SU1765903A1 - Method of signal processing in hearing aid - Google Patents

Abstract

Use: medical equipment. SUMMARY OF THE INVENTION: The received acoustic signal is divided into frequency bands and separately amplified, followed by summation, the signal is divided into frequency bands with highlighting of the envelopes, which are amplified and summed with the original signal, and the resulting sum signal after adding with separately amplified components The signal in the frequency ranges is supplied in acoustic form to the human ear. 1 il.

Description

 The invention relates to medical technology and can be used to rehabilitate the hearing of people with hearing loss of various kinds.

The known method 1 for adjusting the hearing aid and its use, characterized in that the adjustable amplifier circuit contains at least one regulating element operated by the consumer, which can change the slope of the frequency response at frequencies higher than the cut-off frequency 500-2000 Hz. In this case, it is possible to change the treble boost.

The disadvantage of this method is that in people with a complete loss of sensitivity to sounds, whose frequencies are concentrated in a certain part of the spectrum, hearing aids, constructed according to the described method, do not ensure the recovery of the perception of any information about sounds, whose frequencies lie in the affected parts of the spectrum.

Of the known technical solutions, the closest to the proposed is

the method described in the patent Scheme and method of processing a sound signal in a hearing aid (US Patent No. 4680798, Cl, H 04 R 25/00, H 03 C 3/20. ICM, 1988, Issue 137, No. 4) 2.

In the aforementioned audio signal processing method, an electrical audio signal is generated, the signal is divided into several frequency bands, a separate signal amplification in each band, and the sum of the amplified signals of each band to obtain a converted signal. When controlling the separate amplification, the audio signal is gated in the ranges and, based on the results of the gating, the gain level for each range is determined. The analog and digital conversion and individual detection is performed during the acquisition to obtain an analog signal corresponding to the level of the audio signal in each band.

A common disadvantage of these methods of building hearing aids is ISO

Vj

Os

ate you oh

with

with a low ability to restore the perception of the human auditory system of such sound signals, the processing mechanism of which in the auditory system is violated,

Thus, if the Sukhov system loses the ability to perceive residual sounds in one or another part of the speech frequency spectrum, the existing hearing aids are not able to correct this defect, which leads to the loss of a significant part of information perceived by a person through sound signals in the normal state of his auditory system.

As an example, we can cite the perception of an amplitude-modulated sound signal, the carrier frequency of which falls into the spectral region, the sensitivity of hearing to which in a given person is impaired.

Such a person hears neither the carrier of such a signal nor the modulating tone, although the frequency of the latter is in the range of frequencies whose perception in this patient is not disturbed. Such a defect has a particularly large effect on the perception of complex sounds (for example, speech).

One of the reasons for this drawback is that the existing methods of building hearing aids are based on the selective amplification of those components of the sound signal, the perception of which in a given person is disturbed,

Obviously, in the event of a disturbance in the auditory system of the functioning of any mechanism for processing the sound signal, the amplification of the latter itself cannot compensate for the perceived defect.

Consider some analogues of the proposed method of signal processing in hearing aids.

In the most common hearing aids, the frequency correction gain is currently performed for the entire channel with a relative gain boost in the high frequency range, the degree of which is adjusted both smoothly and in steps. Such hearing aids are domestic devices of types K 10, K 13, etc., as well as foreign ones, for example, type NA-41, HA27D from Rion Co LTD (Japan), 37ILDC from Oticon (Dani), etc.

When the hearing sensitivity deteriorates in the local frequency domain, for example, around 2 kHz, increasing the gain towards higher frequencies by known devices is not effective. At the same time, local loss of sensitivity at a frequency in the region of 2 kHz leads to a deterioration in the intelligibility of human speech perception by 30–40%.

With loss of hearing sensitivity on

any part of the spectrum occupied by speech signals is simultaneously disturbed by the mechanism of the formation of residual tones, which significantly affect the

0 perception of human sound signals (in particular, speech signals),

Since hearing aids of this type do not contain in their composition the formation of residual tones, they are

5 use in many cases does not provide impaired speech intelligibility.

As an analogue of the proposed method of building a hearing aid, you can

0 consider a hearing aid with a transposition of the frequency of sounds, issued by Oticon (model TP72).

In this type of hearing aid, the sounds of higher

5 frequencies (in the range of 4000-8000 Hz) in the low-frequency sounds.

For this, a special high-pass filter with a cut-off frequency of 4000 Hz and a large slope at this frequency (36 dB / octave) is used. It serves to prevent the passage and subsequent transformation of higher formant vowel sounds. The filtered voltage above 4000 Hz is applied to the nonlinear gain stage, after which the low frequencies are further amplified and the voltage of all difference frequencies up to 1500 Hz is extracted. The combination tones thus allocated from the frequency range of 4000-8000 Hz are fed to the common gain channel and to the telephone.

The main disadvantages of the TP72 model are: 1. Transposition into low-frequency part of the spectrum is carried out only for sounds in the frequency range from 4000 Hz and above (up to 8000 Hz). All of these sounds lie outside the frequency band of a standard telephone channel, within which the main part of the voice signals are located. Transformation into the low-frequency region of sounds from a non-speech portion of the frequency spectrum leads to the penetration into the speech segments of the spectrum by additional

5 interference, which significantly impairs the noise immunity of hearing aids of this type.

2. Since, in TP72-type hearing aids, a portion of the spectrum from which signals are transposed is separated from the rest.

a section of the spectrum with a bandpass filter (4000-8000 Hz) with sufficiently steep slopes of the amplitude-frequency characteristic does not underline or emphasize the main residual tones, which largely determine the perception of the auditory system of images of complex acoustic signals (speech). As is known (4), residual tones occur under the condition that the frequencies of the spectral components generating these tones do not exceed 4–5 kHz.

The purpose of the invention is to improve the quality of the output signals of hearing aids by obtaining residual tones in the output signal.

The goal is achieved by processing a signal in a manner that includes receiving an acoustic signal, converting it into an original electrical signal, dividing the signal into frequency bands, separately adjustable amplification of signals in each frequency band, summing the amplified signals of all frequency bands, and converting the total electrical signal, after dividing it into frequency ranges, the envelopes are separated, they are controlled separately and added to the original signal. ohm, and the resulting sum signal is added to the total amplified signal of all frequency ranges. The novelty of the proposed signal processing method in hearing aids is that the envelopes of the amplified frequency components of the electrical equivalent of the acoustic signal are added to the signal and the resulting signal in the non-linear feedback ring is fed to the human ear after electro-acoustic conversion.

A new feature of the proposed method of signal processing in hearing aids is the selection of the total signal envelope in the form of an independent signal fed to a person’s ear, and as a result, the expansion of the number of types of signals perceived by a person suffering from a defect in the auditory system, in particular ensuring the perception of residual tones of complex sounds.

The proposed method makes it possible for humans to perceive a modulating signal of a modulated sound located in the frequency range that is not perceived by a person due to impaired auditory function.

The essential difference of the proposed method of signal processing in auditory

Aids from existing methods of compensating for human hearing impairment are the introduction into the structure of hearing aids of blocks that carry out functional transformations of signals close to it, which are carried out in a normally functioning human hearing system, but which have been for some reason impaired. The converted signal, when applied to the input of the human auditory system, is perceived by the latter as more complete and natural, since it contains components that are not formed by existing hearing aids.

The drawing shows a block diagram of a device that implements the proposed method. In the proposed method, the audio signal is converted into its electric

0 equivalent, is divided into several frequency ranges, each of which amplifies the signal components, extracts their envelopes, and then adds both the amplified signal components and their envelopes with the electric equivalent of the original acoustic signal. In this case, the gain of the signal components in different ranges is different and is determined by

0 the nature of the hearing defect in humans.

The device implementing the proposed method contains the source 1 of the original signal, the adder 2, the filter set 3, the selector envelopes 4, the amplifier converters 5 for the envelopes, the amplifier converters 6 for the signal components, the second adder 7, the electroacoustic converter 8, the control unit 9 converters 5. At the same time

0 source signal source 1 is connected to the first adder 2, the output of which is connected to the input of the second adder 7 and to the inputs of filters 3, the outputs of which are connected to the amplifier converters 6 and

5 with envelope selectors 4, the outputs of which are connected to the inputs of the converters 5, the outputs of which are connected to the inputs of the first adder 2, and the outputs of the amplifiers-converters 6 are connected to

0 inputs of the second adder 7, the output of which is connected to the electro-acoustic transducer 8, and the control inputs of the transducers 5 and transducer amplifiers 6 are connected to the outputs

5 envelope highlighters,

The practical implementation of the proposed method of signal processing in hearing aids is carried out using commercially available circuits. Thus, amplifiers and adders can be made

on KR140UD20A type microcircuits, envelope selectors - on KR544UDD2V, K140UD6, and KD513A microcircuits. The filters included in blocks 3 are made according to the scheme of active RC filters built with KR140UD20A microcircuits according to standard schemes for 4th order filters. The implementation of the control unit is carried out taking into account the performance of the main functions performed by it: estimating the levels of the frequency components of the signal and their envelopes and developing control signals for setting the gain factors (attenuation in blocks 5 and 6) of both the frequency components and their envelopes . In accordance with what has been said, control unit 9 consists of integrator nodes (averagers), peak detectors, control voltage generation circuits. Frequency component levels are estimated at the average voltage at the outputs of the envelope highlighters 4 using averagers (integrators) performed on any type of operational amplifiers in accordance with typical schemes (for example, 140UD6, KR140UD20A operational amplifiers, etc. ). The envelope levels of the components of the signal in the selected bands are estimated using voltages at the outputs of the peak detectors, which determine the amplitudes of the variable components of the envelopes in the channels. Peak detectors are made according to standard schemes for operational amplifiers on IC types KP140UD20A, 140UD6 (or others of the same class). The control voltages from the outputs of the control unit are fed to the control inputs of the amplifiers-converters 5 and 6. The latter are performed according to the typical schemes of operational amplifiers on ICC type IC140UD20A, 140UD6, etc., and as an element regulating the gain factor, an optocoupler (e.g., type ZOTU2E) included in the feedback circuit of an operational amplifier or in an engine arm. In this case, the control voltages from the outputs of the control unit are supplied to the control inputs of the optocouplers in blocks 5 and 6, and thereby the gain (attenuation) of the signal components (in the amplifiers 5) and the envelopes (in the amplifiers) is adjusted. -converters 6) before they arrive at adders 2 and 7. In the simplest case, the functional relationship between the magnitude of the control voltage and the level of the corresponding component of the signal is set in the process of adjusting the output control circuits

stress in the clinic under the characteristics of the auditory system of the patient.

Typically, such a connection includes non-linear constraint-type operations that are performed using diodes (for example, type KD513A) or logarithmic amplifiers (for example, type 1112P1). Control voltage generation schemes

0 can also be performed according to standard circuits for operational amplifiers operating in amplified mode with limiting or in logarithmic mode.

Amplifier converters are

5 substantially scale converters with controllable transform (gain) coefficients. Their implementation is also widely known from the literature (3). All the indicated circuit solutions were made in the form of a valid layout.

The implementation of the hearing aid in accordance with the proposed method using microelectronics allows it to be performed in a volume of about 125

5 cu. cm with a weight not exceeding 100 g, which makes it possible to use it in the form of a pocket-type hearing aid, whose ability to restore the human auditory function is significantly higher than that of

0 existing devices of the same class. The proposed method of signal processing in hearing aids allows for more complex control functions (e.g.

5 normalization of signal component levels with respect to the maximum, underlining individual components according to some criterion, etc.). Such operations generally require use as a microprocessor control unit, as was done in the prototype of the proposed method (i.e., in US patent No. 4,680,798, Cl, H 04 R 25/00).

Hearing Aid with Functional

5 by the scheme shown in figure 1, implemented in accordance with the proposed method, works as follows. The acoustic signal is converted into an electrical equivalent using an acoustoelectric transducer 1. The output signal of an acoustoelectric transducer 1 is fed to one of the inputs of the first multi-channel adder 2, the other inputs of which receive signals from the transducer amplifiers 5. The output of the first adder is total. the signal simultaneously arrives at the second multi-channel adder and at the inputs of the band-pass filters 3. Each of the output signals of the band-pass filters 3 is fed simultaneously to the input of the envelope selector 4 and to the input of the corresponding amplifier - converter "1.0: 2" p 6, outputs

which are connected to the inputs of the second adder 7. The output voltages of the envelope highlighters, after being converted into amplifiers 5, are fed to the inputs of the first multi-channel adder 2, they add up to the output voltage of the acoustoelectric converter 1. Signals from the outputs of the envelope selection blocks 4 are received control unit 9, which estimates the energy levels in the respective frequency channels and the energy levels of the variable components of the envelopes in the same frequency channels. Based on the estimates obtained and the data entered from a person using the control unit, the gain factors of the transducer amplifiers 5 and 6 are set. Thus, as in the prototype, it is possible to separately amplify the frequency components of the signal in accordance with the nature of the defect in the auditory system. this person. Ensuring that the functions of the control unit 9 are performed in the most complete form can be realized with the help of a microprocessor operating in the same mode as described in US patent specification No. 4,680,798, cl. H 04 R 25/00.

A feature of the proposed method of signal processing in hearing aids is the introduction of so-called residual (additional) tones into the sound-shaping scheme in their structure. As shown in general, a number of physiological and psychoacoustic studies, when a person perceives complex sounds (speech, music, tone modulation, etc.), residual tones that are not contained in the original signal itself, but are generated by nonlinear signals, play a significant role. signal conversion at different levels of the auditory system (4, 5).

When the mechanism of formation of residual tones is disturbed, the nature of the perception of complex sounds changes dramatically.

Since by themselves the residual sounds in the original acoustic signal are not contained, their perception at the aforementioned hearing defect disappears. To the greatest extent, this defect manifests itself in impaired speech intelligibility, music perception, etc.

To eliminate the consequences of disturbing the formation of residual signals in the human auditory system

It is proposed to introduce into the hearing instruments a circuit that performs similar functions, and add residual signals thus generated to the original

sound. After the residual signals appear to be inputted into a sound signal arriving at the ear of the person, the latter will perceive them, even if the mechanism of their formation in his auditory system does not work.

Schematic implementation of the mechanism for the formation of residual signals, in particular, should ensure the formation of intense combination tones of the form

f 2fi-f2,

the relative level of the combination frequencies of the type should not depend on the level of the signal.

The fulfillment of such requirements is achieved by converting the signal according to the feedback scheme along an envelope.

Consider the processes in this scheme in more detail. We introduce the following notation in the diagram shown in FIG. 1: Uc is the electrical equivalent of the sound signal at the output of the block (1); Y (t) is the signal at the output of block 2; Vi - signals at the outputs of filters 3; Wi (t) is the voltage at the outputs of the envelope highlighters; Ф (т.) - voltage at the outputs of the converters.

When analyzing the processes in the scheme shown in Fig. 1, we first make the following simplifications:

1. Filters 3 have rectangular frequency and linear phase characteristics.

The bandwidth of the filters is greater than the frequency spacing of the adjacent harmonic components included in the amplitude-modulated signal, or in the two-tone signal,

2. The passbands of the filters separating the envelopes in blocks 4 are much smaller than the values of the resonant frequencies of the respective channels,

3. The transfer functions of the envelope selection blocks 4 are 0 for high-frequency filling of the signal and 1 for the modulating signal. 4, Transducers a have transfer function

"I (t) + 1,

where A (t) is the modulating signal (envelope signal).

In view of the simplifications made, the following expression can be derived for the voltage at the output of the 1st adder.

Let be

Uc (t) A (t) f (t),

where f (t) is the high-frequency filling of the signal. Then

y (t) A A (t) f (t) + (o (t) + 1. Consider special cases: 1. Let

A (t) Umo (1 + mcos Q t) “1;

f (t) sin (wt + p).

Then

y (t) Umo (1 + mcos Q t) sin (ut) +

+ Umo (1 + mcos Q t).

Thus, when an amplitude-modulated oscillation is presented to the input of the proposed hearing aid, the modulating voltage is released as a separate low-frequency signal. The latter, after electro-acoustic conversion, enters the human ear and is perceived by the auditory system as a self-contained signal, even if in the high-frequency filling area (i.e., at the frequency (o), the sensitivity of the auditory system is completely impaired.

2. To describe the process of forming the so-called residual tones, i.e. combination tones with frequencies of 2fi-T2, we assume that the transfer function of the selection of the envelope provides at its output a voltage of the form

Wi (t) Wo

Y1 Vi (t) + ya Vi2 (t) + uz | 3 (x).

Then the envelope AVyh voltage Uc for the case when it consists of two tones, i.e.

Uc U-jCOS ІL t + UL COS UJ2 t

1M2

will have the form (without components, having frequencies sludge and higher)

AvichM + + Y2 U "CiU% COS (Ul - Wfl} 4+ 3 nodes and Ј" (2 W1 в) gProved corresponding calculations, you can get the following approximate expression for the signal formed at the output of the first adder 2

y (t) 1 Cs al t + 1 Cs 0) 21 + + (g) + 1,

Thus, in the considered case, the input of the auditory system is oscillated at a frequency of 2 villages per second, i.e. the so-called cubic tone, clearly perceived by a person when listening to complex sounds. Amplitude dependence

the generated combination tone with a frequency of 2 yL - y from the amplitudes of the components of the original stimulus U o and U yr turns out to be close to that observed in psychoacoustic experiments on healthy subjects (6).

In a more general case, when the above simplifications are not performed in the flowchart shown in Fig. 1, the appearance of the input signal causes a more complex response generated by the interaction of the signal with the sum of its envelopes in different frequency channels occurring in the circuit. non-linear feedback. A more detailed theoretical and experimental analysis of the signals generated at the same time showed that in this case independent low-frequency tones appear, perceived by a person as residual sounds,

In the process of implementing the proposed

method of building hearing aids

a laboratory model of the apparatus was made,

included all items included

in the block diagram shown in figure 1.

Testing the effectiveness of a hearing aid built using the proposed method was performed on people with significant hearing loss in one or another area of sound frequencies.

Initially, an amplitude-modulated sound signal was applied to a person's ear, the whole spectrum of which lies in the frequency range, to which this person is not sensitive.

The lack of human perception of such a signal was recorded by its inability to detect a modulating signal or correctly estimate its frequency. The signal amplitude was set at 80 dB ultrasound,

After that, the same acoustic signal of the same level was fed to the input of the layout of the hearing aid, from the output of which the acoustic signal came to the human ear.

It is shown that for a normally hearing person, the signal from the output of the proposed hearing aid differs little from the original audio signal.

At the same time, people with a pronounced hearing loss in the carrier frequency of the original sound signal without a hearing aid did not hear or heard a weak illegible sound (depending on the degree of loss of sensitivity in the corresponding frequency range).

When such people fed the signal after it was converted in the proposed hearing aid, all the subjects clearly heard the modulating frequency signal.

The most important effect for improving the quality of the output signal of the layout due to the formation of residual tones in it is a significant increase in the amplitudes of the components of the spectrum of vowel sounds of speech signals.

Thus, in the spectrum of sound a, at the output of the layout, with the feedback circuit on, the spectral envelope components were 2–5 times higher in amplitude (depending on their frequency location) than the spectrum components of the same sound, while the feedback on envelope. The intensity of the sound was equalized in both cases. A similar effect was observed for all other vowel speech sounds. An increase in the amplitudes of the components of the vowel sound spectra is the result of the formation and addition of the residual tones of speech signals to the initial signal, which causes an improvement in the quality of the output speech signal at the layout output.

Thus, the proposed method for constructing a hearing aid provides a qualitatively new effect.

Claims (1)

A method for processing a signal in a hearing aid, including receiving an acoustic signal, converting it into an original electrical signal, dividing the signal into frequency bands, separately adjustable amplification of the signal in each frequency band, summing the amplified signals of all frequency bands, and converting the total electrical signal into an acoustic signal , characterized in that, in order to improve the quality of the output signal by obtaining residual tones in the output signal, from After dividing the signal into frequency ranges, the envelopes are separated, their separate adjustable gain is made and summed with the original signal, and the resulting sum signal is added to the total amplified signal of all frequency ranges. ABOUT