A method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones

Abstract

The present invention discloses a method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones. First, a dual microphone array pickup module is used to collect ambient sound signals and obtain a main microphone signal and a secondary microphone signal. Then, the main microphone signal and the secondary microphone signal are processed to obtain the maximum energy value of each frequency point of the two microphone signals. Next, the modulation amount of the maximum energy value of each frequency point at each frequency point is calculated. Based on the modulation amount of the maximum energy value of each frequency point at each frequency point; the present invention realizes the function of fully utilizing the spatial orientation information of the dual microphones of the hearing aid to identify different usage scenarios, and can refine and calibrate the algorithm architecture of the hearing aid according to the specific scenario so as to provide a better usage experience, significantly improve the speech clarity and wearing comfort, enhance the communication confidence and quality of life of users in daily life and social occasions, and is suitable for being widely promoted and used.

Description

Technical Field

[0001] The present invention relates to the technical field of hearing assistance devices, and particularly relates to a method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones. Background Art

[0002] With the development of society and the aggravation of the problem of population aging, hearing impairment has become an increasingly serious social problem. As an important tool to improve the quality of life of hearing-impaired people, hearing aids have evolved from the initial analog hearing aids to today's digital hearing aids, and the signal processing technology has been continuously improved. Modern hearing aids can not only amplify sounds, but also perform complex signal processing to provide a clearer and more natural auditory experience. The core algorithms of current mainstream digital hearing aids include WDRC and noise reduction, which are respectively used to compress and compensate for the hearing loss of patients in specific frequency bands and reduce background noise to improve wearing comfort. In addition, some hearing aids are equipped with a feedback suppression algorithm to prevent the hearing aid from whistling feedback, and a directional pickup algorithm based on multiple microphones.

[0003] However, the usage scenarios of hearing aids are complex and variable. In some usage scenarios of patients, the hearing aids cannot obtain a good usage experience. For example, the hearing aids are prone to produce feedback whistling during wearing, and in a noisy environment, the noise reduction algorithm will reduce the intelligibility of the patient's listening, easily introduce and amplify wind noise, and easily amplify the patient's own voice too much; existing hearing aids generally rely on a single microphone for sound collection and processing. Although this method can meet the basic needs in simple usage scenarios, it is difficult to effectively identify and utilize the spatial information contained in the sound signal, thus unable to obtain a better listening experience; therefore, it is necessary to design a method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to better and effectively solve the problem that in some usage scenarios of patients, the hearing aids cannot obtain a good usage experience. For example, the hearing aids are prone to produce feedback whistling during wearing, and in a noisy environment, the noise reduction algorithm will reduce the intelligibility of the patient's listening, easily introduce and amplify wind noise, and easily amplify the patient's own voice too much. The present invention provides a method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones, which realizes the function of fully utilizing the spatial orientation information of the dual microphones of the hearing aid to identify different usage scenarios, and can refine and calibrate the algorithm architecture of the hearing aid according to specific scenarios so as to provide a better usage experience, significantly improve the speech clarity and wearing comfort, and enhance the user's communication confidence and quality of life in daily life and social occasions.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for identifying and adjusting spatial information of a hearing aid based on dual microphones, comprising the following steps:

[0007] Step A: Use a dual microphone array pickup module to collect ambient sound signals and obtain the main microphone signal s1 and the secondary microphone signal s2. Then process the main microphone signal s1 and the secondary microphone signal s2 to obtain the larger energy value of each frequency point of the two microphone signals.

[0008] Step B: Calculate the modulation amount of the larger energy value of each frequency point at each frequency point.

[0009] Step C: Based on the modulation amount of the larger energy value of each frequency point at each frequency point, use the least mean square error method to calculate the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point respectively.

[0010] Step D: Calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient with the respective scene parameters stored in the spatial information identification module to obtain the spatial scene corresponding to the ambient sound signal.

[0011] Step E: Perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the ambient sound signal to complete the operation of identifying and adjusting the spatial information of the hearing aid.

[0012] For the aforementioned method for identifying and adjusting spatial information of a hearing aid based on dual microphones, in step A, use a dual microphone array pickup module to collect ambient sound signals and obtain the main microphone signal s1 and the secondary microphone signal s2. Then process the main microphone signal s1 and the secondary microphone signal s2 to obtain the larger energy value of each frequency point of the two microphone signals. The specific steps are as follows:

[0013] Step A1: Perform short-time Fourier transform on the main microphone signal s1 and the secondary microphone signal s2 respectively, and then decompose the main microphone signal s1 and the secondary microphone signal s2 into complex representations in the frequency domain, as shown in formula (1):

[0014] S1 = STFT(s1), S2 = STFT(s2) (1)

[0015] Wherein, S1 is the complex number in the frequency domain of the main microphone signal s1, S2 is the complex number in the frequency domain of the secondary microphone signal s2, and STFT represents the short-time Fourier transform operator.

[0016] Step A2: Calculate the energy of each frequency point of the complex number S1 in the frequency domain of the main microphone signal s1 and the complex number S2 in the frequency domain of the secondary microphone signal s2, as shown in formula (2):

[0017] P1 = abs(S1) 2 , P2 = abs(S2) 2 (2)

[0018] Among them, P1 is the energy at each frequency point of the complex number S1, P2 is the energy at each frequency point of the complex number S2, and abs is to take the modulus of the complex number;

[0019] Step A3, compare the magnitudes of the energy P1 at each frequency point of the complex number S1 and the energy P2 at each frequency point of the complex number S2, and obtain the larger value P3 of the energy at each frequency point, as shown in formula (3):

[0020] P3 = max(P1, P2) (3)

[0021] Among them, max is to take the maximum value between the two for the sequence.

[0022] For the aforementioned method for identifying and calibrating spatial information of a hearing aid based on two microphones, in step B, calculate the modulation amount of the larger value of the energy at each frequency point at each frequency point. The specific steps are as follows:

[0023] Step B1, calculate the short-time envelope of the signal energy by first-order smoothing calculation, as shown in formula (4):

[0024] P a = a1 P3+(1 - a1) P a (4)

[0025] Among them, P a is the short-time energy envelope at each frequency point, and a1 is the smoothing factor;

[0026] Step B2, perform long-term minimum value statistics on the short-time energy envelope P a at each frequency point and obtain the long-term minimum value P b , as shown in formula (5):

[0027] P b = a2 P a +(1 - a2) P b , when P b ≤P a ;

[0028] P b = P a , when P b > P a (5)

[0029] Among them, a2 is the smoothing factor;

[0030] Step B3, calculate the short-time energy envelope P a at each frequency point and the long-term energy minimum value P bThe difference is obtained to get the modulation amount M, as shown in formula (6).

[0031] M = 10 lg(P a ) - 10 lg(P b ) (6);

[0032] Step B4, set the threshold Th1 and threshold Th2 for judging the energy activation state of each frequency point. If the modulation amount M is not greater than the threshold Th1, the frequency point signal is in an inactive state. If the modulation amount M is not less than the threshold Th2, the frequency point signal is in an active state. If the threshold Th1 < modulation amount M < threshold Th2, the frequency point signal is in a transition state.

[0033] For the aforementioned method for identifying and calibrating spatial information of a hearing aid based on dual microphones, in step C, based on the modulation amount of the maximum value of energy at each frequency point, the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point are calculated respectively using the least mean square error method. Specifically, the background noise coherence coefficient W1 and the activation signal coherence coefficient W2 in each state of each frequency point are statistically calculated respectively using the least mean square error method, as shown in formula (7).

[0034] W1 = W1 + d1 (S2 - W1 S1) conj(S1), if M ≤ Th1;

[0035] W2 = W2 + d2 (S2 - W2 S1) conj(S1), if M ≥ Th2 (7)

[0036] Among them, d1 is the background noise update rate, and d2 is the activation signal update rate.

[0037] For the aforementioned method for identifying and calibrating spatial information of a hearing aid based on dual microphones, in step D, the background noise coherence coefficient and the activation signal coherence coefficient are respectively calculated and judged with the respective scene parameters stored in the spatial information recognition module to obtain the spatial scene corresponding to the environmental sound signal. The specific steps are as follows.

[0038] Step D1, calculate the first scene discrimination parameter y1 and judge whether the environmental sound signal is a coherent noise scene, as shown in formula (8).

[0039] y1 = sum(abs(W1) 2 C1) + b1 (8)

[0040] Among them, sum is the sequence summation operator, C1 and b1 are the corresponding spatial information recognition parameters. If the result of the first scenario discrimination parameter y1 is greater than 0, it is determined that the environmental sound signal is in a coherent noise scenario. If the result of the first scenario discrimination parameter y1 is less than 0, it is determined that the environmental sound signal is in an incoherent noise scenario;

[0041] Step D2, if it is determined that the environmental sound signal is in an incoherent noise scenario, then continue to calculate the second scenario discrimination parameter y2, the third scenario discrimination parameter y3, and the fourth scenario discrimination parameter y4. The specific steps are as follows.

[0042] Step D21, calculate the second scenario discrimination parameter y2, as shown in formula (9).

[0043] y2 = sum(abs(W2) 2 C2)+b2 (9)

[0044] Among them, C2 and b2 are both the corresponding spatial information recognition parameters. If the calculation result of the second scenario discrimination parameter y2 is greater than 0, it is determined that there is a wind noise scenario in the environmental sound signal;

[0045] Step D22, calculate the third scenario discrimination parameter y3, as shown in formula (10).

[0046] y3= sum(abs(W2 - W3) 2 C3)+b3 (10)

[0047] Among them, W3, C3, and b3 are all the corresponding spatial information recognition parameters. If the calculation result of the third scenario discrimination parameter y3 is greater than 0, it is determined that there is a howling scenario in the environmental sound signal;

[0048] Step D23, calculate the fourth scenario discrimination parameter y4, as shown in formula (11).

[0049] y4 = sum(abs(W2 - W4) 2 C4)+b4 (11)

[0050] Among them, W4, C4, and b4 are all the corresponding spatial information recognition parameters. If the calculation result of the fourth scenario discrimination parameter y4 is greater than 0, it is determined that the environmental sound signal is a wearer's own voice scenario.

[0051] The above-mentioned method for identifying and calibrating spatial information of a hearing aid based on dual microphones, step E, performs hearing aid adjustment processing on the corresponding spatial scene according to the obtained environmental sound signal, and completes the operation of identifying and calibrating the spatial information of the hearing aid. If the environmental sound signal is identified as a coherent noise scene, the microphone array is adjusted to the omnidirectional pickup signal of the main microphone signal s1 and the noise reduction depth of the noise reduction algorithm is reduced. If the environmental sound signal is identified as a non-coherent noise scene, the directional pickup signals of the main microphone signal s1 and the secondary microphone signal s2 are turned on and the noise reduction depth of the noise reduction algorithm is increased. If the environmental sound signal is identified as a scene with wind noise, the single microphone pickup signal of the secondary microphone signal s2 is used and the low-frequency compensation gain in the noise reduction algorithm is reduced. If the environmental sound signal is identified as a scene with a risk of howling, the high-frequency compensation gain in the noise reduction algorithm is reduced and the tracking speed of the howling suppression algorithm is increased. If the environmental sound signal is identified as the self-sound scene of the wearer, the low-frequency compensation gain in the noise reduction algorithm is reduced.

[0052] In the above-mentioned method for identifying and calibrating spatial information of a hearing aid based on dual microphones, the directional pickup output signal s3 in step E is shown in formula (12).

[0053] S3 = S2 + W2 S1 k - W1 S2 (1 - k) (12)

[0054] Wherein, S3 is the complex number in the frequency domain of the directional pickup output signal s3, and the larger the modulation amount M, the larger the parameter k, and the smaller the modulation amount M, the smaller the parameter k.

[0055] A system for identifying and calibrating spatial information of a hearing aid based on dual microphones includes a signal acquisition module, a modulation amount calculation module, a coherence coefficient calculation module, a spatial scene discrimination module, and a spatial scene adjustment module. The signal acquisition module is used to collect environmental sound signals by using a dual microphone array pickup module and obtain the main microphone signal and the secondary microphone signal, and then process the main microphone signal and the secondary microphone signal to obtain the larger value of the energy of each frequency point of the two microphone signals. The modulation amount calculation module is used to calculate the modulation amount of the larger value of the energy of each frequency point at each frequency point. The coherence coefficient calculation module is used to calculate the background noise coherence coefficient and the activation signal coherence coefficient in each state at each frequency point respectively by using the least mean square error method based on the modulation amount of the larger value of the energy of each frequency point at each frequency point. The spatial scene discrimination module is used to calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient with the respective scene parameters stored in the spatial information recognition module respectively to obtain the spatial scene corresponding to the environmental sound signal. The spatial scene adjustment module is used to perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the environmental sound signal and complete the operation of identifying and calibrating the spatial information of the hearing aid.

[0056] The beneficial effects of the present invention are as follows: A method and system for identifying and calibrating spatial information of a hearing aid based on dual microphones according to the present invention first use a dual-microphone array pickup module to collect environmental sound signals and obtain a main microphone signal and a secondary microphone signal, then process the main microphone signal and the secondary microphone signal and obtain the maximum energy values of each frequency point of the two microphone signals. Next, calculate the modulation amount of the maximum energy value of each frequency point at each frequency point, and then calculate the background noise coherence coefficient and the activation signal coherence coefficient of each frequency point and each state respectively based on the modulation amount of the maximum energy value of each frequency point at each frequency point by using the least mean square error method. Subsequently, calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient with the respective scene parameters stored in the spatial information recognition module to obtain the spatial scene corresponding to the environmental sound signal, and then perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the environmental sound signal to complete the spatial information recognition and calibration operation of the hearing aid. It effectively realizes that the hearing aid spatial information recognition and calibration method and system have the function of making full use of the spatial orientation information of the dual microphones of the hearing aid to identify different usage scenarios, and can refine and calibrate the algorithm architecture of the hearing aid according to the specific scenario to provide a better user experience. At the same time, the hearing aid can automatically identify the corresponding usage scenario through the spatial information of the dual microphones during the complex and changeable usage process, and then through the adaptive parameter adjustment module, automatically adjust the pickup method and various algorithm parameters so that there is no need for the user to manually adjust frequently, significantly improving the speech clarity and wearing comfort, and enhancing the user's communication confidence and quality of life in daily life and social occasions. Description of the Drawings

[0057] Figure 1 is a flowchart of a method for identifying and calibrating spatial information of a hearing aid based on dual microphones according to the present invention;

[0058] Figure 2 is a schematic diagram of the spatial information recognition process according to the present invention;

[0059] Figure 3 is a schematic diagram of the spatial information calibration process according to the present invention. Detailed Embodiments

[0060] The present invention will be further described below in conjunction with the accompanying drawings of the specification.

[0061] As Figure 1 shown, a method for identifying and calibrating spatial information of a hearing aid based on dual microphones according to the present invention includes the following steps:

[0062] Step A, use a dual-microphone array pickup module to collect environmental sound signals and obtain a main microphone signal s1 and a secondary microphone signal s2, and then process the main microphone signal s1 and the secondary microphone signal s2 to obtain the maximum energy values of each frequency point of the two microphone signals. The specific steps are as follows:

[0063] Step A1: Perform short-time Fourier transform on the main microphone signal s1 and the secondary microphone signal s2 respectively, and then decompose the main microphone signal s1 and the secondary microphone signal s2 into complex representations in the frequency domain, as shown in formula (1).

[0064] S1 = STFT(s1), S2 = STFT(s2) (1)

[0065] Where S1 is the complex number in the frequency domain of the main microphone signal s1, S2 is the complex number in the frequency domain of the secondary microphone signal s2, and STFT represents the short-time Fourier transform operator.

[0066] Step A2: Calculate the energy at each frequency point of the complex number S1 in the frequency domain of the main microphone signal s1 and the complex number S2 in the frequency domain of the secondary microphone signal s2, as shown in formula (2).

[0067] P1 = abs(S1) 2 , P2 = abs(S2) 2 (2)

[0068] Where P1 is the energy at each frequency point of the complex number S1, P2 is the energy at each frequency point of the complex number S2, and abs is to find the modulus of the complex number.

[0069] Step A3: Compare the magnitudes of the energy P1 at each frequency point of the complex number S1 and the energy P2 at each frequency point of the complex number S2 and obtain the larger value P3 of the energy at each frequency point, as shown in formula (3).

[0070] P3 = max(P1, P2) (3)

[0071] Where max is to find the maximum value between the two in the sequence.

[0072] As Figure 2 shown, in step B, calculate the modulation amount of the larger value of the energy at each frequency point at each frequency point. The specific steps are as follows.

[0073] Step B1: Calculate the short-time envelope of the signal energy by first-order smoothing calculation, as shown in formula (4).

[0074] P a = a1 P3+(1 - a1) P a (4)

[0075] Where P a is the short-time energy envelope at each frequency point, and a1 is the smoothing factor.

[0076] Step B2: For the short-time energy envelope P aPerform long-term minimum value statistics and obtain the long-term minimum value P b , as shown in Equation (5),

[0077] P b = a2 P a +(1 - a2) P b , when P b ≤ P a ;

[0078] P b = P a , when P b > P a (5)

[0079] where a2 is the smoothing factor;

[0080] Step B3, calculate the difference between the short-term energy envelope P a of each frequency point and the long-term energy minimum value P b and obtain the modulation amount M, as shown in Equation (6),

[0081] M = 10 lg(P a ) - 10 lg(P b ) (6);

[0082] Step B4, set the thresholds Th1 and Th2 for determining the energy activation state of each frequency point. If the modulation amount M is not greater than the threshold Th1, the signal at this frequency point is in the unactivated state. If the modulation amount M is not less than the threshold Th2, the signal at this frequency point is in the activated state. If the threshold Th1 < modulation amount M < threshold Th2, the signal at this frequency point is in the transition state.

[0083] Step C, based on the modulation amounts of the larger values of the energies at each frequency point, use the least mean square error method to calculate the background noise coherence coefficient and the activated signal coherence coefficient for each state at each frequency point. Specifically, use the least mean square error method to separately count the background noise coherence coefficient W1 and the activated signal coherence coefficient W2 for each state at each frequency point, as shown in Equation (7),

[0084] W1 = W1 + d1 (S2 - W1 S1) conj(S1), if M ≤ Th1;

[0085] W2 = W2 + d2 (S2 - W2 S1) conj(S1), if M ≥ Th2 (7)

[0086] Among them, d1 is the background noise update rate, and d2 is the activation signal update rate.

[0087] As Figure 3 shown, in step D, the background noise coherence coefficient and the activation signal coherence coefficient are respectively calculated and discriminated from each scene parameter stored in the spatial information recognition module to obtain the spatial scene corresponding to the environmental sound signal. The specific steps are as follows:

[0088] Step D1, calculate the first scene discrimination parameter y1 and determine whether the environmental sound signal is in a coherent noise scene, as shown in formula (8):

[0089] y1 = sum(abs(W1) 2 C1)+b1 (8)

[0090] Among them, sum is the sequence summation operator, and C1 and b1 are the corresponding spatial information recognition parameters. If the result of the first scene discrimination parameter y1 is greater than 0, it is determined that the environmental sound signal is in a coherent noise scene. If the result of the first scene discrimination parameter y1 is less than 0, it is determined that the environmental sound signal is in a non - coherent noise scene.

[0091] Step D2, if it is determined that the environmental sound signal is in a non - coherent noise scene, then continue to calculate the second scene discrimination parameter y2, the third scene discrimination parameter y3, and the fourth scene discrimination parameter y4. The specific steps are as follows:

[0092] Step D21, calculate the second scene discrimination parameter y2, as shown in formula (9):

[0093] y2 = sum(abs(W2) 2 C2)+b2 (9)

[0094] Among them, C2 and b2 are both the corresponding spatial information recognition parameters. If the calculation result of the second scene discrimination parameter y2 is greater than 0, it is determined that there is a wind noise scene in the environmental sound signal.

[0095] Step D22, calculate the third scene discrimination parameter y3, as shown in formula (10):

[0096] y3 = sum(abs(W2 - W3) 2 C3)+b3 (10)

[0097] Among them, W3, C3, and b3 are all the corresponding spatial information recognition parameters. If the calculation result of the third scene discrimination parameter y3 is greater than 0, it is determined that there is a whistling scene in the environmental sound signal.

[0098] Step D23: Calculate the fourth-scenario discrimination parameter y4 as shown in formula (11).

[0099] y4 = sum(abs(W2 - W4) 2 C4)+b4 (11)

[0100] Where W4, C4, and b4 are all corresponding spatial information recognition parameters. If the calculated result of the fourth-scenario discrimination parameter y4 is greater than 0, it is determined that the environmental sound signal is the wearer's own voice scenario.

[0101] Step E: Perform hearing aid adjustment processing according to the corresponding spatial scenario of the obtained environmental sound signal to complete the spatial information recognition and calibration operation of the hearing aid. If the environmental sound signal is recognized as the coherent noise scenario, adjust the microphone array to the omnidirectional pickup signal of the main microphone signal s1 and reduce the noise reduction depth of the noise reduction algorithm. If the environmental sound signal is recognized as the incoherent noise scenario, turn on the directional pickup signals of the main microphone signal s1 and the secondary microphone signal s2 and increase the noise reduction depth of the noise reduction algorithm. If the environmental sound signal is recognized as the wind noise scenario, use the single-microphone pickup signal of the secondary microphone signal s2 and reduce the low-frequency compensation gain in the noise reduction algorithm. If the environmental sound signal is recognized as having a risk of howling, reduce the high-frequency compensation gain in the noise reduction algorithm and increase the tracking speed of the howling suppression algorithm. If the environmental sound signal is recognized as the wearer's own voice scenario, reduce the low-frequency compensation gain in the noise reduction algorithm.

[0102] The directional pickup output signal s3 in Step E is as shown in formula (12).

[0103] S3 = S2 + W2 S1 k - W1 S2 (1 - k) (12)

[0104] Where S3 is the complex number in the frequency domain of the directional pickup output signal s3, and the larger the modulation amount M, the larger the parameter k, and the smaller the modulation amount M, the smaller the parameter k.

[0105] A hearing aid spatial information recognition and calibration system based on dual microphones, comprising a signal acquisition module, a modulation amount calculation module, a coherence coefficient calculation module, a spatial scene discrimination module, and a spatial scene adjustment module. The signal acquisition module is used to collect ambient sound signals by means of a dual microphone array pickup module to obtain a main microphone signal and a secondary microphone signal, and then process the main microphone signal and the secondary microphone signal to obtain the maximum energy value of each frequency point of the two microphone signals. The modulation amount calculation module is used to calculate the modulation amount of the maximum energy value of each frequency point at each frequency point. The coherence coefficient calculation module is used to calculate the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point respectively based on the modulation amount of the maximum energy value of each frequency point at each frequency point and by using the least mean square error method. The spatial scene discrimination module is used to calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient respectively from the various scene parameters stored in the spatial information recognition module to obtain the spatial scene corresponding to the ambient sound signal. The spatial scene adjustment module is used to perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the ambient sound signal and complete the spatial information recognition and calibration operation of the hearing aid.

[0106] In summary, for the hearing aid spatial information recognition and calibration method and system of the present invention, first, a dual microphone array pickup module is used to collect ambient sound signals to obtain a main microphone signal and a secondary microphone signal, and then the main microphone signal and the secondary microphone signal are processed to obtain the maximum energy value of each frequency point of the two microphone signals. Next, the modulation amount of the maximum energy value of each frequency point at each frequency point is calculated. Then, based on the modulation amount of the maximum energy value of each frequency point at each frequency point and by using the least mean square error method, the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point are calculated respectively. Subsequently, the background noise coherence coefficient and the activation signal coherence coefficient are calculated and discriminated respectively from the various scene parameters stored in the spatial information recognition module to obtain the spatial scene corresponding to the ambient sound signal. Then, hearing aid adjustment processing is performed according to the obtained spatial scene corresponding to the ambient sound signal and the spatial information recognition and calibration operation of the hearing aid is completed. It effectively realizes that the hearing aid spatial information recognition and calibration method and system have the function of fully utilizing the spatial orientation information of the dual microphones of the hearing aid to identify different usage scenarios, and can refine and calibrate the algorithm architecture of the hearing aid according to the specific scenario to provide a better user experience. At the same time, the hearing aid can automatically identify the corresponding usage scenario through the spatial information of the dual microphones during the complex and changeable usage process, and then through the adaptive parameter adjustment module, automatically adjust the pickup method and various algorithm parameters so that there is no need for the user to frequently manually adjust, significantly improving the speech clarity and wearing comfort, and enhancing the user's communication confidence and quality of life in daily life and social occasions.

[0107] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A method for identifying and calibrating spatial information of a hearing aid based on dual microphones, characterized in that: It includes the following steps: Step A: Use a dual-microphone array pickup module to collect ambient sound signals, obtain the main microphone signal s1 and the secondary microphone signal s2, and then process the main microphone signal s1 and the secondary microphone signal s2 to obtain the maximum energy value of each frequency point of the two microphone signals. The specific steps are as follows: Step A1: Perform short-time Fourier transform on the main microphone signal s1 and the secondary microphone signal s2 respectively, and then decompose the main microphone signal s1 and the secondary microphone signal s2 into complex representations in the frequency domain, as shown in formula (1): S1 = STFT(s1), S2 = STFT(s2) (1) Where S1 is the complex number in the frequency domain of the main microphone signal s1, S2 is the complex number in the frequency domain of the secondary microphone signal s2, and STFT represents the short-time Fourier transform operator; Step A2: Calculate the energy of each frequency point of the complex number S1 in the frequency domain of the main microphone signal s1 and the complex number S2 in the frequency domain of the secondary microphone signal s2, as shown in formula (2): P1 = abs(S1) 2 , P2 = abs(S2) 2 (2) Where P1 is the energy of each frequency point of the complex number S1, P2 is the energy of each frequency point of the complex number S2, and abs is to find the modulus of the complex number; Step A3: Compare the magnitudes of the energy P1 of each frequency point of the complex number S1 and the energy P2 of each frequency point of the complex number S2, and obtain the maximum energy value P3 of each frequency point, as shown in formula (3): P3 = max(P1, P2) (3) Where max is to take the maximum value between the two for the sequence; Step B: Calculate the modulation amount of the maximum energy value of each frequency point at each frequency point. The specific steps are as follows: Step B1: Calculate the short-time energy envelope of the first-order smoothed calculation signal energy, as shown in formula (4): P a = a1 * P3+(1 - a1)*P a (4) Among them, P a is the short-time energy envelope of each frequency point, and a1 is the smoothing factor; Step B2: Perform long-term minimum value statistics on the short-time energy envelopes P of each frequency point a and obtain the long-term minimum value P b , as shown in formula (5). P b = a2 * P a + (1 - a2) * P b when P b ≤ P a ; P b = P a When P b > P a (5) Where a2 is the smoothing factor; Step B3, calculate the short-time energy envelope P of each frequency point a and the difference from the minimum long-time energy P b to obtain the modulation amount M, as shown in formula (6). M = 10 * lg(P a ) - 10 * lg(P b )(6); Step B4: Set the threshold Th1 and threshold Th2 for judging the activation state of each frequency point energy. If the modulation amount M is not greater than the threshold Th1, the frequency point signal is in the unactivated state. If the modulation amount M is not less than the threshold Th2, the frequency point signal is in the activated state. If the threshold Th1 < modulation amount M < threshold Th2, the frequency point signal is in the transition state; Step C: Based on the modulation amount of the maximum energy value of each frequency point at each frequency point, use the least mean square error method to calculate the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point respectively. Specifically, use the least mean square error method to statistically calculate the background noise coherence coefficient W1 and the activation signal coherence coefficient W2 in each state of each frequency point, as shown in formula (7): W1 = W1 + d1 * (S2 - W1 * S1) * conj(S1), if M ≤ Th1; W2 = W2 + d2 * (S2 - W2 * S1) * conj(S1), if M ≥ Th2 (7) Where d1 is the background noise update rate and d2 is the activation signal update rate; Step D: Calculate and judge the background noise coherence coefficient and the activation signal coherence coefficient respectively with the respective scene parameters stored in the spatial information recognition module to obtain the spatial scene corresponding to the ambient sound signal; Step E: Perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the ambient sound signal to complete the spatial information recognition and calibration operation of the hearing aid.

2. The method for identifying and calibrating spatial information of a hearing aid based on dual microphones according to claim 1, wherein: Step D: Calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient against each scene parameter stored in the spatial information recognition module respectively to obtain the spatial scene corresponding to the environmental sound signal. The specific steps are as follows: Step D1: Calculate the first scene discrimination parameter y1 and determine whether the environmental sound signal is a coherent noise scene, as shown in formula (8): y1 = sum(abs(W1) 2 *C1) + b1(8) where sum is the sequence summation operator, and C1 and b1 are the corresponding spatial information recognition parameters. If the result of the first scene discrimination parameter y1 is greater than 0, it is determined that the environmental sound signal is in a coherent noise scene; if the result of the first scene discrimination parameter y1 is less than 0, it is determined that the environmental sound signal is in a non - coherent noise scene. Step D2: If it is determined that the environmental sound signal is in a non - coherent noise scene, continue to calculate the second scene discrimination parameter y2, the third scene discrimination parameter y3, and the fourth scene discrimination parameter y4. The specific steps are as follows: Step D21: Calculate the second scene discrimination parameter y2, as shown in formula (9): y2 = sum(abs(W2) 2 *C2)+b2(9) where C2 and b2 are the corresponding spatial information recognition parameters. If the calculation result of the second scene discrimination parameter y2 is greater than 0, it is determined that there is a wind noise scene in the environmental sound signal. Step D22: Calculate the third scene discrimination parameter y3, as shown in formula (10): y3 = sum(abs(W2 - W3) 2 *C3) + b3(10) where W3, C3, and b3 are the corresponding spatial information recognition parameters. If the calculation result of the third scene discrimination parameter y3 is greater than 0, it is determined that there is a howling scene in the environmental sound signal. Step D23: Calculate the fourth scene discrimination parameter y4, as shown in formula (11): y4 = sum(abs(W2 - W4) 2 *C4) + b4(11) where W4, C4, and b4 are the corresponding spatial information recognition parameters. If the calculation result of the fourth scene discrimination parameter y4 is greater than 0, it is determined that the environmental sound signal is the wearer's own voice scene.

3. A method for identifying and calibrating spatial information of a hearing aid based on dual microphones according to claim 2, characterized in that: Step E: Perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the environmental sound signal to complete the spatial information recognition and calibration operation of the hearing aid. If it is recognized that the environmental sound signal is a coherent noise scene, adjust the microphone array to the omnidirectional pickup signal of the main microphone signal s1 and reduce the noise reduction depth of the noise reduction algorithm; if it is recognized that the environmental sound signal is a non - coherent noise scene, turn on the directional pickup signals of the main microphone signal s1 and the secondary microphone signal s2 and increase the noise reduction depth of the noise reduction algorithm; if it is recognized that the environmental sound signal is a wind noise scene, use the single - microphone pickup signal of the secondary microphone signal s2 and reduce the low - frequency compensation gain in the noise reduction algorithm; if it is recognized that the environmental sound signal has a howling risk, reduce the high - frequency compensation gain in the noise reduction algorithm and increase the tracking speed of the howling suppression algorithm; if it is recognized that the environmental sound signal is the wearer's own voice scene, reduce the low - frequency compensation gain in the noise reduction algorithm.

4. A method for identifying and calibrating spatial information of a hearing aid based on dual microphones according to claim 3, characterized in that: The directional pickup output signal s3 in Step E is as shown in formula (12): S3 = S2 + W2 * S1 * k - W1 * S2 * (1 - k)(12) where S3 is the complex number in the frequency domain of the directional pickup output signal s3, and the larger the modulation amount M, the larger the parameter k; the smaller the modulation amount M, the smaller the parameter k.

5. A hearing aid spatial information recognition and calibration system based on dual microphones, wherein the spatial information recognition and calibration process of the spatial information recognition and calibration system is based on the spatial information recognition and calibration method according to any one of claims 1-4, and is characterized in that: It includes a signal acquisition module, a modulation amount calculation module, a coherence coefficient calculation module, a spatial scene discrimination module, and a spatial scene adjustment module. The signal acquisition module is used to pick up environmental sound signals by using a dual microphone array pickup module to obtain a main microphone signal and a secondary microphone signal, and then process the main microphone signal and the secondary microphone signal to obtain the maximum energy value of each frequency point of the two microphone signals. The modulation amount calculation module is used to calculate the modulation amount of the maximum energy value of each frequency point at each frequency point. The coherence coefficient calculation module is used to calculate the background noise coherence coefficient and the activation signal coherence coefficient in each state of each frequency point respectively based on the modulation amount of the maximum energy value of each frequency point at each frequency point and by using the least mean square error method. The spatial scene discrimination module is used to calculate and discriminate the background noise coherence coefficient and the activation signal coherence coefficient with the respective scene parameters stored in the spatial information recognition module respectively to obtain the spatial scene corresponding to the environmental sound signal. The spatial scene adjustment module is used to perform hearing aid adjustment processing according to the obtained spatial scene corresponding to the environmental sound signal and complete the spatial information recognition and calibration operation of the hearing aid.

Images