Bluetooth earphone wearer voice detection method and system, and Bluetooth earphone

Abstract

The invention relates to the field of audio recognition, in particular to a Bluetooth headset wearer voice detection method and system and a Bluetooth headset, and the method comprises the steps: obtaining a first sound signal and a second sound signal; wherein the first sound signal is from the call microphone, and the second sound signal is from the feed-forward microphone; synthesizing the first sound signal and the second sound signal based on a preset time difference of receiving the voice of the wearer by the feed-forward microphone and the call microphone to obtain a synthesized signal; and detecting whether the synthetic signal comprises a voice signal according to a voice activity detection algorithm. Compared with a traditional single-microphone scheme, extra hardware equipment does not need to be introduced, so that the complexity of the system is reduced, the calculation overhead is reduced, and the hardware cost is also greatly reduced.

Description

Technical Field

[0001] The present invention relates to the field of audio recognition, and in particular to a method and system for detecting the voice of a wearer of a Bluetooth headset, and a Bluetooth headset. Background Art

[0002] Voice activity detection is a technology used to determine whether the current audio signal contains voice. It can monitor in real time whether the wearer of the Bluetooth headset is speaking. Once it detects that the wearer of the Bluetooth headset is speaking, the Bluetooth headset can automatically trigger a series of intelligent responses, such as turning off active noise reduction, pausing music playback, switching to ambient sound mode, etc., to enhance the user's auditory experience during the conversation.

[0003] Existing solutions mainly rely on bone vibration sensors to capture the tiny vibrations of the user's head bones when speaking, and convert them into electrical signals to determine whether the user is speaking. However, this method relies on the performance and accuracy of the bone vibration sensor and is relatively costly. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a method and system for detecting the voice of a wearer of a Bluetooth headset and a Bluetooth headset.

[0005] The first aspect of the present invention discloses a method for detecting voice of a wearer of a Bluetooth headset, which is used for a Bluetooth headset including a feedforward microphone and a call microphone. The voice detection method comprises:

[0006] Acquire a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone;

[0007] Based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, synthesizing the first sound signal and the second sound signal to obtain a synthesized signal;

[0008] According to a voice activity detection algorithm, it is detected whether the synthesized signal includes a voice signal.

[0009] Furthermore, the time difference is Δt:

[0010]

[0011] Among them, d is the distance between the feedforward microphone and the call microphone, θ is the angle between the preset transmission direction of the wearer's voice signal and the connecting line between the feedforward microphone and the call microphone, and c is the propagation speed of sound in the air.

[0012] Furthermore, based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, the first sound signal and the second sound signal are synthesized to obtain a synthesized signal, including:

[0013] Calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal based on the time difference;

[0014] The first sound signal and the second sound signal are synthesized according to the first weighting coefficient and the second weighting coefficient:

[0015] y(t)=w1*x mic1 (t)+w2*x mic2 (t);

[0016] Among them, t represents time, w1 represents the first weighting coefficient, x mic1 (t) represents the first sound signal, w2 represents the second weighting coefficient, x mic2 (t) represents the second sound signal, and y(t) represents the synthesized signal.

[0017] Further, based on the time difference, calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal includes:

[0018] Setting the first weighting coefficient to 1;

[0019] The second weighting coefficient is set based on the time difference:

[0020] w2=e -j2πfΔt ;

[0021] Among them, f represents a preset voice frequency, and Δt represents the time difference between the feedforward microphone and the call microphone receiving the wearer's voice.

[0022] Furthermore, the step of detecting whether the synthesized signal includes a speech signal according to a speech activity detection algorithm comprises:

[0023] Performing frequency domain conversion on the synthesized signal to obtain a frequency domain synthesized signal;

[0024] According to a plurality of preset frequency band ranges, the frequency domain synthesized signal is divided into a plurality of sub-band signals;

[0025] Calculate the speech prediction probability of each subband signal;

[0026] Based on the speech prediction probabilities of all the sub-band signals, it is determined whether the synthesized signal includes a speech signal.

[0027] Furthermore, the step of calculating the speech prediction probability of each subband signal includes:

[0028] Calculating the energy of each of the sub-band signals;

[0029] The energy of the sub-band signal is input into a Gaussian speech detection model to obtain the speech prediction probability of each sub-band signal.

[0030] Furthermore, the step of inputting the energy of the subband signal into a Gaussian speech detection model to obtain the speech prediction probability of each subband signal includes:

[0031] Inputting the energy of the subband signal into a speech signal Gaussian mixture submodel of a Gaussian speech detection model to obtain an initial speech prediction probability of each subband signal;

[0032] Inputting the energy of the subband signal into the noise signal Gaussian mixture submodel of the Gaussian speech detection model to obtain the noise prediction probability of each subband signal;

[0033] The log-likelihood ratio of the initial speech prediction probability and the noise prediction probability of each sub-band signal is calculated to obtain the speech prediction probability of each sub-band signal.

[0034] Further, based on the speech prediction probabilities of all the sub-band signals, the step of determining whether the synthesized signal includes a speech signal comprises:

[0035] Based on a preset weighting coefficient of the subband signal, modifying the corresponding speech prediction probability;

[0036] Summing up the speech prediction probabilities corrected by all the sub-band signals;

[0037] According to the summation result, it is determined whether the synthesized signal includes a speech signal.

[0038] The second aspect of the present invention discloses a voice detection system for a wearer of a Bluetooth headset, comprising:

[0039] An acquisition module, used to acquire a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone;

[0040] A synthesis module, configured to synthesize the first sound signal and the second sound signal based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, to obtain a synthesized signal;

[0041] The detection module is used to detect whether the synthesized signal includes a speech signal according to a speech activity detection algorithm.

[0042] A third aspect of the present invention discloses a Bluetooth headset, comprising a feedforward microphone, a call microphone and a processor, wherein the processor can execute the wearer voice detection method of the Bluetooth headset as described in any one of the first aspect of the present invention.

[0043] The present invention achieves speech enhancement and noise suppression at the signal level by making full use of the spatial information of the microphone array, creating favorable conditions for subsequent speech activity detection. Compared with the traditional single-microphone solution, the present invention does not require the introduction of additional hardware devices, which not only reduces the complexity of the system and reduces the computational overhead, but also greatly reduces the hardware cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 It is a flowchart of a method for detecting voice of a wearer of a Bluetooth headset disclosed in an embodiment of the present invention;

[0046] Figure 2 It is a schematic diagram of the structure of a Bluetooth headset and a diagram showing the relationship between the signal and the headset position disclosed in an embodiment of the present invention;

[0047] Figure 3 It is a structural schematic diagram of a voice detection system for a wearer of a Bluetooth headset disclosed in an embodiment of the present invention;

[0048] Figure 4 It is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, or product end including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0051] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0052] The present invention cleverly utilizes the existing feedforward microphone and call microphone on the headset. Through the phase difference and amplitude difference between them, combined with beamforming technology, a directional beam can be formed to accurately enhance the voice signal from the wearer's mouth, and effectively suppress noise and interference from other directions, such as environmental noise, conversations of people around, etc., thereby significantly improving the signal-to-noise ratio of the wearer's voice. In addition, on the basis of beamforming technology, the present invention also introduces a voice activity detection (VAD) algorithm, VAD (Voice Activity Detection) algorithm. The VAD algorithm further screens out the real voice segment by analyzing the energy, spectral characteristics, fundamental period and other parameters of the voice signal, and eliminates the interference of non-voice segments, such as background noise, breathing sound, keystroke sound, etc. The realization of this technical effect not only further improves the signal-to-noise ratio of the wearer's voice, but also significantly reduces the misrecognition rate of the wearer's voice, making the wearer's voice recognition more accurate and reliable.

[0053] See also Figure 1 As shown, Figure 1 1 is a flow chart of a method for detecting the voice of a wearer of a Bluetooth headset disclosed in an embodiment of the present invention, which is used for a Bluetooth headset including a feedforward microphone and a call microphone. Figure 1 As shown, the wearer voice detection method of the Bluetooth headset may include the following operations:

[0054] S101, obtaining a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone;

[0055] In this optional embodiment, the call microphone is usually located inside the headset, close to the wearer's mouth. The main function of the call microphone is to receive the wearer's voice signal for calls and voice interaction. Because it is close to the mouth, it can clearly capture the wearer's voice. The signal received by the call microphone contains a strong wearer's voice and a certain degree of environmental noise.

[0056] The feedforward microphone is usually located outside the headset, away from the wearer's mouth. The main function of the feedforward microphone is to receive ambient noise signals for active noise reduction. By receiving external noise, an anti-phase noise signal with equal amplitude and opposite phase to the noise signal can be generated, thereby canceling the noise.

[0057] The call microphone and the feedforward microphone play different but complementary roles in a Bluetooth headset. By properly utilizing the signal characteristics of the two microphones, the effect of voice detection and enhancement can be improved, thereby providing users with a clearer and more comfortable call and audio experience. At present, most Bluetooth headsets on the market are equipped with a call microphone and a feedforward microphone. The present invention does not require additional hardware equipment and can reduce hardware costs.

[0058] S102, synthesizing the first sound signal and the second sound signal based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice to obtain a synthesized signal;

[0059] In an optional embodiment, the time difference is Δt:

[0060]

[0061] Among them, d is the distance between the feedforward microphone and the call microphone, θ is the angle between the preset transmission direction of the wearer's voice signal and the connecting line between the feedforward microphone and the call microphone, and c is the propagation speed of sound in the air.

[0062] Figure 2 It shows the schematic diagram of the structure of Bluetooth headset and the relationship between the signal and the headset position. Figure 2 Where mic1 is the call microphone, mic2 is the feedforward microphone, θ is the angle between the transmission direction of the wearer's voice signal and the line connecting the feedforward microphone and the call microphone, d is the distance between the feedforward microphone and the call microphone, s(t) is the wearer's voice signal, n1(t) is the noise signal of the call microphone, and n2(t) is the noise signal of the feedforward microphone.

[0063] In this optional embodiment, the preset value of θ is determined according to the wearing method of the earphone and the positions of the two microphones on the earphone. An optimal preset value of θ can be obtained by statistically analyzing the wearing conditions of multiple users.

[0064] In an optional embodiment, based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, synthesizing the first sound signal and the second sound signal to obtain a synthesized signal includes:

[0065] Calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal based on the time difference;

[0066] The first sound signal and the second sound signal are synthesized according to the first weighting coefficient and the second weighting coefficient:

[0067] y(t)=w1*x mic1 (t)+w2*x mic2 (t);

[0068] Where t represents time, w1 represents the first weighting coefficient, x mic1 (t) represents the first sound signal, w2 represents the second weighting coefficient, x mic2 (t) represents the second sound signal, and y(t) represents the synthesized signal.

[0069] First sound signal x mic1 (t) is:

[0070] x mic1 (t) = s(t) + n1(t);

[0071] Second sound signal x mic2 (t) Based on the first sound signal x mic1 (t) and the time difference Δt are estimated:

[0072] x mic2 (t) = s(t + Δt) + n2(t);

[0073] In this optional embodiment, the two microphones in the Bluetooth headset form a small sensor array that points toward the wearer's mouth. This means that the speech signal s(t) will mainly come from this direction. The noise signals n1(t) and n2(t) may come from all other directions.

[0074] For the speech signal s(t) coming from the direction of the wearer's speech, since it has a certain delay between the two microphones, its amplitude will be enhanced after weighted processing. For the noise signals n1(t) and n2(t) coming from other directions, since they have different delays between the two microphones, their amplitudes will be suppressed after weighted processing. After synthesizing the first sound signal and the second sound signal through the above formula, the amplitude of the speech signal coming from the direction of the wearer's speech in the synthesized signal can be significantly increased, thereby improving the signal-to-noise ratio.

[0075] In an optional embodiment, calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal based on the time difference includes:

[0076] Setting the first weighting coefficient to 1;

[0077] The second weighting coefficient is set based on the time difference:

[0078] w2=e -j2πfΔt ;

[0079] Among them, f represents a preset voice frequency, and Δt represents the time difference between the feedforward microphone and the call microphone receiving the wearer's voice.

[0080] In this optional embodiment, a wearer voice detection experiment is performed by taking values ​​from the main frequency range of human voice between 300 Hz and 3000 Hz, and the frequency with the best experimental effect is selected as the preset voice frequency f.

[0081] In the frequency domain, the delay of a signal is equivalent to multiplying a complex exponential factor, which is proportional to the preset voice frequency f and the time difference Δt. By setting the first weighting coefficient and the second weighting coefficient, the delay effect of the language signal can be simulated in the frequency domain, so that the first sound signal and the second sound signal can be more accurately aligned when synthesizing the signal.

[0082] S103: Detect whether the synthesized signal includes a speech signal according to a speech activity detection algorithm.

[0083] The voice activity detection algorithm is a technology used to identify in real time whether a sound signal contains a voice segment. It can be a Mel-Frequency Cepstral Coefficients (MFCC) algorithm, a Spectral Entropy (SE) algorithm, a Short-time Energy (STE) algorithm, etc., and the present invention does not limit it.

[0084] In an optional embodiment, the step of detecting whether the synthesized signal includes a speech signal according to a speech activity detection algorithm comprises:

[0085] Performing frequency domain conversion on the synthesized signal to obtain a frequency domain synthesized signal;

[0086] According to a plurality of preset frequency band ranges, the frequency domain synthesized signal is divided into a plurality of sub-band signals;

[0087] Calculate the speech prediction probability of each subband signal;

[0088] Based on the speech prediction probabilities of all the sub-band signals, it is determined whether the synthesized signal includes a speech signal.

[0089] In this optional embodiment, performing frequency domain conversion on the synthesized signal refers to performing short-time Fourier transform on the synthesized signal to obtain a frequency domain synthesized signal.

[0090] In an optional embodiment, the pre-set multiple frequency band ranges are six frequency bands: 80-250 Hz, 250-500 Hz, 500 Hz-1 KHz, 1 K-2 KHz, 2 K-3 KHz, and 3 K-4 KHz.

[0091] In an optional embodiment, the step of calculating the speech prediction probability of each subband signal includes:

[0092] Calculating the energy of each of the sub-band signals;

[0093] The energy of the sub-band signal is input into a Gaussian speech detection model to obtain the speech prediction probability of each sub-band signal.

[0094] In this optional embodiment, calculating the energy of each sub-band signal means square the modulus of the spectrum value in the frequency band corresponding to the sub-band signal and then adding them.

[0095] In an optional embodiment, the step of inputting the energy of the subband signal into a Gaussian speech detection model to obtain the speech prediction probability of each subband signal comprises:

[0096] Inputting the energy of the subband signal into a speech signal Gaussian mixture submodel of a Gaussian speech detection model to obtain an initial speech prediction probability of each subband signal;

[0097] Inputting the energy of the subband signal into the noise signal Gaussian mixture submodel of the Gaussian speech detection model to obtain the noise prediction probability of each subband signal;

[0098] The log-likelihood ratio of the initial speech prediction probability and the noise prediction probability of each sub-band signal is calculated to obtain the speech prediction probability of each sub-band signal.

[0099] In this optional embodiment, the energy of each subband signal is used as the input of the Gaussian mixture submodel of the speech signal to obtain the initial speech prediction probability of the subband signal; the energy of each subband signal is used as the input of the Gaussian mixture submodel of the noise signal to obtain the noise prediction probability of the subband signal.

[0100] The ratio of the predicted probabilities of each subband signal under the two hypotheses of speech and noise, i.e., the likelihood ratio, is calculated. The likelihood ratio is a method of measuring the probability that data belongs to one category relative to the probability that it belongs to another category. In order to facilitate calculation and comparison, the present invention takes the logarithm of the likelihood ratio to obtain the logarithmic likelihood ratio. The logarithmic function is monotonically increasing, so taking the logarithm will not change the magnitude relationship of the likelihood ratio, but it can convert the multiplication calculation into the addition calculation, simplifying the calculation process.

[0101] In an optional embodiment, the step of determining whether the synthesized signal includes a speech signal based on the speech prediction probabilities of all the sub-band signals comprises:

[0102] Based on a preset weighting coefficient of the subband signal, modifying the corresponding speech prediction probability;

[0103] Summing up the speech prediction probabilities corrected by all the sub-band signals;

[0104] According to the summation result, it is determined whether the synthesized signal includes a speech signal.

[0105] In this optional embodiment, a weighting coefficient is preset for each sub-band signal, and then the speech prediction probabilities of all sub-band signals are weighted averaged. The selection of weights can be determined according to the importance and credibility of the frequency band. For example, if the results of the low frequency band are considered more credible, then the results of the low frequency band can be given a greater weight.

[0106] Judging whether the synthesized signal includes a speech signal according to the summation result refers to judging whether the summation result is greater than 0, which indicates that the synthesized signal includes a speech signal, otherwise, the synthesized signal does not include a speech signal.

[0107] like Figure 3 As shown, the present invention provides a voice detection system for a wearer of a Bluetooth headset, comprising:

[0108] An acquisition module 301 is used to acquire a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone;

[0109] A synthesis module 302, configured to synthesize the first sound signal and the second sound signal based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice to obtain a synthesized signal;

[0110] The detection module 303 is configured to detect whether the synthesized signal includes a speech signal according to a speech activity detection algorithm.

[0111] For the specific definition of the voice detection system for the wearer of the Bluetooth headset, please refer to the definition of the voice detection method for the wearer of the Bluetooth headset mentioned above, which will not be repeated here. Each module in the above-mentioned voice detection system for the wearer of the Bluetooth headset can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in hardware format, or can be stored in the memory in the electronic device in software format, so that the processor can call the operations corresponding to the above modules.

[0112] It should be noted that, in order to highlight the innovative part of the present invention, the present embodiment does not introduce modules that are not closely related to solving the technical problem proposed by the present invention, but this does not mean that there are no other modules in the present embodiment.

[0113] like Figure 4 As described above, the electronic device 1 provided by the present invention may include a memory 12, a processor 13 and a bus, and may also include a computer program stored in the memory 12 and executable on the processor 13, such as a voice detection program for the wearer of a Bluetooth headset.

[0114] Among them, the memory 12 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 12 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 12 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 12 can also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 12 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code for voice detection of the wearer of the Bluetooth headset, etc., but also can be used to temporarily store data that has been output or is to be output.

[0115] In some embodiments, the processor 13 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 13 is the control core (Control Unit) of the electronic device 1, and uses various interfaces and lines to connect the various components of the entire electronic device 1, and executes or executes programs or modules stored in the memory 12 (such as a voice detection program for the wearer of a Bluetooth headset, etc.), and calls the data stored in the memory 12 to execute various functions of the electronic device 1 and process data.

[0116] The processor 13 executes the operating system and various installed applications of the electronic device 1. The processor 13 executes the applications to implement the steps in the above-mentioned method for detecting the voice of the wearer of the Bluetooth headset.

[0117] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory 12 and executed by the processor 13 to complete the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the electronic device 1. For example, the computer program may be divided into an acquisition module 301, a synthesis module 302, and a detection module 303.

[0118] The above-mentioned integrated unit implemented in the form of a software function module can be stored in a computer-readable storage medium, and the storage medium can be non-volatile or volatile. The above-mentioned software function module is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a computer device, or a network device, etc.) or a processor to perform part of the functions of the Bluetooth headset wearer voice detection method described in each embodiment of the present application.

[0119] In one embodiment, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps implemented when the processor executes the computer program can also be implemented.

[0120] In summary, the present invention discloses a method, system and Bluetooth headset for detecting the voice of the wearer of a Bluetooth headset, which does not require the introduction of additional hardware devices, which not only reduces the complexity of the system and reduces the computing overhead, but also greatly reduces the hardware cost. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has a high industrial utilization value.

[0121] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for detecting voice of a wearer of a Bluetooth headset, characterized in that: For a Bluetooth headset including a feedforward microphone and a call microphone, the voice detection method comprises: Acquire a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone; Based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, synthesizing the first sound signal and the second sound signal to obtain a synthesized signal; According to a voice activity detection algorithm, it is detected whether the synthesized signal includes a voice signal.

2. The method for detecting voice of a wearer of a Bluetooth headset according to claim 1, characterized in that: The time difference is Δt: Among them, d is the distance between the feedforward microphone and the call microphone, θ is the angle between the preset transmission direction of the wearer's voice signal and the connecting line between the feedforward microphone and the call microphone, and c is the propagation speed of sound in the air.

3. The method for detecting voice of a wearer of a Bluetooth headset according to claim 1, characterized in that: The step of synthesizing the first sound signal and the second sound signal based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice to obtain a synthesized signal includes: Calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal based on the time difference; The first sound signal and the second sound signal are synthesized according to the first weighting coefficient and the second weighting coefficient: y(t)=w1*x mic1 (t)+w2*x mic2 (t); Among them, t represents time, w1 represents the first weighting coefficient, x mic1 (t) represents the first sound signal, w2 represents the second weighting coefficient, x mic2 (t) represents the second sound signal, and y(t) represents the synthesized signal.

4. The method for detecting voice of a wearer of a Bluetooth headset according to claim 3, characterized in that: Calculating a first weighting coefficient of the first sound signal and a second weighting coefficient of the second sound signal based on the time difference includes: Setting the first weighting coefficient to 1; The second weighting coefficient is set based on the time difference: w2=e -j2πfΔt ; Among them, f represents a preset voice frequency, and Δt represents the time difference between the feedforward microphone and the call microphone receiving the wearer's voice.

5. The method for detecting voice of a wearer of a Bluetooth headset according to claim 1, characterized in that: The step of detecting whether the synthesized signal includes a speech signal according to a speech activity detection algorithm comprises: Performing frequency domain conversion on the synthesized signal to obtain a frequency domain synthesized signal; According to a plurality of preset frequency band ranges, the frequency domain synthesized signal is divided into a plurality of sub-band signals; Calculate the speech prediction probability of each subband signal; Based on the speech prediction probabilities of all the sub-band signals, it is determined whether the synthesized signal includes a speech signal.

6. The method for detecting voice of a wearer of a Bluetooth headset according to claim 5, characterized in that: The step of calculating the speech prediction probability of each subband signal comprises: Calculating the energy of each of the sub-band signals; The energy of the sub-band signal is input into a Gaussian speech detection model to obtain the speech prediction probability of each sub-band signal.

7. The method for detecting voice of a wearer of a Bluetooth headset according to claim 6, characterized in that: The step of inputting the energy of the subband signal into a Gaussian speech detection model to obtain the speech prediction probability of each subband signal comprises: Inputting the energy of the subband signal into a speech signal Gaussian mixture submodel of a Gaussian speech detection model to obtain an initial speech prediction probability of each subband signal; Inputting the energy of the subband signal into the noise signal Gaussian mixture submodel of the Gaussian speech detection model to obtain the noise prediction probability of each subband signal; The log-likelihood ratio of the initial speech prediction probability and the noise prediction probability of each sub-band signal is calculated to obtain the speech prediction probability of each sub-band signal.

8. The method for detecting voice of a wearer of a Bluetooth headset according to claim 5, characterized in that: The step of determining whether the synthesized signal includes a speech signal based on the speech prediction probabilities of all the sub-band signals comprises: Based on a preset weighting coefficient of the subband signal, modifying the corresponding speech prediction probability; Summing up the speech prediction probabilities corrected by all the sub-band signals; According to the summation result, it is determined whether the synthesized signal includes a speech signal.

9. A voice detection system for a Bluetooth headset wearer, characterized in that: include: An acquisition module, used to acquire a first sound signal and a second sound signal; wherein the first sound signal comes from the call microphone, and the second sound signal comes from the feedforward microphone; A synthesis module, configured to synthesize the first sound signal and the second sound signal based on a preset time difference between the feedforward microphone and the call microphone receiving the wearer's voice, to obtain a synthesized signal; The detection module is used to detect whether the synthesized signal includes a speech signal according to a speech activity detection algorithm.

10. A Bluetooth headset, comprising a feedforward microphone, a call microphone and a processor, wherein the processor can execute the wearer voice detection method of the Bluetooth headset according to any one of claims 1 to 8.

Images