EP1523219A2 - Method for training and operating a hearingaid and corresponding hearingaid - Google Patents

Abstract

The training of a hearing aid to individual situations should be made easier and more comprehensive for the hearing aid wearer. Therefore, it is provided that the hearing aid wearer only has to assign a current acoustic situation to a predetermined hearing situation identifier (3 '). This assignment learns a classifier, e.g. B. neural network (5). After training, the neural network (5) can then reliably assign the corresponding hearing situation identifier (3 ') to an acoustic input signal (2). Based on this assignment, a current parameter set (4 ') is varied or supplemented accordingly.

Description

The present invention relates to a method for exercising at night a hearing aid by providing an acoustic Input signal, providing a plurality of listening situation identifiers and assigning the acoustic input signal to a the Hörsituationskennungen by a hearing aid wearer. Moreover, the present invention relates to a corresponding Hearing aid, which is nachtrainierbar, and a method for operating such a hearing aid after a Retraining.

Classifiers are found in hearing aids for detecting different Situations use. The default parameters However, for the appropriate situations in an individual hearing aid wearer not necessarily be optimal. By night training, as is usual used in speaker-related speech recognition systems can, in certain situations, the detection rate regarding the individual framework conditions are increased. This is especially true for the situation of the performance of the own voice of great importance. Likewise, the classifier on certain noise situations, typical of the acoustic environment of the hearing aid wearer are optimally adjusted become.

In this connection, the document EP 0 681 411 A1 a programmable hearing aid known that independently adapts to changing environmental situations. The hearing aid parameters be continuously connected to the present Ambient noise adjusted, with "blurred" inputs from Hearing aid wearers in addition to the measured input signals can be used. The objective here is the parameters to optimize directly, taking into account the listening situation is not explicitly stated.

The document EP 0 814 634 A1 further describes a method described with which the hearing aid wearer the hearing aid optimally adjusts itself by a caused by the hearing aid wearer Night training is performed. The hearing aid wearer gets a lot of predefined parameter sets for that one Listening situation, which he tells the hearing aid to choose from. Out this limited set of parameter sets, each one Hearing aid preset, it selects the one he feels as optimal for himself. The appropriate hearing aid setting is learned from a set of rules, so that for a similar acoustic input signal, the same hearing aid setting he follows. This means that the rules a picture of the acoustic input variables on the optimal Hearing aid parameter set. At this night training if the hearing situation is taken into account indirectly, that only those corresponding to this hearing situation Parameter sets are made available for selection. A direct adaptation of the listening situation to the acoustic Input data does not occur. The disadvantage is that the hearing aid wearer during such night training the Sound of the hearing aid, by the parameter set used is determined, must judge. For example, he has to judge whether he gets the sound brighter or darker want. For certain complex algorithms and dynamic, adaptive parameters, such as B. the control of an adaptive Directional microphones, however, is the differentiation between different parameter sets difficult for the hearing aid wearer, if not impossible.

The object of the present invention is therefore to the night training of a hearing aid for the hearing aid wearer Simplify and operate the hearing aid accordingly improve.

According to the invention, this object is achieved by a method for training a hearing aid by providing a Acoustic input signal, providing multiple Hörsituationskennungen and assigning the acoustic input signal to one of the Hörsituationskennungen by a hearing aid wearer as well as automatic learning of the assignment of the acoustic Input signal to the one of the Hörsituationskennungen.

Furthermore, the invention provides a hearing aid with a Recording device for recording an acoustic input signal, a memory device for storing a plurality of Listening situation identifiers and an input device for assigning the acoustic input signal to one of the Hörsituationskennungen by a hearing aid wearer and a learning device to automatically learn the assignment of the acoustic Input signal to the one of the Hörsituationskennungen through the input device.

Moreover, the above object is achieved by a method for operating a hearing aid by receiving a hearing aid acoustic input signal, automatic assignment of a a hearing situation characterizing Hörsituationskennung to the acoustic input signal and automatic setting of the Hearing aid as a function of Hörsituationskennung.

The invention is based on the knowledge that for the hearing aid wearer although the differentiation between different Parameter sets is difficult, whereas calling a acoustic situation that is currently occurring, e.g. For example, the situation "own voice" or "entertainment in the car", in the Most cases very reliable by the hearing aid wearer is possible. These situations go beyond the conventional ones Hearing aids used hearing situations such as "speech at rest" and "speech in noise". Ie. the more differentiated Listening situations can turn on for signal processing relevant aspects of these "classic" situations Respectively. The acoustic representations that this novel, may be based on more diverse situations simple manner individually trained by specific naming become. For example, the sound of your own Voice or the specific sound of your own car from the hearing aid for example, be learned by a neural network. The neural network thus takes in contrast to the mentioned Prior art according to EP 0 813 634 A1 an illustration the acoustic input variables do not affect the resulting Overall setting (parameter set) of the hearing aid, but on the internal situation representation (hearing situation identification) in front. This will then be based on audiological Expert knowledge of the hearing aid parameter set to be used derived or the relevant parameters varied and / or supplemented. In particular, the adaptive algorithms reuse this information without the hearing aid wearer evaluate the result. This simple assignment the acoustic input signal to predetermined listening situations is because of the adaptivity of the algorithms and the so associated temporal dynamics are far less difficult for the hearing aid wearer as the direct sound evaluation, such as for example, the frequency response or the compression ratios / knee points, according to the prior art.

In a specific embodiment according to the invention can one of the listening situations of the performance of one 's own voice Hearing aid wearer, so after the automatic Learning your own voice can be recognized. This is in many situations, for example, for the directional microphone setting of great importance.

The automatic learning of the at least one Hörgeräteeinstellparameters for the assigned listening situation on the Basis of automatic evaluation can be during (online) or after (offline) the presentation of the acoustic input signal respectively. In online night training, the acoustic Input signal can not be completely saved, however The hearing aid requires more computing power to the night training perform. There is no need for offline night training this additional computing needs in the hearing aid, however, is a Memory device for the acoustic input signal needed. The online evaluation avoids the time-consuming reading, Processing and reprogramming the data or the Hearing aid.

The input device for assigning the acoustic input signal An auditory situation can also be used to start and stop of the night training. This will be the Handling the hearing aid or carrying out the night training simplified for the hearing aid wearer.

In addition, the input device may consist of one in the Hearing aid integrated receiver and an external remote control consist. The remote control can be for wired or wireless be designed wireless communication with the hearing aid. It is also conceivable that the remote control exclusively is used for the night training of the hearing aid. alternative can the remote control as a multifunction device, for example as a mobile phone or portable computer with radio interface speed, be designed.

The input device may also include a programmable arithmetic unit, in particular, include a PC, so that the operation via an appropriate programming software.

Finally, the input device at a special Embodiment verbal and in particular by means of one or several keywords be operable. This will be the operation the hearing aid for the hearing aid wearer even more comfortable designed.

Furthermore, the acoustic input signal can be a manual or automatically processed speech signal. That's it possible to train the classifier very specific.

When operating the hearing aid, d. H. after the night training, can a currently valid parameter set by the automatic Assigning the current hearing situation to a hearing situation identifier to be influenced. In particular, a parameter of the parameter set is varied by the automatic assignment and / or supplemented. This makes it possible for that acoustic input signal of a complex signal processing is subjected based on expert knowledge when the neural Net a learned hearing situation, z. B. your own Voice, recognizes. This is the currently used in the hearing aid Parameter set modified accordingly or equivalent Filtering performed.

FIG 1

ein Blockschaltbild zum Verfahren gemäß dem Stand der Technik;

FIG 2

ein Blockschaltbild des erfindungsgemäßen Verfahrens;

FIG 3

eine prinzipielle Darstellung eines Hörgeräts mit Fernbedienung zum Eingeben einer Hörsituation in einem ersten Schritt; und

FIG 4

die Situation des Hörgeräts gemäß FIG 3 während der Trainingsphase.

The present invention will now be explained in more detail with reference to the accompanying drawings, in which:

FIG. 1

a block diagram of the method according to the prior art;

FIG. 2

a block diagram of the method according to the invention;

FIG. 3

a schematic representation of a hearing aid with remote control for entering a hearing situation in a first step; and

FIG. 4

the situation of the hearing aid according to FIG 3 during the training phase.

The embodiment described in more detail below a preferred embodiment of the present invention For better understanding of the invention, however, first the procedure for night training according to the state of Technique explained in more detail with reference to FIG 1.

The hearing aid wearer or user 1 is located, as in FIG. 1 shown in a special acoustic situation, in the hearing aid an acoustic input signal 2 available stands. Since the hearing aid for the hearing aid wearer 1 is subjective not optimally adjusted, he takes a night training in front. He classifies the noise and tells the hearing aid the corresponding very general hearing situation or hearing situation identification, z. As "speech in noise", with. Everyone Each of these listening situations 3 is a plurality of parameter sets 4 assigned. Due to the selected listening situation 3, the hearing aid wearer 1 has, for example, seven parameter sets to select. He can now choose the parameter set 4 with which the hearing aid is set, that in this acoustic situation the subjectively best Sound produced.

A neural network 5 learns the desired parameter set 4 for the applied acoustic input signal 2, so it this parameter set 4 also for a similar acoustic situation will choose again after the training phase.

The subjective assessment of the sounds caused by the different parameter sets for the hearing aid setting, is very difficult for the hearing aid wearer 1, since this has a lot of detailed knowledge about the effects of hearing aid parameters presupposes.

According to the present invention is therefore intended by the hearing aid not the use of special parameter sets, but only the recognition of the current situation trains become. This is done according to the method of FIG. 2. Again, the hearing aid wearer or user 1 receives the acoustic Input signal 2. For night training of the neural network 5 in the hearing aid, the hearing aid wearer 1 the acoustic situation, in which he is currently, only one of a variety of predetermined, specific listening situations 3 'assign. The number of specific listening situations 3 ' is usually larger in the case of the present invention as the number of general listening situations 3 according to FIG. 1, because they should be differentiated from the beginning. Of the The reason for this is that the general listening situation "Speech in noise", for example, the specific hearing situation "own voice" includes.

The neural network 5 does not learn the assignment of a Parameter set to the acoustic input signal 2, but the assignment of a differentiated hearing situation or one Hörsituationskennung 3 'to the acoustic input signal. 2 (see arrows with solid lines in FIG. This means, that the neural network in contrast to the prior art learns at a higher level. On the example of the listening situation "own voice in your own car" is explained in more detail. According to the state of the art, this complex situation a fixed parameter set starting, for example, from assigned to the parameter set group "Language in noise". Because only a number of parameter sets are available for selection Hearing aid wearer for such situations "speech in noise" is meaningful, certainly none of the available standing parameter sets for own voice and additionally optimized for your own car.

According to the invention, on the other hand, the situation "own voice" and learned the other situation "in your own car" separately. These listening situations each have a specific influence the complex signal processing. For example, at the situation "own voice" a specific reinforcement, possibly coupled with a special setting of Directivity of the hearing aid, and in the situation "in his own Auto "again a very specific noise reduction caused in the hearing aid.

It is particularly advantageous that your own voice from the hearing aid can be learned. This is done by the acoustic Input signal with the voice of a special one Processing is subjected and appropriate parameters for the hearing aid specifically set and the hearing situation "own The same applies to learning, for example the listening situation "own car", creating a very specific noise suppression can be achieved can. So here, learning is not just the input signal assigned to a listening situation, but it also finds a very specific determination of parameters, such as Filter or gain parameter, instead.

When using the hearing aid after night training is the neural network 5 is an acoustic input signal 2 one or assign a plurality of specific listening situation identifiers 3 ', so that the currently valid parameter set 4 '(including Filter parameter) is influenced accordingly. A complex Signal processing unit 6 z. B. with adaptive directional microphone is the signal processing based on the influenced Carry out parameter set 4 '. If the neural network According to the example above, the input signal "own Voice in your own car ", it shows this to both the Listening situation identification "own voice" as well as the Hörsituationskennung "in your own car" too, so the current parameter set for example, in terms of specific reinforcement for your own voice and specific Filtering to suppress the noise of your own car varied or supplemented.

Below are two concrete application examples of the present Invention shown:

Example 1: An adaptive directional microphone is aimed at the Direction from which the maximum useful sound, z. B. a Speech signal, is incident. Entertains the hearing aid wearer with a person walking beside him, that should be Set directional microphone on the conversation partner, d. H. on a maximum gain at an angle of about 90 °. As soon as However, the hearing aid itself speaks, comes the useful sound signal from the own mouth, d. H. from an angle of 0 °. The own language thus draws the directional microphone characteristic away from the actual caller and usually with a certain delay. If that Hearing aid, however, is trained on one's own voice, and thus the adaptive microphone control is known which acoustic properties can belong to one's own voice Signals classified as "own voice" for the tracking of the directional characteristic is disregarded stay. In contrast, the adjustment of the hearing aid would be according to the prior art of EP 0 814 634 A1 according to FIG 1, after which the hearing aid wearer assess several sets of parameters would have, due to the dynamics and adaptivity of the processes little promise. In particular, your own could Voice can not be recognized.

Example 2: A noise canceling method may specifically be used trained on a complex, time-varying noise become. This noise is then optimally suppressed, although it may have similar spectral components or a Modulation spectrum such as speech, which further processed as a useful signal should be, has. Through individual training to this acoustic situation, z. B. the above-mentioned Situation "in the car", can the noise reduction method automatically optimally adjusted by, for example special weighting factors for single spectral Tapes are set or the dynamic behavior optimally adapted to the noise characteristic. Also in this case are the differences between the settings dynamic noise canceling only difficult to assess directly, but the situation is very reliable.

In certain acoustic situations it may be advantageous if in addition to the night training according to the invention Night training according to the state of the art under assessment various parameter sets done by the hearing aid wearer.

The night training, as it is for the hearing aid wearer, will now be explained in more detail with reference to FIGS 3 and 4. Of the Hearing aid wearer wants, for example, the situation "own Voice "into his hearing aid 10. He concludes via a line 11, a remote control 12 to the hearing aid 10th at. The remote control has a button 13 as a control on.

In the classifier several listening situations are stored. The hearing aid wearer knows that the hearing situation "own Voice "for example corresponds to situation 3. Therefore he pushes button 13 three times to tell the classifier that situation 3 should be retrained.

In a subsequent step, an acoustic signal, here your own voice, for recording for the hearing aid 10 according to FIG 4 presented. The hearing aid wearer must now the hearing aid 10 announce the beginning and the end of the training phase. This is done by holding down the button 13 while he speaks himself. This means that he steps for both use only a single control element 13 of the training got to. If there are many hearing situation identifications, you can another design be more user-friendly, for B. with a Display and a slider (slider, trackball, etc.), with which the corresponding situation can be selected quickly can.

The actual night training of the hearing aid 10 may be during the performance of the acoustic signal 14 take place. alternative the acoustic signal 14 is recorded in the hearing aid and evaluated after the recording or the selected listening situation due to characteristic acoustic properties assigned. In the case of online night training is one permanent or temporary storage of the acoustic signal 14 not necessary.

Since the hearing aid 10 only the information about the current situation must be communicated, unlike the state the technique according to EP 0 814 634 A1 an external operating unit not necessary. It can, however, according to the 3 and 4 are used for example for reasons of comfort. However, it can also be a recording button on the hearing aid itself to be appropriate.

By night training, the recognition rate of the Classifier for certain situations clearly opposite Increase the presetting, so that the hearing aid more reliable to adjust to this situation. By the independent Starting and ending the night training phase by the Hearing aid wearers can also reliably relieve situations because the hearing aid wearer decides when the signal can be assigned to the situation.

Claims (27)

A method for training a hearing aid (10) by Providing an acoustic input signal (2), Providing a plurality of listening situation identifiers (3 ') and Assigning the acoustic input signal (2) to one of the hearing situation identifications (3 ') by a hearing aid wearer (1), marked by automatically learning the assignment of the acoustic input signal (2) to the one of the listening situation identifiers (3 '). The method of claim 1, wherein one of the listening situation identifiers (3 ') a listening situation of the performance of their own Voice or noise of the car of the hearing aid wearer (1), so that after automatic learning your own voice or the sound of your own car can be recognized. Method according to claim 1 or 2, wherein learning during the performance of the acoustic input signal (2) takes place. The method of claim 1 or 2, wherein learning to the performance of the acoustic input signal (2) takes place. Method according to one of the preceding claims, wherein Starting and stopping the night training and assigning of the acoustic input signal (2) via a remote control (12). Method according to claim 5, wherein the remote control (12) wirelessly communicates with the hearing aid (10). Method according to one of claims 1 to 4, wherein the operation of the hearing aid (10) for exercising night verbally. The method of claim 7, wherein the operation of the hearing aid (10) by means of one or more keywords. Method according to one of the preceding claims, wherein the acoustic input signal (2) on manually or automatically processed voice signal includes. Hearing aid with a recording device for recording an acoustic input signal (2), a memory device for storing a plurality of listening situation identifiers (3 ') and an input device (12) for assigning the acoustic input signal (2) to one of the hearing situation identifications (3 ') by a hearing device wearer, marked by a learning device (5) for automatically learning the assignment of the acoustic input signal (2) to the one of the hearing situation identifiers (3 ') by the input device (12). Hearing aid according to claim 10, wherein one of the Hörsituationskennungen (3 ') the listening situation of the performance of their own Voice or noise of the car of the hearing aid wearer indicates that after automatic learning recognize your own voice or the sound of your own car can be. Hearing aid according to claim 10 or 11, wherein the learning in the learning device (5) during the recording of the acoustic Input signal (2) in the receiving device feasible is. Hearing aid according to one of claims 10 to 12, wherein the Learning after recording the acoustic input signal (2) in the recording device in the learning device (5) or in an external device with subsequent transmission in the hearing aid is feasible. Hearing aid according to one of claims 10 to 13, wherein the Input device (12) for starting and stopping the recording or the Nachrainierens is usable. Hearing aid according to one of claims 10 to 14, wherein the Input device (12) has an external remote control. Hearing aid according to claim 15, wherein the remote control wireless communication with the hearing aid is feasible. Hearing aid according to claim 15 or 16, wherein the remote control designed exclusively for the night training of the hearing aid is. Hearing aid according to one of claims 10 to 16, wherein the Remote control is designed as a mobile device. Hearing aid according to one of claims 10 to 18, wherein the Input device (12) a programmable arithmetic unit, in particular a PC. Hearing aid according to one of claims 10 to 19, wherein the Input device (12) is verbally operated. Hearing aid according to claim 20, wherein the input device (12) operable by means of one or more keywords is. Hearing aid according to one of claims 10 to 21, wherein the Acoustic input signal (2) on manually or automatically processed voice signal includes. A method for operating a hearing aid Receiving an acoustic input signal (2), marked by automatic assignment of a hearing situation identification (3 ') characterizing a hearing situation to the acoustic input signal (2) and automatic adjustment of the hearing aid as a function of the hearing situation identifier (3 '). The method of claim 23, wherein a currently valid Parameter set (4 ') for adjusting the hearing aid by the automatic assignment of Hörsituationskennung (3 ') influenced becomes. The method of claim 24, wherein at least one parameter of the parameter set (4 ') by the automatic assignment varied and / or supplemented. Hearing aid according to one of claims 10 to 22, which is a Signal processing device, whose parameter set (4 ') by means of the learning device (5) by assigning the acoustic input signal (2) to a learned Hörsituationskennung (3 ') can be influenced. Hearing aid according to claim 26, wherein at least one parameter of the parameter set (4 ') by the learning device (5) during the automatic assignment can be varied and / or supplemented.

Images