A method for generating tinnitus therapeutic sound based on notch and audio equalization

Abstract

The purpose of the present invention is to provide a method for generating tinnitus therapeutic sounds based on notching and audio equalization, which redistributes audio energy before enhancing and notching the audio, thereby ensuring both a lateral inhibition effect and a pleasant sound, thereby providing patients with a better treatment experience and ensuring efficacy, and is characterized in that it includes the following steps: Step 1, obtaining the type of the patient's tinnitus sound and the tinnitus center frequency f; Step 2, determining the cut-off frequency band range according to the type of the tinnitus sound; Step 3, selecting the original audio of the therapeutic sound; Step 4, performing energy equalization on the original audio of the therapeutic sound; Step 5, enhancing the therapeutic sound after equalization in Step 4; Step 6, performing notching processing on the therapeutic sound after enhancement in Step 5.

Description

Technical Field

[0001] The present invention belongs to the technical field related to sound data production, and in particular relates to a method for generating tinnitus therapeutic sounds based on notching and audio equalization. Background Art

[0002] Prior art CN113380215A discloses a method for generating notched music for tinnitus treatment, which obtains the central frequency of the tinnitus sound for each patient to obtain an enhanced frequency interval, enhances the data in the natural sound data located in the enhanced frequency interval, and then performs notching processing on the enhanced natural sound data to ensure personalized notched music. The notched music obtained based on the above method can effectively treat both high-frequency and low-frequency tinnitus.

[0003] The above method is to directly enhance the audio in the high-frequency part and then make a notch. However, for many audios with low energy in the high-frequency part, the energy of the audio may still be very low even after enhancing it by 20dB, so the lateral suppression effect is not good. In addition, if the energy of the high-frequency part of the audio is very high, the sound may not be pleasant enough after enhancing it by 20dB, which affects the treatment experience. Summary of the invention

[0004] The purpose of the present invention is to provide a method for generating tinnitus therapeutic sounds based on notching and audio equalization to solve the problems raised in the above-mentioned background technology. It redistributes the audio energy before enhancing and notching the audio, which not only ensures the lateral inhibition effect, but also ensures the pleasant sound, so that patients have a better treatment experience and ensures the efficacy.

[0005] To achieve the above object, the present invention provides the following technical solutions.

[0006] A method for generating tinnitus therapeutic sound based on notch and audio equalization, characterized in that it comprises the following steps:

[0007] Step 1, obtaining the type of tinnitus sound and the tinnitus center frequency f of the patient;

[0008] Step 2, determining the cut-off frequency band range according to the type of the tinnitus sound;

[0009] Step 3, select the original audio of the therapeutic sound;

[0010] Step 4, performing energy equalization on the original audio of the therapeutic sound;

[0011] Step 5, enhancing the therapeutic sound after equalization in step 4;

[0012] Step 6: Perform notching processing on the therapeutic sound enhanced in step 5.

[0013] Furthermore, the step 1 is specifically as follows:

[0014] The types of the tinnitus sounds include pure tones and narrowband white noise;

[0015] When the type of the tinnitus sound is a pure tone, the tinnitus center frequency f is the frequency of the pure tone; when the type of the tinnitus sound is narrowband white noise, the tinnitus center frequency f is the square root of the product of the lower frequency limit fl and the upper frequency limit fu of the narrowband white noise, where fl < f < fu.

[0016] Further, step 2 is specifically as follows:

[0017] The excision frequency band range is a continuous frequency interval between the excision frequency band lower limit frequency mfl and the excision frequency band upper limit frequency mfu;

[0018] The value range of the excision frequency band range is: when the type of the tinnitus sound is a pure tone, mfl < f < mfu, f / 2^(1 / 2) < mfl < f / 2^(1 / 8), f*2^(1 / 8) < mfu < f*2^(1 / 2); when the type of the tinnitus sound is narrowband white noise, mfl < fl < fu < mfu, fl / 2^(1 / 2) < mfl < fl / 2^(1 / 8), fu*2^(1 / 8) < mfu < fu*2^(1 / 2), and the product of mfl and mfu is always equal to the square of f.

[0019] Further, step 4 is specifically as follows:

[0020] First, calculate the total energy e of the original treatment sound audio;

[0021] Then, perform frequency division processing on the original treatment sound audio using a bank of band-pass filters. The number of filters in the bank of band-pass filters is n, and each filter satisfies the following characteristics: the ratio of the upper cut-off frequency to the lower cut-off frequency of the filter is r, and the lower cut-off frequency of each filter is the same as the upper cut-off frequency of the previous filter;

[0022] After the frequency division processing, calculate the audio energy ei (i = 1, 2,..., n) filtered by each filter, adjust the audio energy ei filtered by each filter to 1 / n of the total energy e of the original treatment sound audio, and then superimpose the audio filtered by each filter to obtain a treatment sound with balanced energy;

[0023] When the tinnitus sound is a pure tone, r < 2^(1 / 8); when the tinnitus sound is narrowband white noise, r < (fu / fl)^0.25;

[0024] The ratio r of the upper cut-off frequency to the lower cut-off frequency of the filter, the number n of filters, and the relationship between the upper cut-off frequency ffu and the lower cut-off frequency ffl of the band-pass filter bank are: ffu / ffl = r^n; where ffu > 8000 Hz and ffl < 125 Hz.

[0025] Further, step 5 is specifically as follows:

[0026] The frequency band near the cut-off frequency band range is an enhanced frequency band, and the enhancement amplitude is a; for the therapeutic sound equalized in step 4, a band-pass filter is used to obtain the audio s(t) within the enhanced frequency band, a band-stop filter is used to obtain the audio o(t) outside the enhanced frequency band, then s(t) is enhanced to u(t) = s(t) * 10^(a / 10), and finally u(t) and o(t) are superimposed to obtain the enhanced therapeutic sound m(t) = u(t) + o(t); where m(t), u(t), o(t), and s(t) are all time-domain expressions.

[0027] The lower cut-off frequency of the enhanced frequency band is afl = mfl / 2^s, the upper cut-off frequency of the enhanced frequency band is afu = mfu * 2^s, and the value range of s is generally 0.01 < s < 0.99; the enhancement amplitude a is generally 5 dB - 20 dB.

[0028] Further, step 6 is specifically as follows:

[0029] According to the cut-off frequency band range, a band-stop filter is used to filter the enhanced therapeutic sound in step 5 to obtain the final therapeutic sound.

[0030] Compared with the prior art, the present invention provides a method for generating a tinnitus treatment sound based on notch and audio equalization, having the following beneficial effects:

[0031] The present invention equalizes the energy of the tinnitus treatment sound music on the logarithmic frequency axis before enhancement and notch, that is, the energy of the audio within each octave is equal. In this way, when the energy of the audio is low in the high-frequency part, after the energy is redistributed and then notch is performed, a good lateral inhibition effect can be obtained. When the energy is high in the high-frequency part, after the energy is redistributed and then notch is performed, it will not cause the sound to be unpleasant, which conforms to the auditory habit of the human ear, so that the patient has a better treatment experience and also takes into account the treatment effect. Description of the Drawings

[0032] Figure 1 It is a spectrum relationship diagram of the original audio and the tinnitus sound;

[0033] Figure 2 It is a spectrum relationship diagram of the equalized therapeutic sound and the tinnitus sound;

[0034] Figure 3Spectrum relationship diagram of enhanced treatment sound and tinnitus sound;

[0035] Figure 4 Spectrum relationship diagram of the final treatment sound and tinnitus sound. Detailed implementation manners

[0036] A method for generating tinnitus treatment sound based on notch and audio equalization, comprising the following steps:

[0037] Step 1: Obtain the type of the patient's tinnitus sound and the tinnitus center frequency f.

[0038] The type of the tinnitus sound includes pure tone and narrowband white noise.

[0039] When the type of the tinnitus sound is pure tone, the tinnitus center frequency f is the frequency of the pure tone; when the type of the tinnitus sound is narrowband white noise, the tinnitus center frequency f is the logarithmic center of the upper and lower limit frequencies of the narrowband white noise, that is, the square root of the product of the frequency lower limit fl of the narrowband white noise and the frequency upper limit fu of the narrowband white noise, where fl < f < fu.

[0040] Use the existing tinnitus detector to detect the type of the patient's tinnitus sound and the tinnitus center frequency. Repeatedly adjust the device center frequency and the sound type to enable the patient to find the sound closest to their tinnitus sound as the approximate tinnitus sound.

[0041] Step 2: Determine the notch frequency band range according to the type of the tinnitus sound.

[0042] The notch frequency band range is a continuous frequency interval between the notch frequency band lower limit frequency mfl and the notch frequency band upper limit frequency mfu.

[0043] The notch frequency band range is taken as: when the type of the tinnitus sound is pure tone, mfl < f < mfu, f / 2^(1 / 2) < mfl < f / 2^(1 / 8), f*2^(1 / 8) < mfu < f*2^(1 / 2); when the type of the tinnitus sound is narrowband white noise, mfl < fl < fu < mfu, fl / 2^(1 / 2) < mfl < fl / 2^(1 / 8), fu*2^(1 / 8) < mfu < fu*2^(1 / 2), and the product of mfl and mfu is always equal to the square of f.

[0044] Step 3: Select the original audio of the treatment sound; in a preferred example, the relationship between the original audio of the treatment sound and the tinnitus sound is as shown in the appendix Figure 1 as follows.

[0045] Step 4: Equalize the energy of the original audio of the treatment sound.

[0046] First, calculate the total energy e of the original audio of the treatment sound.

[0047] Then, a band - pass filter bank is used to perform frequency division processing on the original audio of the therapeutic sound. The number of filters in the band - pass filter bank is n, and each filter satisfies the following characteristics: the ratio of the upper cut - off frequency to the lower cut - off frequency of the filter is r, and the lower cut - off frequency of each filter is the same as the upper cut - off frequency of the previous filter.

[0048] After frequency division processing, the audio energy ei (i = 1, 2,..., n) filtered by each filter is calculated. The audio energy ei filtered by each filter is adjusted to 1 / n of the total energy e of the original audio of the therapeutic sound, and then the audio filtered by each filter is superimposed to obtain the therapeutic sound with balanced energy. The spectral relationship between the balanced therapeutic sound and the tinnitus sound is as shown in the appendix. Figure 2 Calculating audio energy is a conventional technique in the art and will not be elaborated here.

[0049] When the tinnitus sound is a pure tone, r < 2^(1 / 8); when the tinnitus sound is narrow - band white noise, r < (fu / fl)^0.25, and r can take any value within the above range. The relationship between the ratio r of the upper cut - off frequency to the lower cut - off frequency of the filter, the number of filters n, and the upper cut - off frequency ffu and the lower cut - off frequency ffl of the band - pass filter bank is: ffu / ffl = r^n; where ffu > 8000Hz, ffl < 125Hz, and ffu and ffl can take any value within the above range.

[0050] Step 5: Enhance the therapeutic sound balanced in step 4.

[0051] The frequency band near the cut - off frequency band range is the enhancement frequency band, and the enhancement amplitude is a. For the therapeutic sound balanced in step 4, a band - pass filter is used to obtain the audio s(t) within the enhancement frequency band, and a band - stop filter is used to obtain the audio o(t) outside the enhancement frequency band. Then, s(t) is enhanced to u(t)=s(t)*10^(a / 10), and finally, u(t) and o(t) are superimposed to obtain the enhanced therapeutic sound m(t)=u(t)+o(t); where m(t), u(t), o(t), and s(t) are all time - domain expressions. The spectral relationship between the enhanced therapeutic sound and the tinnitus sound is as shown in the appendix. Figure 3 as shown.

[0052] The lower cut - off frequency of the enhancement frequency band is afl = mfl / 2^s, the upper cut - off frequency of the enhancement frequency band is afu = mfu*2^s, and the value range of s is generally 0.01 < s < 0.99; the enhancement amplitude a is generally 5dB - 20dB. In specific implementation, the enhancement amplitude a is generally set to 5dB, specifically, the amplitude of the time - domain signal within the frequency band is increased by 10^0.5 times, and s is generally set to 1 / 6.

[0053] In step 5, the lower cutoff frequency of the bandpass filter and the bandstop filter is the lower limit frequency afl of the enhanced frequency band, and the upper cutoff frequency of the bandpass filter and the bandstop filter is the upper limit frequency afu of the enhanced frequency band.

[0054] Step 6: Perform notching processing on the therapeutic sound enhanced in step 5.

[0055] According to the cut-off frequency range, the enhanced therapeutic sound in step 5 is filtered using a band-stop filter to obtain the final therapeutic sound. The frequency spectrum relationship between the final therapeutic sound and the tinnitus sound is shown in the attached figure. Figure 4 shown.

[0056] In step 6, the lower cutoff frequency of the band stop filter is the lower limit frequency mf1 of the cut-off frequency band, and the upper cutoff frequency of the band stop filter is the upper limit frequency mf of the cut-off frequency band.

[0057] The present invention customizes tinnitus treatment sounds suitable for each patient according to the tinnitus sound type and tinnitus frequency of each patient. The tinnitus treatment sound music is first energy balanced on the logarithmic axis according to the frequency before being enhanced and notched, that is, the energy of the audio is equal in each octave. In this way, when the energy of the high-frequency part of the audio is low, the energy can be redistributed before notching, which can obtain a good lateral inhibition effect. When the energy of the high-frequency part is high, the energy can be redistributed before notching, which will not cause unpleasant sounds and conform to the auditory habits of the human ear, thereby allowing patients to have a better treatment experience while taking into account the treatment effect.

Claims

1. A method for generating tinnitus therapeutic sounds based on notch and audio equalization, characterized in that: It includes the following steps: Step 1, obtain the type of the patient's tinnitus sound and the tinnitus center frequency f; Step 2, determine the excision frequency band range according to the type of the tinnitus sound; Step 3, select the original audio of the treatment sound; Step 4, perform energy equalization on the original audio of the treatment sound; Step 5, enhance the treatment sound equalized in Step 4; Step 6, perform notch processing on the treatment sound enhanced in Step 5; The specific content of Step 1 is: The types of the tinnitus sound include pure tone and narrowband white noise; When the type of the tinnitus sound is pure tone, the tinnitus center frequency f is the frequency of the pure tone. When the type of the tinnitus sound is narrowband white noise, the tinnitus center frequency f is the square root of the product of the lower frequency limit fl and the upper frequency limit fu of the narrowband white noise, where fl < f < fu; The specific content of Step 2 is: The excision frequency band range is a continuous frequency interval between the lower excision frequency limit mfl and the upper excision frequency limit mfu; The value of the excision frequency band range is: when the type of the tinnitus sound is pure tone, mfl < f < mfu, f / 2^(1 / 2) < mfl < f / 2^(1 / 8), f*2^(1 / 8) < mfu < f*2^(1 / 2); when the type of the tinnitus sound is narrowband white noise, mfl < fl < fu < mfu, fl / 2^(1 / 2) < mfl < fl / 2^(1 / 8), fu*2^(1 / 8) < mfu < fu*2^(1 / 2), and the product of mfl and mfu is always equal to the square of f; The specific content of Step 4 is: First, calculate the total energy e of the original audio of the treatment sound; Then, use a bank of band-pass filters to perform frequency division processing on the original audio of the treatment sound. The number of filters in the bank of band-pass filters is n, and each filter satisfies the following characteristics: the ratio of the upper cut-off frequency to the lower cut-off frequency of the filter is r, and the lower cut-off frequency of each filter is the same as the upper cut-off frequency of the previous filter; After frequency division processing, calculate the audio energy ei (i = 1, 2,..., n) filtered by each filter, adjust the audio energy ei filtered by each filter to 1 / n of the total energy e of the original audio of the treatment sound, and then superimpose the audio filtered by each filter to obtain the treatment sound with energy equalization; When the tinnitus sound is pure tone, r < 2^(1 / 8). When the tinnitus sound is narrowband white noise, r < (fu / fl)^0.25; the relationship between the ratio r of the upper cut-off frequency to the lower cut-off frequency of the filter, the number of filters n, and the upper cut-off frequency ffu and the lower cut-off frequency ffl of the bank of band-pass filters is: ffu / ffl = r^n; where ffu > 8000 Hz and ffl < 125 Hz.

2. According to the method for generating a tinnitus treatment sound based on notch and audio equalization as described in claim 1, the specific content of Step 5 is: The frequency band near the excision frequency band range is the enhancement frequency band, and the enhancement amplitude is a; for the therapeutic sound after equalization in step 4, a band-pass filter is used to obtain the audio s(t) within the enhancement frequency band, and a band-stop filter is used to obtain the audio o(t) outside the enhancement frequency band. Then, s(t) is enhanced to u(t)=s(t)*10^(a / 10), and finally, u(t) and o(t) are superimposed to obtain the enhanced therapeutic sound m(t)=u(t)+o(t); where m(t), u(t), o(t), and s(t) are all time-domain expressions. The lower limit frequency of the enhancement frequency band is afl=mfl / 2^s, and the upper limit frequency of the enhancement frequency band is afu=mfu*2^s. The value range of s is generally 0.01 < s < 0.99; the enhancement amplitude a is generally 5dB - 20dB.

3. A tinnitus therapeutic sound generation method based on notch and audio equalization according to claim 2, wherein step 6 is specifically: According to the excision frequency band range, a band-stop filter is used to filter the enhanced therapeutic sound in step 5 to obtain the final therapeutic sound.

Images