EP1101390B1 - Hearing aid having an improved speech intelligibility by means of frequency selective signal processing, and a method for operating such a hearing aid - Google Patents

Abstract

The invention relates to a hearing aid comprising a microphone (1), a signal processing unit (2) and a listening device (3), whereby the signal processing unit (2) has a filter element (6) for splitting the signal into a number of partial signals (12). In addition, the hearing aid comprises an analyzing element (7) for detecting speech information found in the partial signals (12), and has a conditioning element (8) for boosting the partial signals (12) during the availability of speech. The invention also relates to a method for operating such a hearing aid.

Description

The invention relates to a method for operating a hearing aid with the features of the preamble of the claim 1 and a hearing aid with the features of the preamble of Claim 5.

Hearing aids which are wearable on the head are known as prior art, what special hearing programs to suppress Have noise, certain by activating a filter Frequency components are damped. As a result, a Improvement in general noise levels achieved which, however, is not sufficient for many applications.

A hearing aid with at least one is known from EP 577 441 A2 Input converter of a signal processing unit and one Output converter. The signal processing unit comprises an A / D converter at the input, several parallel signal channels with one filter, one analysis element and one Adaptation element and at the output a D / A converter. The signal processing unit enables recognition and lifting of speech signals. The signal analysis is based on measurements the amplitude, frequency, duration and frequency changes of a signal.

From US 5,046,102 a hearing aid emerges, in which between a microphone and a handset an arrangement of several Frequency selection channels is switched on. Farther this hearing aid includes a device for measuring the strength of a signal in each individual frequency selection channel and to influence the channels with each other in the sense of suppression from weak signal channels to the advantage of strong ones Signal channels.

WO 91/03042 describes a method and a device for distinguishing known from speech and interference signals. Doing so divides the input signal into several band-limited partial signals, each of which has at least two harmonic frequencies of the speech signal contains. It will be the envelopes of the partial signals as well as the synchronism between the envelopes determined. Depending on the value of the determined synchronism the partial signals are further processed. The disadvantage of this method is the high level associated with it Computational effort.

A multi-channel hearing aid emerges from US Pat. No. 4,852,175. in which the individual channels independently on the Existence of noise is investigated. Will in one Channel noise is detected, so there is a reduction in Reinforcement for this channel. The level of noise in one Channel is determined based on an amplitude histogram.

The invention has for its object a hearing aid, such as a hearing aid that can be worn on the head or an implantable hearing aid as well as a method for operating a hearing aid create an improvement in speech intelligibility in enable a variety of listening situations.

The task becomes for the procedure for operating a hearing aid by the features of the characterizing part of the claim 1 in conjunction with the features of the preamble solved. Advantageous embodiments of the method are realized by subclaims 2 -8. For the hearing aid the task is characterized by the characteristics of the characteristic part of claim 9 solved with the features of the preamble.

The hearing aid according to the invention has an analysis element for Detection of speech information in which the filter element of the Hearing aid split partial signals. Thereby in the Partial signals the mean and the course of the envelope determined. If the level peaks, the amount exceeds of the envelope the amount of the mean. Such one Excess can be used as a criterion for the existence of a Level peak are used, so that this in the analysis of the partial signal the presence of speech information becomes detectable. Thereupon the considered partial signal be raised to face the detected speech information to amplify low-level noise.

In contrast to the prior art, in the case of the invention Hearing aid no general increase in certain Frequency ranges in which there is a higher probability Voice information is located, but it will be the individual Partial signals are first analyzed and their presence checked by voice information. Only when actually Voice information is available, the respective Partial signals raised and thus the speech information in the present partial signal compared to the (not raised) Noise is amplified in other partial signals.

In an advantageous embodiment of the hearing aid the split partial signals have the lowest possible frequency range. The lower the frequency range of the partial signals: is, the more accurate speech information can be from noise can be separated and can be an isolated and targeted Detected speech information is raised, thereby speech intelligibility can be significantly improved.

When the total signal is split into narrow frequency bands can also be achieved within a short settling time an increase in the respective partial signal during detection of language take place.

The analysis element advantageously comprises several analysis programs for the detection of different speech information and speech patterns. For example, an analysis program for the detection of sudden level peaks ("bursts") serve. Such peaks are often speech information, the level of which is the constant and low level Noise exceeds.

Other analysis programs can detect certain ones Individual speech elements, e.g. Vowels, sibilants or consonants concern what exactly detects these speech elements and can be specifically raised to a clear To improve speech intelligibility. This can be of great importance especially for hearing impaired children because these are often due to overheard consonants and Sibilants have great difficulty in language learning.

In general, the hearing aid according to the invention enables any short-term and only occasionally occurring Detect and specifically raise voice information. At the same time, the control process is ended immediately when from Analysis element determines that no speech information there is more. This avoids that after the voice information re-emerging noise was also raised becomes.

Further advantages and designs of the hearing aid according to the invention can be found in the description of the invention Method and the embodiments of the invention.

In the method according to the invention, an analysis is carried out first for the detection of speech information in the partial signals, by splitting the total signal in a filter element be won. In the presence of voice information the respective partial signal is raised and thus the Speech intelligibility of the detected speech information improved. In an advantageous process variant Speech information in the partial signals by detection determined from level peaks ("bursts"). When such occurs Level peaks are highly likely to be speech information in front of the low noise level (e.g. background noise).

The respective partial signal can be used in further process variants in its signal course with mathematical methods and characteristic values be analyzed in the manner of a curve discussion.

In a particularly advantageous method variant, an increase in reliable detection of level peaks of the partial signal only when the amount of the envelope the amount of the mean of the partial signal by one adjustable minimum value to increase the Avoid partial signal with smaller level fluctuations.

In addition to the waveform analysis of the partial signal ("curve discussion" e.g. e.g. Determination of minima, maxima and pole positions), with mathematical methods and key figures to analyze the partial signal also auditory parameters (e.g. Depth of modulation, frequency of modulation etc.) are determined, to detect certain speech information and speech patterns and the partial signal in coordination with the determined Speech information type or speech pattern type by a corresponding one appropriate amount to raise.

The analysis of the partial signals according to the invention can be arbitrary short time intervals take place, even to individual only briefly occurring language elements (e.g. vowels, sibilants, consonants etc.) to be detected and practically with immediate effect and only short settling times to improve the Increase speech intelligibility. In an extreme case, a spoken word is just a single letter starting with matches a certain language pattern, especially highlighted while the other letters and syllables transmitted unreinforced or with less gain.

The method according to the invention thus enables selective Improve speech intelligibility. The invention Control and regulation jumps within a very short time Settling times (<than 2 ms) and switches within equally short decay times if there is no more language is detected.

FIG 1

ein Prinzipschaltbild einer erfindungsgemäßen Hörhilfe,

FIG 2

eine Analyse eines Teilsignals sowie

FIG 3

ein ergänzter Ausschnitt aus FIG 2.

The invention is explained in more detail using exemplary embodiments in the drawing figures. Show it:

FIG. 1

2 shows a basic circuit diagram of a hearing aid according to the invention,

FIG 2

an analysis of a partial signal and

FIG 3

an expanded section from FIG. 2.

The signal picked up by the microphone 1 passes through the hearing aid 1 shows the signal processing unit 2 to the listener 3. There is a preamplifier in the signal processing unit 2 4, an A / D converter 5 and a filter element 6 for Splitting the total signal into several partial signals 12, in present case in the three channels 11, 11 ', 11' '. In the analysis element 7, the partial signals of the individual channels 11, 11 ', 11 "for the presence of language. If A detection of speech is made over the connection between analysis element 7 and adaptation device 8 the respective Sub-signal 12 raised in the adaptation element 8 (increased), to improve speech intelligibility. In the signal processing element 9 are the partially raised partial signals processed (e.g. by AGC circuits). After a Merging of the partial signals 12 and a D / A conversion in D / A converter 10, the overall signal with improved speech intelligibility fed to the listener 3.

2 shows a partial signal 12 which has been checked for speech detection in the analysis element 7 and which has a sudden level increase ("burst") between t ₁ and t ₂ . It is assumed in analysis element 7 that speech information is present between times t ₁ and t ₂ and that low-level noise occurs before t ₁ and after t ₂ .

2 is a schematic illustration and shows for reasons clarity, a relatively long lasting level increase. The method according to the invention also enables Detection of short-term level peaks (not shown).

The course of the partial signal 12 can now be mathematical Methods and parameters e.g. with those from the curve discussion known methods can be assessed in the analysis element 7. there can have minima, maxima and pole positions as well as first or higher Derivatives are calculated and based on the results Conclusions about certain existing types of language information or speech patterns (e.g. vowels, sibilants, consonants etc.) are drawn.

The analysis of the partial signal 12 of the respective channel 11, 11 ', 11 '' serves for the detection of speech, since only when it is present of speech information, the raising of the invention respective partial signal 12 in the considered analysis period should take place in order to avoid that also a noise level is raised.

FIG. 2 shows the analysis of the profile of the partial signal 12 in the analysis element 7 the mean 13 and the (schematic shown) envelope 14 of the partial signal 12 is formed.

3 shows the course for reasons of clarity of the partial signal 12 removed, so that the amount course the envelope 14 and the mean value 13 are more evident.

Before t ₁ and after t ₂ , the envelope 14 reaches the value B, which only slightly exceeds the value C of the mean value 13.

Between t ₁ and t ₂ , the amount of the envelope 14 increases to the value A, as a result of which the coverage area 15 between the envelope 14 and the mean value 13 increases. It can be assumed that the value A is above a limit value X between the value B and the value A.

If the limit value X is exceeded by the value A of the envelope 14, speech is detected between times t ₁ and t ₂ and the partial signal 12 is raised in this period to improve speech intelligibility.

The increase occurs with the shortest possible settling time of less than 2 ms after t ₁ and is ended after t ₂ with the shortest possible settling time of likewise less than 2 ms.

The method according to the invention thus reacts briefly and spontaneously to the speech information present between the times t ₁ and t ₂ in the partial signal 12 and raises the partial signal 12 accordingly between the times t ₁ and t ₂ .

For reasons of clarity, a clear distance between the times t ₁ and t _{2 is} provided in FIGS. 2 and 3.

In the case of brief level peaks, the level increase (at t ₁₎ and level drop (at t ₂ ) can also follow one another briefly, as a result of which the partial signal 12 can be raised and lowered again practically without delay due to the particularly short settling / decay times (not shown). ,

Claims (9)

1. Method for operating a hearing aid with a microphone, a signal processing unit and an earpiece, the signal being split by a filter element of the signal processing unit into several partial signals and speech information being detected in the partial signals by at least one analysis element (7) and partial signals being boosted by at least one adapter element (8) given the existence of speech, characterised in that when the partial signals are analysed their mean values and their envelope are determined and the values compared with one another and the relevant partial signal is boosted if the value of the envelope exceeds the mean value.
2. Method according to Claim 2, characterised in that the relevant partial signal is boosted if the value of the envelope exceeds the mean value by an adjustable minimum value.
3. Method according to Claim 1 or 2, characterised in that when the partial signals are analysed a signal characteristic analysis of the partial signals is performed.
4. Method according to Claim 3, characterised in that when the partial signals are analysed the first and higher derivations are determined.
5. Method according to Claim 3 or 4, characterised in that when the partial signals are analysed characteristics of the partial signals such as minima, maxima and pole locations are determined.
6. Method according to one of Claims 1 to 5, characterised in that when the partial signals are analysed parameters such as modulation depth, modulation frequency and other auditory parameters are determined.
7. Method according to one of Claims 1 to 6, characterised in that the partial signals are analysed at adjustable time intervals.
8. Method according to one of Claims 1 to 7, characterised in that the relevant partial signals are boosted within maximum transient periods and decay times of 2 ms.
9. Hearing aid with means for carrying out the method steps according to one of Claims 1 to 8, with a microphone, a signal processing unit and an earpiece, the signal processing unit having a filter element to split the signal into several partial signals, and at least one analysis element (7) being provided for detecting speech information in the partial signals and at least one adapter element (8) for boosting the partial signals given the existence of speech, characterised by at least one envelope generator for generating the envelope of a partial signal and a mean value generator for generating the mean value of the partial signal and a comparison element for comparing the envelope with the mean value.

Images