Adaptive, programmable signal processing and filtering for hearing aids

• US 5,027,410 A
• Filed: 11/10/1988
• Issued: 06/25/1991
• Est. Priority Date: 11/10/1988
• Status: Expired due to Fees

• Current Assignee: Wisconsin Alumni Research Foundation (University of Wisconsin System)
• Original Assignee: Wisconsin Alumni Research Foundation (University of Wisconsin System)
• Inventors: Cummins, Kenneth L., Hecox, Kurt E., Williamson, Malcolm J.
• Primary Examiner(s): Isen, Forester W.
• Application Number:
  US07/269,937
• Time in Patent Office: 957 Days
• Field of Search: 381/68.4, 381/68, 381/68.2, 381/71, 381/73.1, 381/94, 381/104, 381/106, 381/107, 381/99, 333/166, 333/167, 333/168, 333/173, 333/14, 328/167, 364/724.01, 364/724.08, 364/724.09
• US Class Current: 381/320
• CPC Class Codes: H03G 9/005 of digital or coded signals
• CPC Class Codes: H03G 9/025 frequency-dependent volume compression or expansion, e.g. multiple-band systems H03G9/10, H03G9/18 take precedence
• CPC Class Codes: H04R 2225/43 Signal processing in hearing aids to enhance the speech intelligibility
• CPC Class Codes: H04R 2430/03 Synergistic effects of band splitting and sub-band processing
• CPC Class Codes: H04R 25/453 electronically
• CPC Class Codes: H04R 25/505 using digital signal processing

Abstract

A hearing aid system utilizes digital signal processing to correct for the hearing deficit of a particular user and to maximize the intelligibility of the desired audio signal relative to noise. An analog signal from a microphone is converted to digital data which is operated on by a digital signal processor, with the output of the digital signal processor being converted back to an analog signal which is amplified and provided to the user. The digital signal processor includes a time varying spectral filter having filter coefficients which can be varied on a quasi-real time basis to spectrally shape the signal to match the hearing deficit of the user and to accommodate ambient signal and noise levels. The coefficients of the spectral filter are determined by estimating the energy in several frequency bands within the frequency range of the input signal, and using those energy estimates to calculate desired gains for the frequency bands and corresponding spectral filter coefficients. The spectral energy analysis may be carried out using pairs of high pass and low pass digital filters in cascade relation, with the output of each low pass filter being provided to the next pair of high pass and low pass filters. The rate at which output data is provided from the filters in each pair may be reduced from the sample rate of input data by one half for succeeding pairs of filters in the cascade to thereby reduce the computation time required.

Claims

1. A digital signal processing hearing aid system comprising:

(a) input means for providing an electrical input signal corresponding to a sound signal;

(b) analog to digital converter means for converting the input signal from the input means to input signal digital data at a selected sample rate;

(c) digital signal processing means for receiving the input signal digital data from the analog to digital converter means and providing processed output data, the digital signal processing means including;

(1) means for analyzing the input signal data and estimating the energy contained in a plurality of selected frequency bands which cover the range of frequency content of the input signal data, wherein the means for analyzing the input signal data includes a first high pass digital filter for passing the higher half of the input signal frequency range, a first low pass digital filter for passing the lower half of the input signal frequency range and providing a second frequency range signal data, a second high pass digital filter selected to pass the higher half of the second frequency range signal and a second low pass digital filter selected to pass the lower half of the second frequency range signal to provide a third frequency range signal, a third high pass digital filter selected to pass the higher half of the third frequency range signal and a third low pass digital filer selected to pass the lower half of the third frequency range signal to provide a fourth frequency range signal, the fourth high pass filter selected to pass the higher half of the fourth frequency range signal and a fourth low pass digital filter selected to pass the lower half of the fourth frequency range signal, and wherein the means for analyzing further includes means for estimating the energy in the data corresponding to the outputs of the first, second, third, and fourth high pass filters and the fourth low pass filter;

(2) gain calculation means for calculating a gain for each band which is a function of estimated energy in each band;

(3) digital spectral filter means, having a plurality of filter coefficients which are changeable, for receiving the input signal digital data and providing spectrally shaped output data for output from the digital signal processing means;

(4) coefficient calculation means for calculating the coefficients of the digital spectral filter means based on the gains for each band calculated by the gain calculation means, the coefficients of the digital spectral filter means being periodically changed to the new coefficients calculated by the coefficient calculation means;

(d) digital to analog converter means for converting the processed output data from the digital signal processing means to an analog signal; and

(e) means for converting the analog signal to a corresponding sound.

2. The hearing aid system of claim 1 wherein the means for estimating the energy in the frequency bands of the input signal data further comprises means for taking the absolute values of the filter outputs and means for low pass filtering these absolute values.

3. The hearing aid system of claim 1 wherein the gain calculation means also includes means for selecting the spectral shaping of the digital spectral filter means to provide spectral shaping to the signal data passed through the spectral filter to compensate for spectrally related deficiencies in the hearing of the user of the hearing aid.

4. The hearing aid system of claim 1 including means for delaying the input signal data provided to the spectral filter means to allow time for the coefficients of the spectral filter to be changed in response to the characteristic of the input signal at the time that the data is provided to the spectral filter.

5. The hearing aid system of claim 1 wherein each high pass filter and low pass filter is a three tap finite impulse response digital filter.

6. The hearing aid system of claim 5 wherein the filter coefficients of each high pass filter are -1, 2, -1 and the filter coefficients of each low pass filter are 1, 2, 1.

7. The hearing aid system of claim 1 wherein the computation of output values of the first high pass and low pass filters is performed at half the sample rate, the computation of output values for the second high pass and low pass filters is performed at one quarter the sample rate, the computation of output values for the third high pass and low pass filters is performed at one 150 of the sample rate, and the computation of output values for the fourth high pass and low pass filters is performed at 1/16 of the sample rate.

8. The hearing aid system of claim 1 wherein the gain calculation means within the digital signal processing means computes gains for each of the frequency bands in accordance with a gain function of the estimated energy including at least a section providing amplitude expansion for selected low level input signals, a linear amplification section for providing constant amplification for intermediate level input signals, and a section providing amplitude compression for high level input signals.

9. The hearing aid system of claim 8 wherein the digital signal processing means includes a programmable memory selectably loaded with processing variables adapted to an individual user.

10. The hearing aid system of claim 9 wherein the processing variables in the programmable memory include the slopes of the expansion section and the compression section for each of the frequency bands in the gain calculation means, and the digital signal data energy envelope positions of the knees at which the expansion section joins the linear amplification section and at which the linear amplification section joins the compression section.

11. The hearing aid system of claim 8 wherein the gain calculation means calculates the gains also as a function of desired base line gains for each of the bands so as to cause the digital input signal passed through the spectral filter means to be spectrally shaped in a manner selected to compensate for the spectral hearing deficit of a user.

12. The hearing aid system of claim 8 wherein the gain calculation means of the digital signal processing means estimates the level of noise as a function of the energy in each of the bands and wherein a knee dividing the expansion section of the nonlinear gain function from the linear section for each band is moved higher or lower as a function of the noise level estimate for that band.

13. The hearing aid system of claim 9 wherein the programmable memory in the digital signal processing means is provided with parameters that define the minimum and maximum values for the position of the knee dividing the expansion section of the gain function from the linear section for each band and wherein the digital signal processing means includes means for estimating the level of noise as a function of the energy estimate of the digital input signal data for each frequency band, and wherein the digital signal processing means moves the knee higher or lower as a function of the noise level estimate but within the limits set by the minimum and maximum values.

14. The hearing aid system of claim 1 including automatic gain control means for controlling the magnitude of the analog signal passed from the input means to the analog to digital converter means, the signal magnitude controlled at a slow rate relative to speech to be within a desired range of magnitudes.

15. The hearing aid system of claim 8 wherein each of the gain functions used in each of the frequency bands by the gain calculation means has a piecewise linear gain function in which the logarithm of the gain is a linearly increasing function of the logarithm of the energy estimate in each band up to a first knee, a constant between the first and a second knee, and a linearly decreasing function above the second knee.

16. The hearing aid system of claim 15 wherein the positions of the first and the second knees are selected so that the constant gain between the first and second knee lies over a preferred dynamic range for the hearing of the individual user in each of the frequency bands.

17. The hearing aid system of claim 15 wherein the gain calculation means includes means for estimating the level of noise as a function of the energy estimate of the digital input signal data in each frequency band and wherein the first knee in each gain function in each frequency band is moved higher or lower as a function of the noise level estimate for that band while the position of the second knee and the slopes of the gain function remain fixed in each band.

18. The hearing aid system of claim 17 wherein the position of the first knee is set equal to the noise level estimate plus a selected constant.

19. The hearing aid system of claim 17 wherein the first knee is moved higher or lower in direct relation to the noise level estimate but no higher than a selected maximum value and no lower than a selected minimum value in determining the gain calculated by the gain calculation means for each frequency band in the digital signal processing means.

20. The hearing aid system of claim 19 wherein the digital signal processor has a programmable memory and wherein the programmable memory is provided with parameters that define the minimum and maximum values of the position of the first knee, the position of the second knee, and the slopes of the gain function.

21. The hearing aid system of claim 17 wherein the means for estimating the level of noise in each frequency band in the gain calculations means estimates the noise level by selecting a percentile of the distribution of the energy estimate for each of the frequency bands.

22. The hearing aid system of claim 19 wherein the position of the first knee is set equal to the noise level estimate plus a selected constant.

23. The hearing aid system of claim 1 including pre-emphasis filtering means for equalizing the frequency spectrum of the analog signal passed from the input means to the analog to digital converter means to minimize the required dynamic range of the analog to digital converter means.

24. The hearing aid system of claim 1 wherein the gain calculation means in the digital signal processing means determines the gain function for each frequency band periodically, and the coefficient calculation means calculates new filter coefficients periodically based on the new gains, with a time constant in the range of about 1 millisecond to about 2 milliseconds.

25. The hearing aid system of claim 8 wherein for each band the section of the gain function up to the first knee provides an expansion ratio of up to 1:

2 and a section above the second knee provides a compression ratio of up to 3.3;

1.

26. A digital signal processing hearing aid system comprising:

(a) input means for providing an electrical signal corresponding to a sound signal;

(b) analog to digital converter means for converting the signal from the input means to digital data at a selected sample rate;

(c) digital signal processing means for receiving the input signal digital data from the analog to digital converter means and providing processed output data, the digital signal processing means including;

(1) N pairs of high pass and low pass digital filters, the high pass filter in each pair selected to pass the highest half of the frequency range of the input signal to the pair and the low pass filter in each pair selected to pass the lowest half of the frequency range of the input signal to the pair, the pairs of filters cascaded so that the input signal to each pair after the first pair is the output of the low pass filter of the prior pair of filters, and means for estimating the energy in the frequency bands corresponding to the outputs of the N high pass filters and the last low pass filter, where N is at least two;

(2) gain calculation means for calculating a gain for each band which is a function of the estimated energy in each band;

(3) digital spectral filter means, having a plurality of filter coefficients which are changeable, for receiving the input signal data and providing spectrally shaped output data for output from the digital signal processing means;

(4) coefficient calculation means for calculating the coefficients of the digital spectral filter means based on the gains for each band calculated by the gain calculation means, the coefficients of the digital spectral filter means being periodically changed to the new coefficients calculated by the coefficient calculation means;

(d) digital to analog converter means for converting the processed output data from the digital signal processing means to an analog signal; and

(e) means for converting the analog signal to a corresponding sound.

27. The hearing aid system of claim 26 wherein N=4.

28. The hearing aid system of claim 26 wherein the means for estimating the energy of the output signal data of the N high pass filters and the last low pass filter comprises means for taking the absolute values of the filter outputs and means for low pass filtering these absolute values.

29. The hearing aid system of claim 26 wherein the gain calculation means also includes means for selecting the spectral shaping of the digital spectral filter means to provide spectral shaping to the signal data passed through the spectral filter to compensate for spectrally related deficiencies in the hearing of the user of the hearing aid.

30. The hearing aid system of claim 26 including means for delaying the input signal data provided to the spectral filter means to allow time for the coefficients of the spectral filter to be changed in response to the characteristic of the input signal at the time that the data is provided to the spectral filter.

31. The hearing aid system of claim 26 wherein each high pass filter and low pass filter is a three tap finite impulse response digital filter.

32. The hearing aid system of claim 31 wherein the filter coefficients of each high pass filter are -1, 2, -1 and the filter coefficients of each low pass filter are 1, 2, 1.

33. The hearing aid system of claim 26 wherein the computation of output values of the first pair of high pass and low pass filters is performed at half the sample rate, and the computation of output values for the succeeding pairs of high pass and low pass filters is performed at rate which is a fraction of the sample rate equal to 1/2ⁿ⁺¹, where n is the number of prior filter pairs.

34. The hearing aid system of claim 26 wherein there are four pairs of high pass and low pass filters, with the output of the first high pass filter passing the portion of the input signal from approximately 4,000 Hz to 8,000 Hz, the second high pass filter passing the portion of the input signal data in the range of 2,000 Hz to 4,000 Hz, the third high pass filter passing the portion of the signal data in the range of 1,000 Hz to 2,000 Hz, the fourth high pass filter passing the signal data in the range of 500 Hz to 1000 Hz, and the fourth low pass filter passing the portion of the input signal data in the range from 0 to 500 Hz.

35. The hearing aid system of claim 26 wherein the gain calculation means within the digital signal processing means computes gains for each of the frequency bands in accordance with a gain function of the estimated energy including at least a section providing amplitude expansion for selected low level input signals, a linear amplification section for providing constant amplification for intermediate level input signals, and a section providing amplitude compression for high level input signals.

36. The hearing aid system of claim 35 wherein the digital signal processing means includes a programmable memory selectably loaded with processing variables adapted to an individual user.

37. The hearing aid system of claim 36 wherein the processing variables in the programmable memory include the slopes of the expansion section and the compression section for each of the frequency bands in the gain calculation means, and the digital signal data energy envelope positions of the knees at which the expansion section joins the linear amplification section and at which the linear amplification section joins the compression section.

38. The hearing aid system of claim 35 wherein the gain calculation means calculates the gains also as a function of desired base line gains for each of the bands so as to cause the digital input signal passed through the spectral filter means to be spectrally shaped in a manner selected to compensate for the spectral hearing deficit of a user.

39. The hearing aid system of claim 35 wherein the gain calculation means of the digital signal processing means estimates the level of noise as a function of the energy in each of the bands and wherein a knee dividing the expansion section of the nonlinear gain function from the linear section for each band is moved higher or lower as a function of the noise level estimate for that band.

40. The hearing aid system of claim 36 wherein the programmable memory in the digital signal processing means is provided with parameters that define the minimum and maximum values for the position of the knee dividing the expansion section of the gain function from the linear section for each band and wherein the digital signal processing means includes means for estimating the level of noise as a function of the energy estimate of the digital input signal data for each frequency band, and wherein the digital signal processing means moves the knee higher or lower as a function of the noise level estimate but within the limits set by the minimum and maximum values.

41. The hearing aid system of claim 26 including automatic gain control means for controlling the magnitude of the analog signal passed from the input means to the analog to digital converter means, the signal magnitude controlled at a slow rate relative to speech to be within a desired range of magnitudes.

42. The hearing aid system of claim 35 wherein each of the gain functions used in each of the frequency bands by the gain calculation means has a piecewise linear gain function in which the logarithm of the gain is a linearly increasing function of the logarithm of the energy estimate in each band up to a first knee, a constant between the first and a second knee, and a linearly decreasing function above the second knee.

43. The hearing aid system of claim 42 wherein the positions of the first and the second knees are selected so that the constant gain between the first and second knee lies over a preferred dynamic range for the hearing of the individual user in each of the frequency bands.

44. The hearing aid system of claim 42 wherein the gain calculation means includes means for estimating the level of noise as a function of the energy estimate of the digital input signal data in each frequency band and wherein the first knee in each gain function in each frequency band is moved higher or lower as a function of the noise level estimate for that band while the position of the second knee and the slopes of the gain function remained fixed in each band.

45. The hearing aid system of claim 44 wherein the position of the first knee is set equal to the noise level estimate plus a selected constant.

46. The hearing aid system of claim 44 wherein the first knee is moved higher or lower in direct relation to the noise level estimate but no higher than a selected maximum value and no lower than a selected minimum value in determining the gain calculated by the gain calculation means for each frequency band in the digital signal processing means.

47. The hearing aid system of claim 46 wherein the digital signal processor has a programmable memory and wherein the programmable memory is provided with parameters that define the minimum and maximum values of the position of the first knee, the position of the second knee, and the slopes of the gain function.

48. The hearing aid system of claim 44 wherein the means for estimating the level of noise in each frequency band in the gain calculations means estimates the noise level by selecting a percentile of the distribution of the energy estimate for each of the frequency bands.

49. The hearing aid system of claim 46 wherein the position of the first knee is set equal to the noise level estimate plus a selected constant.

50. The hearing aid system of claim 26 including pre-emphasis filtering means for equalizing the frequency spectrum of the analog signal passed from the input means to the analog to digital converter means to minimize the required dynamic range of the analog to digital converter means.

51. The hearing aid system of claim 26 wherein the gain calculation means in the digital signal processing means determines the gain function for each frequency band periodically, and the coefficient calculation means calculates new filter coefficients periodically based on the new gains, with a time constant in the range of about 1 millisecond to about 2 milliseconds.

52. The hearing aid system of claim 35 wherein the section of the gain function up to the first knee provides an expansion ratio of up to 1:

2 and a section above the second knee provides a compression ratio of up to 3.3;

1.

53. A digital spectral analysis filter comprising:

(a) input means for providing an electrical signal corresponding to a sound signal;

(b) analog to digital converter means for converting the signal from the input means to digital data at a selected sample rate;

(c) digital signal processing means for receiving the input signal digital data from the analog to digital converter means and providing processed output data, the digital signal processing means including N pairs of high pass and low pass digital filters, the high pass filter in each pair selected to pass the highest half of the frequency range of the input signal to the pair and the low pass filter in each pair selected to pass the lowest half of the frequency range of the input signal to the pair, the pairs of filters cascaded so that the input signal to each pair after the first pair is the output of the low pass filter of the prior pair of filters, where N is at least two, and wherein the computation of output values of the first pair of high pass and low pass filters is performed at half the sample rate and the computation of output values for succeeding pairs of high pass and low pass filters is performed at a rate which is a fraction of the sample rate equal to 1/2ⁿ⁺¹, where n is the number of prior filter pairs.

54. The digital spectral analysis filter of claim 53 further including means for estimating the energy in the outputs of the N high pass filters and the last low pass filter.

55. The digital spectral analysis filter of claim 53 wherein each high pass filter and low pass filter is a three tap finite impulse response digital filter.

56. The digital spectral analysis filter of claim 53 wherein the filter coefficients of each high pass ilter are -1, 2, -1 and the filter coefficients of each low pass filter 1, 2, 1.

57. The digital spectral analysis filter of claim 53 wherein the frequency range of interest is from 0 to 8,000 Hz, N=4, and the first high pass filter passes the portion of the input signal from approximately 4,000 Hz to 8,000 Hz, the second high pass filter passes the portion of the input signal data in the range of 2,000 Hz to 4,000 Hz, the third high pass filter passes the portion of the signal data in the range of 1000 to 2000 Hz, the fourth high pass filter passes the portion of the signal data in the range of 500 Hz to 1,000 Hz, and the fourth low pass filter passes the portion of the input signal data in the range of 0 to 500 Hz.

58. The digital spectral analysis filter of claim 53 wherein the means for estimating the energy of the filter output signal data comprises means for taking the absolute values of the filter outputs and means for low pass filtering these absolute values.