JP2001128296A - Hearing aid - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a hearing aid by which a user can clearly hear a voice of a hearing object independently of the quality of ambient noise. SOLUTION: In the hearing aid consisting of an external sound detection section 1 that outputs an electric signal in response to an external sound, a signal processing section 2 that processes the output signal of the external sound detection section 1, and an output section 3 that transduces the output signal of the signal processing section 2 into a sound signal, the signal processing section 2 is provided with steadiness detection means 2b-2g that detect a steadiness of a noise included in the external sound from the signal outputted from the external sound detection section 1 at a prescribed time interval, an input output characteristic processing section 2a that compresses and amplifies the output signal of the external sound detection section 1, and an input output characteristic table 2i that stores compression amplification characteristic parameters suitable for hearing aid wearing people, and the input output characteristic of the input output characteristic processing section 2a is decided based on the compression amplification characteristic parameter corresponding to a steady-state degree outputted from the steady-state degree detection means 2b-2g.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hearing aid which can clearly hear a voice in consideration of the quality of environmental noise.

[0002]

2. Description of the Related Art Noise present in a use environment has a great influence on hearing of a hearing aid user and comfort of output sound. to solve this problem,
A so-called non-linear amplification hearing aid (non-linear hearing aid) has been proposed in which an automatic gain limiting circuit (AGC circuit) is provided in a hearing aid, and the automatic gain limiting circuit appropriately adjusts an acoustic gain or the like according to the level of an input signal.

A non-linear amplification hearing aid has an input / output characteristic having a large gain for a small sound and a small gain for a large sound, so that a small input level existing outside the audible range of a hearing-impaired person can be reduced within the audible range. It is said that the purpose is to reduce the unpleasant loud noise without amplifying it, and to reduce the discomfort of the user in a noisy environment in many cases.

[0004]

However, the above-mentioned nonlinear amplification hearing aid is easy to listen to when there is environmental noise having a high degree of steadyness at a certain level or higher, but in a good listening environment where there is almost no noise. If the sound gain of the hearing aid changes in a short time following the power fluctuation of the voice, the hearing of the words will be degraded, or if the noise is low as a whole when the voice is lost, this There is a problem that noise is emphasized. here,
The degree of steadiness is a measure of short-term fluctuations in power. Power fluctuations are small, as in an air conditioner. In addition, steadyness is low in the noise whose power fluctuates drastically as in a sheet metal factory.

[0005] A nonlinear amplification hearing aid is better under low stationary noise, but a linear amplification hearing aid is considered better under high stationary noise. An analyzer that measures such a measure of "steadiness" is considered. There were no hearing aids, nor did any hearing aids adaptively change characteristics according to such a measure.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the prior art, and an object of the present invention is to clarify a sound to be heard regardless of the quality of environmental noise. It is an object of the present invention to provide a hearing aid that can be listened to.

[0007]

According to a first aspect of the present invention, there is provided an external sound detecting means for outputting an electric signal corresponding to an external sound, and an output signal of the external sound detecting means is processed. In a hearing aid device including a signal processing unit and an audio output unit that converts an output signal of the signal processing unit into an audio signal, the signal processing unit is configured to detect a steady state of noise included in an external sound from a signal output by the external sound detection unit. A degree-of-stationaryness detecting means for detecting the degree at predetermined time intervals, an input / output characteristic processing means for compressing and amplifying an output signal of the external sound detecting means, and a compression amplification characteristic storing a compression amplification characteristic parameter suitable for a hearing aid wearer A parameter table, wherein the input / output characteristic of the input / output characteristic processing means is determined based on the compression / amplification characteristic parameter corresponding to the degree of steadiness output from the steadiness degree detection means. A.

According to a second aspect of the present invention, in the hearing aid according to the first aspect, the compression amplification characteristic parameter is a compression ratio.

According to a third aspect of the present invention, in the hearing aid according to the first aspect, the compression amplification characteristic parameter is a knee point.

According to a fourth aspect of the present invention, in the hearing aid of the first aspect, the compression amplification characteristic parameter is a gain in a linear region.

[0011]

Embodiments of the present invention will be described below with reference to the accompanying drawings. Here, FIG. 1 is a block diagram of a hearing aid according to the present invention, FIG. 2 is an explanatory diagram relating to frame division processing, FIG. 3 is an input / output characteristic diagram showing a relationship between an input sound pressure level and an output sound pressure level, and FIG. Is an input / output characteristic table, FIG. 5 is a flowchart showing the operation of the hearing aid according to the present invention, FIGS. 6 to 8 are frequency distribution diagrams for explaining the operation of the hearing aid, and FIG. 9 is another embodiment of the hearing aid. It is a frequency distribution figure used for description of.

The hearing aid according to the present invention comprises an external sound detector 1, a signal processor 2, and an output unit 3, as shown in FIG. The external sound detection unit 1 includes a microphone 1a that detects an external sound and converts it into an electric signal, and an A / D converter 1b that converts an output voltage of the microphone 1a into a digital signal. And output digital data based on the data.

The signal processing unit 2 includes an input / output characteristic processing unit 2a,
A frame memory 2b, an average effective level calculator 2c, an average effective level memory 2d, an appearance frequency distribution map calculator 2e,
A peak position detecting section 2f, a steady-state calculating section 2g, an input / output characteristic reference section 2h, an input / output characteristic table 2i, an input / output characteristic function generating section 2j, and an input / output characteristic function memory 2k.
The digital data output from the / D converter 1b is signal-processed and output to the output unit 3.

The frame memory 2b sequentially stores digital data output from the external sound detector 1. That is, as shown in FIG. 2, when the storage of the digital data sequence belonging to the first time frame F0 is completed, the digital data sequence of the first time frame F0 is sequentially converted to the digital data sequence belonging to the next time frame F1. Rewriting, and thereafter, are repeated.

When the storage of the digital data sequence belonging to the unit time frame output from the frame memory 2b is completed, the average effective level calculation unit 2c calculates the average effective level of the unit time frame from the digital data sequence belonging to the unit time frame. Is calculated.

The average effective level memory 2d stores the average effective level p of the unit time frame output by the average effective level calculation unit 2c each time.

The frequency-of-appearance-frequency-distribution-calculation section 2e creates a frequency distribution chart representing the number (frequency) of time frames corresponding to the average effective level p of the time frames sequentially calculated by the average effective level calculation section 2c.

The peak position detecting section 2f detects the mode value of the distribution appearing in the region having a small average effective level in the frequency distribution chart created by the appearance frequency distribution chart calculating section 2e.

The stationary degree calculating section 2g includes a peak position detecting section 2
The kurtosis Kw (stationarity) of the frequency distribution within a predetermined range is calculated around the mode detected at f. Here, the kurtosis K
w represents the sharpness of the distribution.

The input / output characteristic reference section 2h includes a stationary degree calculation section 2
The compression amplification characteristic parameter is obtained from the stationary degree output by g, that is, the kurtosis Kw, and the input / output characteristic table 2i storing hearing aid fitting data obtained in advance.

In the case of this embodiment, the compression amplification characteristic parameters include the gain g (i) of the linear part and the input level k1.
(I) and the input level k2 (i). Here, a case where the calculated kurtosis Kw is divided into five levels of steadyness will be described. The input / output characteristic table 2i also stores parameters corresponding to five levels of steadiness.

As shown in FIG. 3, the input / output characteristics indicating the relationship between the input sound pressure level and the output sound pressure level are such that the gain is increased for a low input level and the gain is constant up to the input level k1. The linear amplification is performed so that The gain is gradually reduced for the input of the input level k1 or higher, and the gain is changed so as not to amplify the input of the input level k2 or higher. Here, the input level k1 is called a knee point, and the slope of the characteristic from the input level k1 to the input level k2 is called a compression ratio.

The hearing aid fitting data stored in the input / output characteristic table 2i can be obtained, for example, as described below. However, in this embodiment, a case is considered in which the degree of steadiness of environmental noise is divided into five levels. Here, the stationarity (1) has a low stationarity (noise in which power fluctuates in a short time, a noise in a sheet metal factory, etc.), and the stationarity (5) has a high stationarity (noise with small power fluctuations, Noise, etc.).

Each of these steady-state noises has a speech (for example,
Test sound superimposed with speech sound intelligibility test term sound or any sentence that is read aloud) is presented to the hearing aid wearer, so that the hearing aid wearer can hear it most clearly. Feel) the gain g (i) of the linear part,
The input level k1 (i) corresponding to the knee point and the input level k2 (i) are obtained and stored in the input / output characteristic table 2i.

The inspection sounds are of the following five types [1] to [5]. [1] Noise of steady state (1) + speech sound [2] Noise of speech level (2) + speech sound [3] Noise of speech level (3) + speech sound [4] Noise + speech sound of steadyness (4) [5] ] Noise of level (5) + speech

Accordingly, the compression amplification characteristic parameters stored in the input / output characteristic table 2i are as shown in FIG. In FIG. 4, the leftmost column is arranged in ascending order of noise steadyness from the top. The second column is the linear gain g
(I) The third column stores the input level k1 (i) corresponding to the knee point, and the fourth column stores the input level k2 (i).

The input / output characteristic function generator 2j outputs the input level k1 (i) corresponding to the gain g (i) of the linear section and the knee point at a certain degree of steadyness (i) output from the input / output characteristic reference section 2h. And an input / output characteristic function f (k) is generated based on the input / output level k2 (i).
(K) is stored in the input / output characteristic function memory 2k. The input / output characteristic function f (k) generated here is the input level k
There are three shapes depending on

That is, when the input level k is lower than the input level k1 (i) corresponding to the knee point, f (k) = g
(I) It becomes [dB].

When the input level k is higher than the input level k1 (i) corresponding to the knee point and lower than the input level k2 (i), the following equation is obtained.

F (k) = g (i) × (k2 (i) -k) /
(K2 (i) -k1 (i)) [dB]

When the input level k is higher than the input level k2 (i), f (k) = 0 [dB].

The input / output characteristic processing unit 2a substitutes the average effective level p of the unit time frame obtained from the average effective level memory 2d into the input / output characteristic function f (k) stored in the input / output characteristic function memory 2k. By doing so, the gain f (p) of the unit time frame is calculated, and the frame memory 2
multiplied by the digital data Xb in b and output to the output unit 3.

The output unit 3 includes a D / A converter 3a and a D / A
Amplifier 3 for amplifying output signal of converter 3a with a predetermined gain
b, and an earphone 3c for converting the output signal of the amplifier 3b into an electroacoustic signal. The digital data signal output from the input / output characteristic processing unit 2a is converted into an analog sound signal having a desired level.

The operation of the hearing aid according to the present invention configured as described above will be described with reference to the flowchart shown in FIG.

First, in step SP1, the hearing aid performs initialization of the average effective level memory 2d, the input / output characteristic function memory 2k, and the frame memory 2b. For example, the input / output characteristic function f (k) at the time of steadyness (3) is set in the input / output characteristic function memory 2k, and the gain becomes 0 dB at the time of steadyness (3) in the average effective level memory 2d. Such k2 (3) is set. All 0s are set in the frame memory 2b.

Next, in step SP2, a predetermined total number of time frames (total frequency = Z) required to create an appearance frequency distribution map is set in a counter (Z), and an average effective level p is calculated. The number of samples (total number Zf) in a predetermined unit time frame required for
Set to f).

Next, in step SP3, the counter (a) for counting the number of time frames for creating the appearance frequency distribution chart is reset (a = 0).

Next, in step SP4, the counter (b) for counting the number of samples in the unit time frame is reset (b = 0).

Next, in step SP5, the microphone signal from the microphone 1a is converted into an A / D converter 1b.
A / D conversion is performed at a predetermined sampling frequency to obtain digital data X'b.

At step SP6, the average effective level p stored in the average effective level memory 2d is read.

Next, in step SP7, the input / output characteristic function f (k) is read from the input / output characteristic function memory 2k.

Next, in step SP8, the gain f (p) is substituted by substituting the average effective level p read from the effective level memory 2d into the input / output characteristic function f (k) read from the input / output characteristic function memory 2k. obtain.

That is, when the average effective level p is smaller than the input level k1 (3) corresponding to the knee point, f
(P) = g (3) [dB].

The average effective level p is larger than the input level k1 (3) corresponding to the knee point and the input level k2
(3) If smaller, the following equation is obtained.

F (p) = g (3) × (k2 (3) -p) /
(K2 (3) -k1 (3)) [dB]

The average effective level p is equal to the input level k2
If it is larger than (3), f (p) = 0 [dB].

Next, in step SP9, the input / output characteristic processing section 2a multiplies the digital data Xb in the frame memory 2b by the gain f (p) to obtain an output value R = Xb × f
(P) is generated.

Next, in step SP10, the input / output characteristic processing 2a outputs the generated output value R to the D / A
Output to the converter 3a. The output of the D / A converter 3a is amplified by the amplifier 3b, and the output of the amplifier 3b is subjected to electroacoustic conversion by the earphone 3c and output as an acoustic signal into the ear canal of the hearing aid wearer.

In step SP11, step S
The digital data X′b obtained in P5 is represented by Xb = X ′
b is stored in the frame memory 2b.

Next, at step SP12, a counter (b) for counting the number of samples within the frame length.
Is incremented (+1).

Next, in step SP13, the value of the counter (b) is compared with the value of the counter (Zf), and b ≧ Z
It is determined whether or not f has been reached. When b <Zf,
Since the digital data sequence per unit time frame has not yet been reached, the flow returns to step SP5 to accumulate digital data. On the other hand, when b ≧ Zf, all digital data strings per unit frame are (x ₀ , x ₁ ,
x ₂ ,..., x _Zf-1 ).
Proceed to.

In step SP14, the average effective level calculation unit 2c calculates the average effective level of the unit time frame from the digital data sequence (x ₀ , x ₁ , x ₂ ,..., X _Zf-1 ) belonging to the unit time frame. Calculate p. The calculation of the average effective level p is performed by the following equation. In the following equation, n
Is the number of digital data strings belonging to the unit frame.

P = {(x ₀ ² + x ₁ ² + x ₂ ² ... X _Zf-1 ² ) / n} ^1/2

Next, the process proceeds to step SP15, where the average effective level p of the unit time frame obtained in step SP14 is stored in the average effective level memory 2d.

In step SP16, the average effective level calculated in step SP14 is created in order to create a frequency distribution chart with the frequency of a time frame having the average effective level p on the horizontal axis and the average effective level p on the vertical axis. A counter corresponding to p is provided, and each time the average effective level p is calculated, the corresponding counter is incremented to count the number of time frames having the average effective level p.

Next, in step SP17, the counter (a) is incremented (+1).

Next, in step SP18, the value of the counter (a) is compared with the value of the predetermined total frequency Z, and a ≧ a
It is determined whether or not Z has been reached. In the case of a ≧ Z, a frequency distribution chart of a predetermined total frequency (Z pieces) was obtained.
Proceed to step SP19. On the other hand, if a <Z, since a complete frequency distribution diagram has not been obtained, step SP4
After setting the counter (b) to b = 0, a predetermined loop is repeated.

The frequency distribution diagram obtained under the condition of a ≧ Z is as follows.
For example, those shown in FIG. 6 and FIG. Here, FIG. 6 is an example of a frequency distribution diagram in a case where conversational speech is included in noise having a low degree of steadyness, and FIG. 7 is a diagram of a frequency distribution in a case where conversational voice is included in noise having a high degree of steadyness. It is an example. 6 and 7
In each case, a distribution A due to environmental noise appears in a region with a low average effective level, and a distribution B due to voice appears in a region with a high average effective level.

In general, during noise, the talker talks at a higher level than the noise level in order to prevent the conversation sound from being masked by the noise. A distribution A appears, and a distribution B due to voice appears in a region where the average effective level is high.

FIG. 8 shows a frequency distribution when the noise distribution and the conversation voice distribution are close to each other and the boundary between the noise distribution and the conversation voice distribution is unclear. Since a part of the conversation voice distribution is included on the right side of the noise distribution (the part of the region where the average effective level is high), it is difficult to calculate the correct steady-state. In order to avoid this, in the frequency distribution diagram shown in FIG. 8, the distribution map shape in the region of the low average effective level is folded back toward the high region when viewed from the mode of the noise distribution.
The step SP19 may be executed after re-creating the symmetric distribution pattern shown in FIG.

The distribution A due to noise in FIG.
Is higher than the distribution A due to noise in FIG.
The distribution is sharp. This is because, as described in the section of the problem to be solved by the invention, the noise in FIG. 6 includes various kinds of acoustic components as compared with the noise in FIG.

Next, in step SP19, the peak position detection unit 2f determines the frequency distribution appearing in the region with the lowest average effective level from the frequency distribution obtained by the frequency-of-appearance distribution map calculation unit 2e, ie, the frequency attributed to noise. The mode (LL in FIGS. 6 and 7) of the distribution is detected.

In step SP20, the continuity calculator 2g determines the kurtosis Kw based on the data within a predetermined range centered on the mode of the distribution in the range A of FIG. 6 detected by the peak position detector 2f. Is calculated. The kurtosis Kw is defined by the following equation. Where n is the total frequency within a predetermined range,
Zi (i = 1,2, ......, n) is the average effective level of each time frame, Zave the average value of zi, V ² is the variance.

Kw = Σ (Zi−Zave) ⁴ / (nV ² ) −3

The value of the kurtosis Kw obtained from this equation is used as a scale representing the degree of steadiness. Also, the larger the value of the kurtosis Kw, the lower the degree of steadiness, and the smaller the value of the kurtosis Kw, the higher the degree of steadiness. The value of the kurtosis Kw is divided into five levels, and the five levels of steadiness from 1 to 5 are set.

Next, in step SP21, the input / output characteristic reference section 2h obtains the stationary level of the environmental noise output from the stationary level calculating section 2g, refers to the input / output characteristic table 2i, and obtains the obtained environment. The compression amplification characteristic parameters (gain g (i) of the linear portion, input level k1 (i) corresponding to the knee point, input level k2)
(I) is obtained from the input / output characteristic table 2i. For example,
When the stationary level of the environmental noise is (5), the gain g (5)
And an input level k1 (5) corresponding to a knee point;
An input level k2 (5) is obtained.

Next, in step SP22, the input / output characteristic function generating section 2j outputs the compression amplification characteristic parameter (gain g of the linear section) currently held by the input / output characteristic reference section 2h.
(5), input level k1 corresponding to the knee point
(5) An input / output characteristic function f (k) is generated based on the input level k2 (5)).

Next, in step SP23, the input / output characteristic function f generated by the input / output characteristic function generator 2j is generated.
(K) is written in the input / output characteristic function memory 2k.

Next, in step SP24, the frequency distribution diagram is cleared, and the flow returns to step SP3. Step S
At P3, the counter (a) is set to a = 0 as described above.

Further, the same processing as described above is executed from step SP4.

As described above, in the hearing aid according to the present invention, the test sounds in which speech sounds are mixed with various noises having different degrees of steadyness are heard, and the speech sounds contained in these test sounds are most easily heard. An input / output characteristic table 2i in which the amplification characteristic parameter is written is prepared, and a compression amplification characteristic parameter is called from the input / output characteristic table 2i in accordance with the steady state of the noise in the actual environment. Since the compression amplification characteristic of the input / output characteristic processing unit 2a is set based on the average effective level p at that time, it is possible to clearly hear speech in various noises having different degrees of steadyness.

The present invention is not limited to the above-described embodiment, and various modified embodiments can be considered. In the above embodiment, the gain g (i) of the linear part, the input level k1 (i) corresponding to the knee point, and the input level k2 are used as the compression amplification characteristic parameters.
The case where the input / output characteristic function f (k) is configured to be generated based on (i) has been described.

However, since it is only necessary to be able to hear speech clearly in the noise, one of the above three parameters may be selected as a parameter, and other parameters such as an attack time, Release time may be selected. Of course, a combination of these parameters may be selected. In this case, the parameters stored in the input / output characteristic table 2i are the selected parameters.

In the embodiment of the present invention, the kurtosis Kw
Is used as a scale representing the stationarity. However, since it is sufficient that the stationarity can be calculated, the standard deviation of the distribution may be calculated, and this value may be used as the stationarity. Also in this case, the larger the value of the standard deviation, the lower the degree of steadiness, and the smaller the value of the standard deviation, the higher the degree of steadiness. Further, the most frequent number in the appearance frequency number distribution chart may be set as the stationary degree. In this case, the larger the most frequent number, the higher the steadiness, and the smaller the most frequent number, the lower the steadiness.

[0075]

As described above, according to the present invention, it is possible to clearly hear a sound to be listened regardless of the quality of environmental noise.

[Brief description of the drawings]

FIG. 1 is a block diagram of a hearing aid device according to the present invention.

FIG. 2 is an explanatory diagram related to a frame division process.

FIG. 3 is an input / output characteristic diagram showing a relationship between an input sound pressure level and an output sound pressure level.

FIG. 4 is an input / output characteristic table

FIG. 5 is a flowchart showing the operation of the hearing aid according to the present invention.

FIG. 6 is a frequency distribution chart for explaining the operation of the hearing aid device.

FIG. 7 is a frequency distribution chart for explaining the operation of the hearing aid device.

FIG. 8 is a frequency distribution diagram for explaining the operation of the hearing aid device.

FIG. 9 is a frequency distribution chart for explaining another embodiment.

[Explanation of symbols]

1: external sound detection unit, 1a: microphone, 1b: A / D
Converter, 2 ... Signal processing unit, 2a ... Input / output characteristic processing unit, 2
b: frame memory, 2c: average effective level calculation unit, 2
d: average effective level memory, 2e: appearance frequency distribution diagram calculation unit, 2f: peak position detection unit, 2g: stationary degree calculation unit,
2h: input / output characteristics reference section, 2i: input / output characteristics table,
2j: input / output characteristic function generator, 2k: input / output characteristic function memory, 3: output unit, 3a: D / A converter, 3b: amplifier, 3c: earphone.

Claims (4)

[Claims] 1. An external sound detecting means for outputting an electric signal corresponding to an external sound, a signal processing means for processing an output signal of the external sound detecting means, and converting an output signal of the signal processing means into an acoustic signal. In a hearing aid device comprising sound output means, the signal processing means comprises: a stationary level detecting means for detecting, at predetermined time intervals, a stationary degree of noise included in the external sound from a signal output by the external sound detecting means; Input / output characteristic processing means for compressing and amplifying the output signal of the detection means, and a compression amplification characteristic parameter table storing compression amplification characteristic parameters suitable for the hearing aid wearer, corresponding to the stationarity outputted by the stationarity detecting means A hearing aid device for determining input / output characteristics of the input / output characteristics processing means based on the compression / amplification characteristic parameters; 2. The hearing aid according to claim 1, wherein the compression amplification characteristic parameter is a compression ratio. 3. The hearing aid according to claim 1, wherein the compression amplification characteristic parameter is a knee point. 4. The hearing aid according to claim 1, wherein the compression amplification characteristic parameter is a gain in a linear region.