and the vibratory signal detected by the filtered osteophonic microphone.

The filter 46 of the first processing unit 38 has an input 46E and an output 46S. The filter 48 is able to generate a signal at the output 46S equal to the signal at the input 48E filtered by a bandpass filter. In this case, the input 46E is connected to the input 38E of the first processing unit 38. The bandpass filter is adapted for the human voice. For example, the bandpass filter is capable of passing frequencies between 300 Hz (Hz) and 3000 Hz.

The first conversion subunit 48 comprises an input 48E and a 48S output. The first conversion subunit 48 is adapted to convert the signal of the input 48E into a DC component with a first time constant rated T1. In the case of FIG. 3, the input 48E of the first conversion subunit 48 is connected to the output 46S of the filter 46 of the first processing unit 38 to receive the output signal of the filter 46 of the first unit treatment 38.

The second conversion subunit 50 comprises an input 50E and a 50S output. The second conversion subunit 50 is adapted to convert the signal of the input 50E into a DC component with a second time constant denoted T2. In the case of FIG. 3, the input 50E of the second conversion subunit 50 is connected to the output 46S of the filter 46 of the first processing unit 38 to receive the output signal of the filter 46 of the first unit treatment 38.

The rectification subunit 52 has two inputs 52E1, 52E2, an output

52S, a comparator 56 and a switch 58.

The comparator 56 is able to compare the value of the signals between the first input 52E1 and the second input 52E2 and to control the opening and closing of the switch 58 as a function of the result of the comparison between the two signals at the two inputs 52E1 and 52E2. More specifically, the comparator 56 controls the closing of the switch 58 when the value of the signal at the first input 52E1 is greater than the value of the signal at the second input 52E2 and the opening of the switch 58 in the opposite case.

The switch 58 is interposed between the first input 52E1 and the output 52S of the rectifier subunit 52. In the open position of the switch 58 (the case of FIG. 3), the output 52S of the subunit of FIG. rectification 52 is not connected to the first input 52E1 of the rectifier subunit 52. On the contrary, in the closed position of the switch 58, the output 52S of the rectifier subunit 52 is connected to the first entry 52E1 of the righting sub-unit 52.

In the case of FIG. 3, the first input 52E1 of the righting subunit 52 is connected to the output 48S of the first conversion subunit 48. The second input 52E2 of the righting subunit 52 is connected to a voltage source 60 capable of delivering a DC voltage corresponding to a threshold V52.

The comparison subunit 54 comprises two inputs 54E1, 54E2, an output 54S, a comparator 62 and a switch 64.

The comparator 62 is able to compare the value of the signals between the first input 54E1 of the comparison subunit 54 and the second input 54E2 of the comparison subunit 54. The comparator 62 is also able to control the opening and the closing of the switch 64 of the comparison sub-unit 54 as a function of the result of the comparison between the two signals at the two inputs 54E1 and 54E2. More specifically, the comparator 56 controls the closing of the switch 64 when the value of the signal at the first input 54E1 is greater than the value of the signal at the second input 54E2 and the opening of the switch 64 in the opposite case.

The switch 64 is interposed between the first input 54E1 and the output 54S of the comparison subunit 54. In the open position of the switch 64 (the case of FIG. 3), the output 54S of the subunit of FIG. comparison 54 is not connected to the first input 54E1 of the comparison sub-unit 54. On the contrary, in the closed position of the switch 64, the output 54S of the comparison sub-unit 54 is connected to the first 54E1 input of the comparison sub-unit 54.

In the case of FIG. 3, the first input 54E1 of the comparison subunit 54 is connected to the output 52S of the righting subunit 52. The second input 54E2 of the comparison subunit 54 is connected to the output 50S of the second conversion subunit 50.

The output 38S is connected to the output 54S of the comparison sub-unit 54 to receive a signal which is the signal at the input 38E of the first processing unit 38 having undergone electronic processing.

The second processing unit 40 comprises elements similar to the elements which have been described for the first processing unit 38. In particular, the second processing unit 40 has an input 40E connected to the second input 16E2 of the voice activity detection module. 16 and an output 40S adapted to receive a signal which is the signal at the input 40E of the second processing unit 40 having undergone electronic processing.

The first amplifier 42 has two inputs 42E1, 42E2 and an output 42S. The first amplifier 42 is adapted to generate at the output 42S a signal amplification signal of the first input 42E1 of the first amplifier 42. The amplification of the signal is characterized by a first gain varying according to the signal present at the second input 42E2 of the first amplifier 42. In the case of Figure 3, the first input 42E1 is connected to the first input 16E1 of the voice activity detection module 16, the second input 42E2 is connected to the output 40S of the second unit of processing 40 and the output 42S of the first amplifier 42 is connected to the first output 16S1 of the voice activity detection module 16.

The second amplifier 44 has two inputs 44E1, 44E2 and 44S output. The second amplifier 44 is adapted to generate at the output 44S a signal amplification signal of the second input 44E1 of the second amplifier 44. The amplification of the signal is characterized by a second gain varying according to the signal present at the second input 44E2 of the second amplifier 44. In the case of FIG. 3, the first input 44E1 is connected to the second input 16E2 of the module 44E2. voice activity detection 16, the second input 44E2 is connected to the output 38S of the first processing unit 38 and the output 44S of the second amplifier 44 is connected to the second output 16S2 of the voice activity detection module 16.

The sound emission module 18 towards the user comprises an input 18E and an osteophonic transducer 66.

The input 18E of the sound emission module 18 is connected to the second output 16S2 of the voice activity detection module 16.

The osteophonic transducer 66 is adapted to be positioned on a lateral side of the user's head and is adapted to emit a vibratory signal converted by bone conduction into a user-perceivable acoustic signal. In this case, the osteophonic transducer 66 is connected to the input 18E of the sound emission module 18 to emit a vibratory signal corresponding to the signal present at the input of the transmission module 18.

The operation of the electronic circuit 10 is now described, in particular with reference to FIGS. 4 to 9.

In the context of a usual conversation, either the user emits sounds or the user hears sounds.

When the user emits sounds, the vibratory signal detected by the osteophonic microphone 20 is greater than a predefined threshold called the switching threshold. As a result, the tilt unit 36 switches the voice detection module 16 into the first mode of operation.

The generation module 14 then produces a signal corresponding to the mixing of a signal equal to the mixture of the second filtered output signal and the vibratory signal detected by the filtered osteophonic microphone. The second filtered output signal is equal to the difference between the vibration signal from the overhead microphone 22 on the one hand and the sum of the vibration signal from the environment 24 and the vibratory signal detected by the osteophonic microphone 20. the vibration signal from the overhead microphone 22 is the sum of the voice of the user of the vibration signal from the environment 24 and the vibratory signal detected by the osteophonic microphone 20, the second signal substantially corresponds to the voice of the 'user.

The signal produced by the generation module 14 is then processed by the first processing unit 38. This first processing unit 38 is able to detect whether the signal produced by the generation module 14 is the voice of the user. For this, the filter 46 eliminates all the frequencies that correspond to the human voice.

Then, the signal at the output 46S of the filter 46 is converted by the two conversion subunits 48, 50 into a DC component. As shown in FIGS. 4 and 5, the rectifier subunit 52 is capable of canceling all the voltages of the signal of the first input 52E1 which are lower than the threshold V52. In the case of Figures 4 and 5, all the voltages between instants t1 and t2 are less than V52. At the output of the rectification subunit 52, the signal obtained between times t1 and t2 is zero. Since the threshold V52 is chosen to eliminate the DC components that correspond to noise, the signal at the output 52S of the rectifier subunit 52 is a sound signal.

To ensure that this beep is a voice signal, the case where the signal is a repetitive noise like a machine should be eliminated. This is the role of the comparison sub-unit 54 as illustrated in FIGS. 6 to 9. FIGS. 6 and 7 correspond to the signal of a human voice whereas FIGS. 8 and 9 correspond to the case of a machine noise. In the first case, at the level of the two continuous components (one in solid lines and the other in dotted lines in Figure 7), it is observed that the continuous components do not have the same temporal variation. On the contrary, in the second case, at the level of the two continuous components (one in solid lines and the other in dashed lines in FIG. 7), it is observed that the continuous components have a similar temporal variation. As a result, the output signal of the processed signal is zero at the output 38S for a repetitive signal while it is not zero for a human voice signal.

Therefore, at the first output 16S1 of the voice activity detection module 16, in the first mode of operation, a non-zero signal is issued only when the user has spoken.

Operation in the second mode is analogous to operation in the first mode. This operation is therefore not further described in the following. Similarly to the previous case, at the second output 16S2 of the voice activity detection module 16, in the second mode of operation, a non-zero signal is transmitted only when the received communication signal corresponds to a voice signal.

The circuit 10 therefore makes it possible to better control the triggering of the voice activity detection module 16. In particular, the switchover from a transmission operating mode to a reception mode is controlled on the basis of a signal coming from the user and not on the basis of a signal comprising the noise of the environment.

According to a second embodiment as visible in Figure 10, the elements identical to the circuit 10 according to the first embodiment of Figure 1 are not repeated. Only the differences are highlighted.

The generation module 14 furthermore comprises a fourth input 14E4 capable of receiving the signal of the sound emission module 18 and a third subtraction unit 70. The third subtraction unit 70 comprises two inputs 70E1, 70E2 and an output 70S. The third subtraction unit 70 is adapted to generate a third output signal equal to the difference between the signal of the first input 70E1 of the third subtraction unit 70 and the signal of the second input 70E2 of the third subtraction unit 70.

In the case of FIG. 10, the first input 70E1 of the third subtraction unit 70 is connected to the third input 14E3 of the generation module 14 to receive the vibratory signal detected by the osteophonic microphone 20. The third input 70E2 of the third subtraction unit 70 is connected to the fourth input 14E4 to receive the sound transmission module signal 18. The third output signal that the third subtraction unit 28 is able to generate, is equal to the difference between the detected vibration signal by the osteophonic microphone 20 and the signal of the sound emission module 18.

The second subtraction unit 28 is then able to generate a second signal equal to the difference between the first signal and the third signal, the second input 28E2 of the second subtraction unit 28 being connected to the output 70S of the third subtraction unit 70.

This additional subtraction in the second embodiment makes it possible to improve the filtering of the aerial signal and therefore the electronic circuit 10.

Claims

1. - Electronic circuit (10) for osteophonic headband comprising:

an osteophonic microphone adapted to be positioned on a lateral flank of the head of a user and suitable for detecting a vibratory signal of bone conduction emitted by the user,

a sound emission module (18) towards the user of a signal, and

a voice activity detection module (16) able to operate in a first mode and in a second mode, the voice activity detection module (16) being able to generate a transmission signal corresponding to the voice of the voice of the user in the first mode and the voice activity detection module (16) being adapted to generate a reception signal to supply the sound emission module (18) in the second mode,

characterized in that the voice activity detection module (16) is adapted to switch between the first mode and the second mode according to the vibratory signal detected by the osteophonic microphone (20).

An electronic circuit according to claim 1, wherein the voice activity detection module (16) has a first input (16E1) adapted to receive an input signal comprising a voice signal from the user and a first amplifier (42) of the input signal, the transmission signal generated in the first mode corresponding to the signal amplified by the first amplifier (42).

An electronic circuit according to claim 1 or 2, wherein the voice activity detection module (16) has a second input (16E2) adapted to receive a communication signal and a second amplifier (44) of the communication signal. , the reception signal generated in the second mode corresponding to the signal amplified by the second amplifier (44).

An electronic circuit according to claims 2 and 3, wherein the voice activity detection module (16) also comprises:

a first processing unit (38) of the input signal adapted to perform a processing on the input signal to obtain a processed input signal, the second amplifier (44) having a gain varying according to the signal of processed entry, and

a second processing unit (40) of the communication signal adapted to perform a processing on the communication signal to obtain a signal of communication processed, the first amplifier (42) having a gain varying according to the communication signal processed.

An electronic circuit according to claim 4, wherein the first signal processing unit (38) comprises:

a first conversion sub-unit (48) of an audio signal into a continuous component capable of generating a first signal converted by filtering with a first time constant (T1),

a second sub-unit for converting (50) an audio signal into a continuous component capable of generating a second converted signal by filtering the audio signal with a second time constant (T2), the first time constant (T1) being different from the second time constant (T2), and

a comparison subunit (54) capable of comparing the two converted signals, the processed input signal being equal to the first converted signal when the comparison subunit (50) detects a difference between the two converted signals.

An electronic circuit according to claim 4 or 5, wherein the second signal processing unit (40) comprises:

a comparison subunit (54) capable of comparing the two converted signals, the processed communication signal being equal to the first converted signal when the comparison subunit (50) detects a difference between the two converted signals.

7. Electronic circuit according to any one of claims 1 to 6, further comprising:

an overhead microphone (22) capable of detecting a signal of vibration of the air,

an environment microphone (24) capable of detecting an air vibration signal corresponding to the noise of the environment, and a signal generation module (14) capable of generating the input signal on the basis of the vibration signal from the overhead microphone (22), the vibration signal from the environment microphone (24) and the vibratory signal; detected by the osteophonic microphone (20).

An electronic circuit according to claim 7, wherein the generation module (14) comprises two subtraction units (26, 28), the first subtraction unit (26) being adapted to generate a first output signal equal to the difference between the vibration signal from the overhead microphone (22) and the vibration signal from the environment microphone (24) and the second subtraction unit (28) being able to generate a second output signal equal to the difference between the first signal and the vibratory signal detected by the osteophonic microphone (24).

An electronic circuit according to claim 7, wherein the generation module (14) comprises three subtraction units (26, 28, 70), the first subtraction unit (26) being adapted to generate a first equal output signal. the difference between the vibration signal from the overhead microphone (22) and the vibration signal from the environment microphone (24), the third subtraction unit (70) being able to generate a third output signal equal to the difference between the vibrating signal detected by the osteophonic microphone (20) and the reception signal and the second subtraction unit (28) being able to generate a second output signal equal to the difference between the first output signal and the third signal Release.

10. - Osteophonic headband comprising an electronic circuit (10) according to any one of claims 1 to 9.