A binaural hearing aid self-sound processing method

Abstract

The application discloses a binaural hearing aid self-sound processing method, which comprises the following steps: a microphone picks up a self-sound signal of a wearer as a training acoustic signal; an interaural relative transfer function corresponding to a mouth position of the wearer is estimated according to the binaural training acoustic signal; the microphone picks up binaural acoustic signals in an actual use scene, and the self-sound components in the left and right ear acoustic signals are suppressed by using the interaural relative transfer function; a compensation filter is adaptively estimated to restore the external sound source components in the self-sound suppressed signal, and finally, an output speech signal with suppressed self-sound is obtained. The method can suppress the self-sound of the hearing aid wearer and retain the naturalness and spatial information of the external sound source.

Description

Technical Field

[0001] This invention relates to an audio signal processing method, and more particularly to a method for processing the self-sound of a binaural hearing aid. Background Technology

[0002] Due to physiological limitations, hearing-impaired individuals experience varying degrees of loss in their ability to recognize and distinguish sounds, hindering their ability to communicate normally and causing difficulties in integrating into society. Currently, hearing aids are widely used to enhance the auditory abilities of hearing-impaired individuals, helping them perceive and understand external sounds. Surveys indicate that the naturalness and intelligibility of external speech after wearing hearing aids are key considerations for hearing-impaired individuals when choosing a hearing aid.

[0003] With the development of speech signal processing technology, hearing aids can enhance speech in various complex acoustic environments. In existing hearing aid technology, techniques such as linear amplification and compression amplification can effectively amplify the acoustic signals picked up by the microphone, ensuring that the output sound pressure level is within the wearer's comfortable hearing threshold. However, hearing aids cannot distinguish between external sound sources and the wearer's own voice. When the wearer speaks, the hearing aid picks up the acoustic signal and amplifies it as well. Studies show that many hearing-impaired patients find the amplified self-sound unnatural after wearing hearing aids, leading to low satisfaction and even causing them to choose not to wear hearing aids altogether.

[0004] This demonstrates that the current hearing aid field still lacks technology for processing the wearer's own voice. Therefore, a method for suppressing self-sound in binaural hearing aids has been designed to prevent abnormal amplification of self-sound by the hearing aid, effectively improving the hearing aid user experience for hearing-impaired patients. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for processing the self-sound of a binaural hearing aid, thereby solving the technical problem that hearing-impaired patients are interfered with by abnormal amplification of self-sound and have low satisfaction with hearing aid use.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for processing the self-sound of a binaural hearing aid, comprising the following steps:

[0007] Step 1: Record training acoustic signals: In a low-noise environment, acquire the binaural microphone signals when the hearing aid wearer speaks, specifically the time-domain acoustic signal of the left ear. and right ear time-domain acoustic signal This serves as the training acoustic signal for both ears. and ;

[0008] Step 2: Estimate the interaural relative transfer function: using training acoustic signals from both ears. and Estimate the interauricular relative transfer function corresponding to the position of the wearer's mouth. and To satisfy and ,in This represents the convolution operation. This indicates the set self-sound retention ratio;

[0009] Step 3: Suppress the wearer's own voice: The microphone picks up the acoustic signal of the left ear in the actual scene. and right ear acoustic signal This includes the wearer's own voice and external sound sources: ,in , It represents the wearer's own voice. Representing external sound sources; for each frame of signal, based on the interauricular relative transfer function and By suppressing the wearer's own voice component in the acoustic signals of the left and right ears respectively, the self-voice suppressed acoustic signal is obtained: , ;

[0010] Step 4: Compensate for external sound sources: For each frame of signal, according to... Acoustic signal after self-suppression Estimating the compensation filter This is used to compensate for the external sound sources that cause distortion in step three, and the final speech signal output by the hearing aid is... ,in .

[0011] Self-sound retention ratio The value range is 0 to 1.

[0012] The training acoustic signal length N ranges from 5 to 10 seconds.

[0013] The relative transfer function length L between the ears ranges from 0.1 to 0.6 s.

[0014] The interaural relative transfer function is only related to the head characteristics of the hearing aid wearer, and is not related to sound characteristics such as timbre.

[0015] The method for estimating the interauricular relative transfer function includes the following steps:

[0016] 1) Training acoustic signals Fourier transform results As input, Fourier transform results As the ideal output;

[0017] 2) Estimation using an adaptive filtering algorithm So that it converges and satisfies ,right The interauricular transfer function is obtained by performing an inverse Fourier transform. ;

[0018] 3) Estimate and save the interauricular transfer function using the same method. .

[0019] Compensation filter The estimation method includes the following steps:

[0020] 1) The acoustic signal after self-sound suppression Fourier transform results As input, Fourier transform results As the ideal output;

[0021] 2) Estimation using an adaptive filtering algorithm The time-domain convergence objective is Then the convergence objective in the frequency domain can be expressed as ,in It represents the mathematical expectation.

[0022] Compared with the prior art, the beneficial effects achieved by the present invention are: personalized estimation of the interaural transfer function of the hearing aid wearer, suppression of the wearer's own sound signal, and preservation of the naturalness of the external sound source; the algorithm has low computational load, is easy to implement, and is suitable for implementation in low-power devices such as hearing aids. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of a self-sound processing method for a binaural hearing aid according to the present invention;

[0024] Figure 2 This is a schematic diagram of the interauricular relative transfer function estimation method described in this invention;

[0025] Figure 3 This is a schematic diagram of the compensation filter estimation method described in this invention;

[0026] Figure 4 This is a schematic diagram of acoustic signals used for left and right ear training.

[0027] Figure 5 This is a schematic diagram of the interauricular transfer function obtained using the training acoustic signal in step one;

[0028] Figure 6 This is a schematic diagram of self-sound suppression signal;

[0029] Figure 7 This is a schematic diagram of the final output signal. Detailed Implementation

[0030] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0031] like Figure 1 As shown, a method for processing the self-sound of a binaural hearing aid includes the following steps:

[0032] Step 1: Record training acoustic signals: Acquire the binaural microphone signals when the hearing aid wearer speaks, specifically the time-domain acoustic signal from the left ear. and right ear time-domain acoustic signal This serves as the training acoustic signal for both ears. and The recording environment here is a quiet indoor space with low noise, where the wearer speaks loudly and continuously. Figure 4 A 10-second audio signal for left and right ear training was displayed.

[0033] Step 2: Estimate the interaural relative transfer function: using training acoustic signals from both ears. and Estimate the interauricular relative transfer function corresponding to the position of the wearer's mouth. and To satisfy and ,in This represents the convolution operation. This indicates the set self-sound retention ratio, here. The value can be adjusted within the range of 0 to 1 according to the wearer's needs, and the relative transfer function between the ears can be re-estimated accordingly. This can suppress self-sound signals to the greatest extent, and is the preferred method. After estimating the interauricular relative transfer function, save the data. and The value is used later and does not need to be calculated repeatedly; the length N of the training acoustic signal ranges from 5 to 20 seconds, preferably... The relative transfer function length L between the ears ranges from 0.1 to 0.6 s, preferably... ; Figure 5 The interauricular transfer function obtained using the acoustic signal trained in step one is shown.

[0034] Step 3: Suppress the wearer's own voice: The microphone picks up the acoustic signal of the left ear in the actual scene. and right ear acoustic signal This includes the wearer's own voice and external sound sources: ,in , which respectively represent the left ear and the right ear; It represents the wearer's own voice. Representing external sound sources; for each frame of signal, based on the interauricular relative transfer function and By suppressing the wearer's own voice component in the acoustic signals of the left and right ears respectively, the self-voice suppressed acoustic signal is obtained: , ;Will Substituting the relative transfer function between the ears into the above equation, we can obtain the binaural acoustic signal at this time as follows: , It can be seen that the self-sound component in the signal is suppressed, while the external sound source component is distorted. Figure 6 The self-sound suppression signal is shown, where the dashed box indicates the segment containing only the user's own voice, the solid box indicates the segment containing only the external sound source, and the rest indicates the segment used for mixing self-sound and external sound sources.

[0035] Step 4: Compensate for external sound sources: For each frame of signal, according to... Acoustic signal after self-suppression Estimating the compensation filter This is used to compensate for the external sound sources that cause distortion in step three. Taking the acoustic signal of the left ear as an example, the distorted external sound source is the speech signal ultimately output by the hearing aid. Then it is necessary to estimate satisfy The final output speech can be represented as . Figure 7 The final output signal of the hearing aid is shown; the part marked by the dashed box indicates the segment containing only the user's own voice, the part marked by the solid box indicates the segment containing only the external sound source, and the rest indicates the segment used to mix the user's own voice and the external sound source.

[0036] like Figure 2 As shown, estimating the interauricular transfer function includes the following steps:

[0037] 1) Based on the binaural microphone signals and the set self-sound retention ratio The signal used to estimate the interauricular transfer function is obtained. and ; Training acoustic signals Fourier transform results As input, Fourier transform results As the ideal output;

[0038] 2) Estimation using an adaptive filtering algorithm So that it converges and satisfies ,right The interauricular transfer function is obtained by performing an inverse Fourier transform. Since the interaural transfer function does not require repeated estimation, a relatively long speech is used to make it converge completely.

[0039] 3) Estimate and save the interauricular transfer function using the same method. .

[0040] like Figure 3 As shown, estimating the compensation filter includes the following steps:

[0041] 1) Obtain the acoustic signals after self-sound suppression. and original external acoustic signals ; the acoustic signal after self-sound suppression Fourier transform results As input, Fourier transform results As the ideal output;

[0042] 2) Estimation using an adaptive filtering algorithm :right Performing the inverse Fourier transform yields Since the location of external sound sources is not fixed, the estimation of the compensation filter is performed in the frequency domain frame by frame, with the time domain convergence target being... In the frequency domain, it can be represented as That is Where k represents the frequency index and m represents the frame index. For the filter coefficients at the k-th frequency point in the m-th frame, This represents the original external acoustic frequency domain signal at frequency k in the m-th frame. Let k be the acoustic frequency domain signal after self-sound suppression at frequency k in the m-th frame; here, the preferred frame length is 32ms and the Fourier transform length is 256 points, so the value of k ranges from 1 to 256.

[0043] This invention can personalize the interaural transfer function of hearing aid wearers, suppress the wearer's own sound signal, and preserve the naturalness of external sound sources. The algorithm has low computational complexity, is easy to implement, and is suitable for implementation in low-power devices such as hearing aids.

[0044] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing the self-sound of a binaural hearing aid, characterized in that, The method includes the following steps: Step 1: Record training acoustic signals: In a low-noise environment, acquire the binaural microphone signals when the hearing aid wearer speaks, specifically the time-domain acoustic signal of the left ear. and right ear time-domain acoustic signal This serves as the training acoustic signal for both ears. and ; Step 2: Estimate the interaural relative transfer function: using training acoustic signals from both ears. and Estimate the interauricular relative transfer function corresponding to the position of the wearer's mouth. and To satisfy and ,in This represents the convolution operation. This indicates the set self-sound retention ratio; Step 3: Suppress the wearer's own voice: The microphone picks up the acoustic signal of the left ear in the actual scene. and right ear acoustic signal This includes the wearer's own voice and external sound sources: ,in , It represents the wearer's own voice. Representing external sound sources; for each frame of signal, based on the interauricular relative transfer function and By suppressing the wearer's own voice component in the acoustic signals of the left and right ears respectively, the self-voice suppressed acoustic signal is obtained: , ; Step 4: Compensate for external sound sources: For each frame of signal, according to... Acoustic signal after self-sound suppression Estimating the compensation filter This is used to compensate for the external sound sources that cause distortion in step three, and the final speech signal output by the hearing aid is... ,in .

2. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The retention ratio of the self-sound The value range is 0 to 1.

3. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The training acoustic signal length N ranges from 5 to 10 seconds.

4. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The interauricular relative transfer function length L ranges from 0.1 to 0.6 s.

5. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The interaural relative transfer function is only related to the head characteristics of the hearing aid wearer, and is not related to the timbre or sound characteristics.

6. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The method for estimating the interauricular relative transfer function includes the following steps: 1) Training acoustic signals Fourier transform results As input, Fourier transform results As the ideal output; 2) Estimation using an adaptive filtering algorithm So that it converges and satisfies ,right The interauricular transfer function is obtained by performing an inverse Fourier transform. ; 3) Estimate and save the interauricular transfer function using the same method. .

7. The self-sound processing method for a binaural hearing aid according to claim 1, characterized in that, The compensation filter The estimation method includes the following steps: 1) The acoustic signal after self-sound suppression Fourier transform results As input, Fourier transform results As the ideal output; 2) Estimation using an adaptive filtering algorithm The time-domain convergence objective is Then the convergence objective in the frequency domain can be expressed as ,in It represents the mathematical expectation.

Images