EP3396978B1 - Hearing aid and method for operating a hearing aid - Google Patents

Description

The invention relates to a method for operating a hearing device which comprises a microphone for converting ambient noise into a microphone signal, a signal processor for processing the microphone signal into an output signal and an output converter for outputting the output signal to the hearing of a wearer of the hearing device. Furthermore, the invention relates to such a hearing device.

Hearing devices are usually used to output a sound signal to the hearing of the wearer of this hearing device. The output takes place by means of an output transducer, usually acoustically via airborne sound using a loudspeaker (also referred to as "receiver" or "receiver"). Such hearing devices are often used as so-called hearing aids (also briefly: hearing aids). For this purpose, the hearing devices normally comprise an acoustic input transducer (in particular a microphone) and a signal processor which is set up to process the input signal (also: microphone signal) generated by the input transducer from the detected ambient noise using at least one signal processing algorithm which is usually stored and / or adapted in a user-specific manner to process that a hearing loss of the wearer of the hearing device is at least partially compensated. In particular in the case of a hearing aid device, the output transducer can alternatively be a so-called bone conduction receiver or a cochlear implant in addition to a loudspeaker, which are set up to mechanically or electrically couple the sound signal into the hearing of the wearer. The term hearing device also includes devices such. B. so-called tinnitus maskers, headsets, headphones and the like.

Modern hearing devices, in particular hearing aids, often include a so-called classifier, which is usually designed as part of the signal processor. This classifier is usually implemented by an algorithm which is used to infer an existing listening situation on the basis of the ambient noise detected by means of the microphone, and thus to enable the respective signal processing algorithm or the respective characteristics of the present hearing situation to be changed. For this purpose, the respective signal processing algorithm usually has a number of changeable parameters. The assigned parameter values are then changed depending on the respective listening situation. Usual listening situations are e.g. B. the presence of speech, the presence of music, (driving in) the vehicle and the like. As is known, the noise present in each case varies, for example (engine and / or wind noise in the vehicle, for example). Likewise, for a clear sound of the music, in contrast to a general operation, for example, other frequency ranges must be amplified (raised) and / or attenuated (lowered) differently, for example in order to better reproduce overtone ranges that are often required for music for a clear sound.

Typically, the classifier is set up to interpret a listening situation recognized in the "foreground" as a "primary" listening situation when listening situations are mixed with one another and thus to ignore or ignore other, simultaneous listening situations. This can disadvantageously lead to incorrect classifications and thus to at least subjectively unfavorable settings of the signal processing algorithm. Such misclassifications of hearing situations can occur, for example, if music with a dominant component in the recorded ambient noise is recognized, but voice components, on the other hand, only with a minor component (for example, if the hearing device wearer is close to a loudspeaker in a music system, but is further away) Third party speaks). In this case, the ambient noise is mostly dominated by music.

To enable conversation between two people, each wearing hearing aids, in a noisy environment, in EP 3 101 919 A1 describes a hearing system formed from the respective hearing devices. The hearing device assigned to one of the two people detects the person's own voice by means of a directional microphone and transmits this to the other hearing device, in particular wirelessly. So this corresponds to a kind of radio connection. In EP 2 603 018 A1 describes, for example, how it can be recognized whether the wearer speaks a hearing device.

The invention has for its object to improve the comfort of use of a hearing device.

This object is achieved according to the invention by a method for operating a hearing device with the features of claim 1. Furthermore, this object is achieved according to the invention by a hearing device with the features of claim 17. Advantageous and partly inventive embodiments and further developments of the invention are in the Subclaims and the description below.

The method according to the invention serves to operate a hearing device, preferably a hearing aid device. The hearing device comprises a microphone for converting ambient noise into a microphone signal, a signal processor for processing the microphone signal into an output signal and an output converter for outputting the output signal to the hearing of a wearer of the hearing device.

According to the method, a characteristic speech measure is determined for a speech component contained in the ambient noise. It is preferably determined whether such a speech component is present and then in particular assigns a current value to the characteristic speech measure. A voice activity detection module comprised by the signal processor is preferably used here. Furthermore, a characteristic additional measure is preferred for an activity of the wearer of the hearing device (in short: “hearing aid wearer”) specifically a current value for this is determined. Here, too, it is preferably determined whether the activity is present and then the current value is assigned to the additional dimension. A relationship between the language measure and the additional measure is then evaluated, in particular determined. On the basis of this evaluation, a probability value - that is, a value of a probability - for the existence of a communication situation between the hearing aid wearer and a third party is increased if the speech measure and the additional measure under at least one predetermined criterion (in particular in each case) one for the presence of the speech component or for the activity of the hearing device wearer assume a representative value. Depending on the probability value for the existence of the communication situation, at least one - preferably predetermined (ie preferably factory-set or user-specifically preset) - signal processing algorithm, which is processed to process the microphone signal, is changed.

To change the or the respective signal processing algorithm, preferably (in particular in the signal processor) at least one parameter, preferably a parameter value assigned to it, is changed, preferably set to a (other) parameter value assigned to the probability value.

The term "communication situation" is understood here and below to mean a conversation, specifically that the hearing aid wearer speaks to a third party. In other words, a communication situation only occurs when at least two people, one of whom is the hearing aid wearer, speak to each other.

“Increase probability value” is understood here and in the following in particular to mean that, compared to not fulfilling the at least one criterion, an increased probability for the existence of the communication situation is assumed and the probability is therefore assigned a correspondingly increased value.

Characteristic here and below means in particular that the speech measure and the additional measure each provide information as to whether (in particular with what probability) a speech component is contained in the ambient noise or whether the hearing aid user is active. Optionally, in addition (in particular as an alternative to specifying the probability), the speech measure and / or the additional measure also provide quantitative information about the speech component (for example a percentage of the total microphone signal) or about the activity. The value representative of the presence of the speech component or the activity is preferably formed by the corresponding probability value itself and / or from a value derived therefrom, for example by means of a threshold value comparison. For example, a probability value (for the presence of the language component or the activity) is only taken into account as a representative value when the threshold value is exceeded.

In an alternative variant, the language measure and / or the additional measure each reflect qualitative information about the presence of language components or the activity. In this case, the language measure or the additional measure is a type of binary or "Boolean" variable that can only assume two states. The state ("first") that is representative of the presence of the language component or the activity is set as "true" (or: "true", "one" or an equivalent, unambiguous designation). Accordingly, the "second", opposite state is set as "false" (or: "false", "zero" or the like) in this case.

The invention is based on the finding that communication situations, ie actual conversations between two people, cannot be clearly or at least not clearly identified with sufficient certainty by means of a known classifier. The fact that the presence of speech in general (ie regardless of the identity of the speaking person) and an activity of the hearing device wearer and that a relationship between these two variables is considered and evaluated can be advantageously achieved in two steps, which are preferably carried out independently of one another a communication situation even if there are several superimposed ones Listening situations, ie when there is not only background noise, but z. B. with dominant music, traffic noise or the like. Advantageously, this may also reduce the need for the hearing aid wearer to make an active change to the settings of their hearing device (for example, due to a misclassification). This in turn advantageously increases the comfort of use of the respective hearing device.

In a preferred variant of the method, the respective signal processing algorithm is continuously changed depending on the probability value for the presence of the communication situation. This is also referred to as "dynamic" starting of a hearing aid setting. For example, the respective parameter value of the respective signal processing algorithm is changed step by step or continuously - d. H. For example, with a comparatively low probability value, there is also a correspondingly slight change. In particular, the or the respective parameter value is represented by a function of the probability value for the existence of the communication situation. As a result, a particularly smooth transition, which is barely noticeable, or at least subjectively pleasant, for the hearing aid wearer in particular, can advantageously be made between different settings - i. H. differently aligned signal processing - enables the hearing device. Artifacts when changing the settings can advantageously be reduced.

In one variant of the method, a measure for the orientation of the head of the hearing device wearer, in particular for a head movement, is used as an additional measure for the activity of the hearing device wearer. For example, this measure is determined by means of a position sensor, in particular by means of an acceleration sensor (preferably included in the hearing device) and / or a gyroscopic sensor (in particular a rotation sensor). When evaluating the relationship, it is determined, for example, on the basis of the additional measure whether the hearing device wearer turns his head, in particular, to a source of the speech component to at least listen to the person who represents this source.

In a particularly preferred variant of the method, as a supplementary measure for the activity of the hearing device wearer, a native language measure characteristic of the presence of the hearing aid wearer's own language, preferably specifically a current value for it, is determined. In particular, it is determined whether the user's own language is available (i.e. whether the hearing aid wearer is speaking himself) and then the current value is assigned to the native language measure. For this purpose, a native language detection module, which is included in particular by the signal processor, is preferably used. Accordingly, the relationship between the language measure and the native language measure is preferably evaluated thereon to determine the probability value for the existence of the communication situation. This means that preferably the language measure and the native language measure are put in relation to each other. The fact that it is determined independently of one another whether there is language and whether the hearing aid wearer speaks and whether a relationship between language and one's own language is considered and evaluated, can advantageously increase the precision of the detection of a communication situation even in the presence of several superimposed hearing situations. The presence of one's own language gives a comparatively strong indication that the hearing aid wearer is taking part in a conversation.

A temporal relationship between the language measure and the additional measure, in particular the native language measure, is used, ie viewed or analyzed, as a predetermined criterion for the existence of the communication situation. For example, it is determined whether the head movement takes place immediately before or after (ie within a few seconds, for example 1 to 3 seconds) the speech measure assumes the representative value. In particular, it can be concluded from this that the hearing aid wearer looks at his interlocutor before he addresses them himself and / or after the interlocutor has addressed the hearing aid wearer.

In a preferred variant of the method, the probability value for the existence of the communication situation is increased if a temporal difference between the language measure and in particular the native language measure is determined, preferably specifically between the respective value representative of the presence of language or one's own language. That is, it is determined whether the respective representative values are available for different time periods. This is based on the knowledge that the own language of the hearing aid wearer and the language of a third party (i.e. the conversation partner) are usually not available at the same time in a conversation.

In a preferred development of the above method variant, the probability value for the existence of the communication situation is increased in particular if the language measure and the native language measure assume the value representative of the presence of the language component or one's own language within a predetermined period of time. In this case, it is thus checked in particular whether the language measure and the native language measure indicate the actual presence of the language of a third party or the own language of the hearing device wearer in a manner characteristic of a conversation between two people. For example, when the speech portion (or the value representative of this) is detected as a predetermined time period, a time window is started, within which it is checked whether the native language measure indicates the presence of one's own language for a shorter time (i.e. time duration) than the language measure does Indicates the presence of a language component. In this development, the specified criterion is, in particular, that the language measure and the native language measure indicate the actual presence of the language of a third party or one's own language within the specified time period.

The method variant described above is particularly expedient in that the presence of any language in the ambient noise is detected by means of the above-mentioned voice activity detection module, ie also the presence of the hearing aid wearer's own language. The native language measure (preferably using the native language detection module described above on the other hand, only indicates the presence of the hearing aid wearer's own language. It is therefore expedient to conclude that the language of a third party (ie that a third party speaks) if in a phase (ie in particular within a part of the period described above) only the language measure assumes the value representative of the language component. If, on the other hand, the language measure and the native language measure have the respective value representative of the presence of the assigned language, it is in particular concluded that the hearing device wearer actually speaks himself. The phases in which the hearing aid wearer or a third party speak in each case are expediently determined as part of this method variant. As a further development of this variant of the method, a contravalence (or a value for this) between the hearing aid wearer's own language and the language of a third party is determined as a criterion for increasing the probability value for the presence of the communication situation. This means that, in particular, the probability value for the presence of the communication situation is increased if either a third party or the hearing aid wearer speaks, but preferably also both speak within the predetermined period of time.

The predetermined time period is preferably at least several seconds, for example 3 to 5 or up to 30, in particular about 20 seconds. The shorter the time period, the faster the respective signal processing algorithm can be changed, so that the hearing aid wearer only has to wait for short reaction times for the signal processing to be adapted to the changed hearing situation. On the other hand, however, a certain duration is also required in order to assume, with sufficient certainty, a certain probability of the existence of the communication situation, in particular to recognize a conversation between the hearing aid wearer and a third person and, in particular, to be able to distinguish between a "mere" greeting exchange between two people.

In an expedient variant of the method, the presence or the detection of the communication situation, in particular on the part of the signal processor excluded - preferably by not increasing the probability value or setting it to a value close to "zero" or from "zero" - if only (ie exclusively) the language measure assumes the value representative of the presence of the language component. This can advantageously prevent the signal processing algorithm from being changed when, for example, there is talk on TV or on the radio. In an optional development of this variant of the method, the presence of the communication situation is particularly excluded if, for example, the native language dimension only takes on the value representative of the presence of the hearing aid wearer's own language for negligibly short durations, while the language dimension in particular continuously or for a multiple - for example at least ten times - longer duration has the value representative of the presence of the language component. Such a temporal distribution of the respective representative values of the language measure or the native language measure is also referred to in this context as "strongly asymmetrical".

In a further expedient variant of the method, the existence of the communication situation is ruled out if the language measure and the native language measure in particular always essentially (i.e. exactly or with only minor deviations) running in the same direction (ie preferably simultaneously and for the same duration in each case) that for the presence of the language component or assume a representative value for the hearing aid wearer's own language. In this case, it can be, for example, a soliloquy, a "singing along" of the hearing aid wearer to a piece of music played, and the like. In this case, however, there is usually no communication situation. H. the hearing aid user does not have a conversation with a third party.

In an advantageous variant of the method, phases of a preferably predetermined duration (for example from 1 to a maximum of 5 seconds) in which the language measure does not assume a value representative of the presence of the speech component are not taken into account within the predetermined time period described above. Such a procedure can advantageously ignore speech pauses that usually occur in every conversation without the The probability value is erroneously reduced or the monitoring of the specified time period is interrupted. Because the language measure also indicates the presence of native language, such a pause in the normal functioning of the entire hearing device can therefore only occur when there is no language at all - ie the native language dimension does not assume the representative value.

In a further expedient method variant, the (optionally further) predefined criterion is used, which preferably speaks at least twice within the predefined period of time of the hearing aid wearer (in particular with a pause in between, in which a third party speaks). In this case, the probability value for the existence of the communication situation is increased if the native language measure assumes the value representative of the existence of one's own language several times within the specified period. Additionally or alternatively, the probability value is increased if a third party speaks at least twice within the predetermined time period described above - preferably with a pause in between, during which the hearing aid wearer speaks. As a (optionally additional) predetermined criterion, it is used in the context of this method variant that within the predetermined time span the language measure and / or the native language measure multiple times, ie. H. at least twice assume the value representative of the presence of the language portion or the wearer's own language. In particular, within the scope of this variant of the method, it is checked during the aforementioned period of time whether an (in particular alternating) "voice sequence" (also referred to as "alternating sequence") takes place at least of the type "hearing aid wearer - third party - hearing aid wearer" or "third party - hearing aid wearer - third party". In this way, for example, it can advantageously be prevented that a simple call with a simple answer, for example also a greeting with a subsequent reply, leads to an increase in the probability value for the communication situation.

In an expedient further development, in particular of the method variant described above, the (optional additional) criterion for increasing is used the probability value for the existence of the communication situation that the hearing aid wearer and / or the third party speak more than twice.

In an optional method variant, the (in particular further) criterion is used that within the time period described above, the respective temporal portions for which it is determined (or a particularly high probability is assumed) that the hearing aid wearer or the third party speak, roughly ( ie preferably with deviations of 10 to 30%). That is, the probability value for the existence of the communication situation is increased (in particular further) when the relationship between the language of the hearing aid wearer and the third party is approximately balanced. In particular, the additional criterion specifically determines whether the duration (possibly the sum of the individual durations detected within the time span, i.e. the total time) for which the language measure assumes the value representative of the presence of the speech component, for example (i.e. within of the aforementioned range) approximates twice the duration (or its total or total time) for which the native language dimension takes on the associated representative value.

In a further optional method variant, several criteria (which are described here in particular here and below) are preferably used to increase the probability value for the presence of the communication situation. In particular, the probability value is increased as a function of several fulfilled or, in particular, only partially fulfilled (e.g., if a multiple change sequence is required after a simple change) - in the case of only partially fulfilled criteria, preferably by a correspondingly smaller predetermined step size. For example, if a smaller number of criteria than the predetermined number are met, the probability value can be set so high that a limit value that is optionally specified for a change is exceeded and then a change in the signal processing algorithm (preferably associated with the current probability value, in particular correspondingly less) is made. In this case, for example, the probability value for the existence of the communication situation is increased if a multiple alternation sequence was recognized within 10 seconds within the predetermined time period, for example 20 seconds, in such a way that the signal processing algorithm is changed. In other words, a cumulative fulfillment of several criteria leads to a cumulative or optionally also progressive increase in the probability value. In the latter case, the probability value for each further fulfilled criterion is increased with a larger step with an increasing number of fulfilled criteria than when this criterion is fulfilled alone. By taking into account that several criteria (possibly also only partially) are met, a faster, ie particularly early, change of the respective signal processing algorithm is advantageously made possible.

In an expedient embodiment, the hearing device comprises a classifier for recognizing different (in particular additional hearing situations which differ from the communication situation), preferably for assigning probability values for the presence of the respective hearing situation. In this case, in a preferred variant of the method, in particular when a predetermined threshold value for the probability value for the presence of the communication situation is exceeded, the signal processing algorithm or algorithms are changed with a higher priority for the communication situation than other hearing situations recognized in parallel. This advantageously makes it possible, in particular in ambiguous hearing situations (ie when there are several hearing situations at the same time) to prioritize the conversation of the hearing device wearer, and the signal processing is thus (if necessary primarily) oriented towards the communication situation. In particular, this means that conversations between the hearing aid wearer and third parties, which are overlaid, for example, by dominant music, traffic noise or the like, can be recognized as such particularly reliably and can lead to a change in the signal processing algorithm. As an alternative or optionally in addition to the aforementioned threshold value comparison, the change in the or the respective signal processing algorithm, in particular the respective parameter depending on the probability value for the communication situation, at least in ambiguous listening situations, weighted more than depending on the probability values for other hearing situations. That is, a change due to the communication situation is preferably more pronounced than a change due to other listening situations. This variant of the method is based in particular on the consideration that, even in the case of superimposed hearing situations, the hearing device wearer will very likely want to devote his attention to a conversation if he himself takes part in it.

In a preferred development of the method variant described above, in the case of ambiguous hearing situations, the signal processing algorithm or algorithms are also changed as a function of a plurality of recognized hearing situations (or at least in each case of a predetermined probability for the respective presence of these hearing situations). For example, a parameter aimed at a clear sound of music and a parameter aimed at speech intelligibility are changed if representative probability values are determined for music as well as for the communication situation for the respective presence. However, as described above, the parameter aimed at intelligibility of speech is preferably changed more than the parameter aimed at the clear sound of music.

In an expedient variant of the method, a “speaking” of the hearing aid wearer and / or the third party (in particular only) is then used as one for a conversation - ie. H. the "utterance" of the respective person associated with the language sequence described above, if it is recognized that the hearing device wearer or the third party speaks for at least 1 or 2 seconds, d. H. the respective representative value of the language measure or the native language measure must be continuously available for this duration. In particular, the criterion used (preferably another predetermined) is that the respective utterance by the third party and / or the hearing device wearer lasts for at least 1 or 2 seconds.

In a further expedient method variant, a (in particular additional) criterion is used that within a (in particular the specified Period of time) of the previous period of predetermined duration, the probability value for the existence of the communication situation has already shown a predetermined value. In particular, the aim here is to determine whether the signal processing algorithm (s) in question has already been adapted to the communication situation in the preceding two to five (optionally also up to ten) minutes. In particular, if (with several criteria to be fulfilled) at the same time part of the specified criteria is already fulfilled or only partially fulfilled (e.g. if only part of the specified language sequence is available, e.g. only one change from "hearing aid wearer - third party" or vice versa ), in this case the probability value for the existence of the communication situation can be increased at an earlier point in time or by a larger increment (or optionally the specified time span can be shortened from 20 seconds to 10 seconds, for example). In this case, it is assumed that the likelihood that the previous conversation will be resumed or that the hearing aid wearer will start a new one, for example at an event. Optionally, this method variant also takes into account how long the communication situation was in the previous period. E.g. It can be provided in the context of the invention that during a previous conversation of, for example, at least three minutes (i.e. such a long communication situation), the predetermined time period is shortened accordingly and / or a simple language change (instead of a multiple change) between the third party and the hearing aid wearer is sufficient to recognize the communication situation.

In an expedient variant of the method, a wearer-specific tendency to participate in conversations, ie whether the hearing aid wearer likes to talk a lot to third parties, for example, is taken into account to identify the communication situation. This is preferably taken into account by means of a corresponding adaptation of at least some of the criteria described above. E.g. the time period described above, within which the alternation sequence must be detected, in particular when the hearing device is adapted for the first time, if there is a strong tendency to participate in the conversation compared to an average one Hearing aid wearers set a shorter duration (for example 8 or 10 seconds instead of 20 seconds). Optionally, additionally or alternatively, a simple change sequence, that is to say a simple change between one's own language and the language of a third party, can also be used within the predetermined time period to identify the communication situation. A "more aggressive", ie faster or "steeper"("in larger steps") increase in the probability value for the presence of the communication situation, and thus a particularly early change of the respective signal processing algorithm can thus advantageously be made possible.

In a further expedient method variant, the (in particular additional) criterion is used that a streaming signal from a communication device (for example a cell phone or the like) is received. Such a streaming signal is usually present in particular when the hearing aid wearer actually wants to communicate.

In a further expedient method variant, which also represents an independent invention, a frequency number for the detection of the communication situation, ie in particular for the change, is preferably stored daily in a memory unit of the hearing device (for example using a clock integrated in the hearing device, specifically in the signal processor) of the respective signal processing algorithm as a function of the probability value for the existence of the communication situation. In other words, the hearing device records how often, in particular, the hearing device wearer communicates, ie participates in conversations, on one day. This frequency number, in particular its course over the duration of therapy by means of the hearing device, can provide information as to how well the therapy of the hearing device wearer “strikes”. The better the hearing aid wearer is supplied with communication-relevant audio signals, the more likely the hearing aid wearer will also take part in conversations (in particular in the future again), in particular get used to taking part in conversations again.

In yet another expedient method variant, a mixing ratio between the microphone signal and the audio signal is changed as a function of the probability value for the presence of the communication situation during playback of an external audio signal to generate the output signal. Usually in a "listening situation", i.e. H. while the audio signal is received as an (in particular non-acoustic) data signal from an external source (in particular a media device such as a TV set, a music system, a lecture audio system or the like), the microphone signal is suppressed when the output signal is generated or at least heavily damped, since it is regularly assumed that the priority of the hearing device wearer in such a situation lies on the external audio signal. By changing the mixing ratio - in particular by adding the microphone signal to the audio signal in the output signal - as a function of the probability value for the presence of the communication situation, a response and adaptation of the signal processing that is independent of a user input is advantageously made possible.

In the above variant of the method, the proportion of the microphone signal compared to the external audio signal is increased (in particular in comparison to the standard mix in the listening situation), in particular when the probability value for the existence of the communication situation for generating the output signal is increased. Optionally, the external audio signal is also withdrawn at the same time. H. damped or not amplified when generating the output signal. As a result, the attention of the hearing device wearer is advantageously automatically and in particular independent of a user input also extended to the present conversation, specifically the (speech) intelligibility for the person opposite is increased. In particular, this enables a particularly robust autonomous "switchover" (adaptation of the signal processing) from the listening situation to the communication situation.

In an advantageous method variant, if the probability value for the presence of the communication situation is reduced, that is, preferably when the requirements for the existence of the communication situation no longer exist - in particular after a predetermined waiting time has elapsed, the change in the signal processing algorithm (s), in particular the changed mixing ratio between the microphone signal and the audio signal, is reset. In other words, in this case the signal processing is preferably "brought back" to the previous setting, in particular if the classifier described above does not detect another new hearing situation. In the case described above of the listening situation prior to the detection of the communication situation and matched to the external audio signal, the external audio signal is (again) prioritized when the communication situation is no longer present.

As described above, the hearing device according to the invention comprises the microphone for converting the ambient noise into the microphone signal, the signal processor for processing the microphone signal into the output signal and the output converter, which outputs the output signal to the hearing of the hearing device wearer. The signal processor of the hearing device is set up according to the invention to carry out the method described above. In other words, the signal processor is set up for the speech component contained in the ambient noise, the characteristic speech measure, specifically its current value, and the additional measure characteristic for the activity of the hearing device wearer, in particular the own language measure, characteristic of the presence of the hearing device wearer's own language, specifically the latter current value. Furthermore, the signal processor is set up to evaluate a relationship between the language measure and the additional measure or the native language measure and, on the basis of this evaluation, to increase the probability value for the existence of the communication situation if the language measure and the additional measure or native language measure are based on at least one of the criteria described above assumes the respective value that is representative of the presence of the language component or the activity (especially one's own language). Furthermore, the signal processor is set up according to the invention as a function of the probability value to change the respective signal processing algorithm (s) for the existence of the communication situation.

The hearing device according to the invention has the same advantages that are described in connection with the method described above.

In a preferred embodiment, the signal processor is formed, at least in the core, by a microcontroller with a processor and a data memory in which the functionality for carrying out the method according to the invention is implemented in the form of operating software (firmware), so that the method - possibly in interaction with the hearing aid wearer - is automatically carried out in the microcontroller when the operating software is executed. Alternatively, the signal processor is a non-programmable electronic component, e.g. an ASIC, in which the functionality for implementing the method according to the invention is implemented using circuitry means.

In a preferred embodiment, the hearing device comprises the voice activity detection module described above (also referred to as "voice activity detector"), which is preferably designed as part of the signal processor. In a likewise preferred embodiment, the hearing device also includes the above-described own language detection module (also referred to as an “own voice detector”), which is preferably also designed as part of the signal processor.

In an expedient embodiment, the hearing device comprises an input interface which is set up to receive the external audio signal described above. For example, the input interface is an antenna, a jack plug, a plug connection for a data cable or the like.

The conjunction "and / or" is to be understood here and below in such a way that the features linked by means of this conjunction can be formed both jointly and as alternatives to one another.

Fig. 1

in einer schematischen Schaltübersicht ein Hörgerät,

Fig. 2

in einem schematischen Ablaufdiagramm einen Ablauf eines von der Hörvorrichtung ausgeführten Betriebsverfahrens,

Fig. 3

in einem Diagramm einen schematischen zeitlichen Verlauf, der bei Ablauf des Betriebsverfahrens gemäß Fig. 2 zur Erkennung einer Kommunikationssituation führt, und

Fig. 4

in Ansicht gemäß Fig. 1 das Hörgerät gemäß eine weiteren Ausführungsbeispiels.

Exemplary embodiments of the invention are explained in more detail below with reference to a drawing. In it show:

Fig. 1

a hearing aid in a schematic circuit diagram,

Fig. 2

a schematic flowchart a sequence of an operating method carried out by the hearing device,

Fig. 3

in a diagram, a schematic course of time, which according to Fig. 2 leads to the detection of a communication situation, and

Fig. 4

in view according to Fig. 1 the hearing aid according to a further embodiment.

Corresponding parts and sizes are always provided with the same reference symbols in all figures.

In Fig. 1 a hearing aid (short: “hearing aid 1”) is shown as the hearing device. The hearing aid 1 comprises, as electrical components, which are housed in a housing 2, two microphones 3, a signal processor 4 and a loudspeaker 5. To supply the electrical components with energy, the hearing aid 1 also comprises a battery 6, which alternatively acts as a primary cell (for example as a button cell) ) or as a secondary cell (ie as a rechargeable battery). By means of the microphones 3, ambient noises are recorded during operation of the hearing device 1 and a microphone signal S _{M is} generated therefrom. These two microphone signals S _M are fed to the signal processor 4, which generates an output signal S _A from these microphone signals S _M while processing a signal processing algorithm and outputs this to the loudspeaker 5. The loudspeaker 5 installs an output converter and converts the output signal S _A into airborne sound, which is transmitted to the housing via a sound tube 7 connected to the housing 2 and an earpiece 8 connected to the end Hearing of a wearer (short: hearing aid wearer) of the hearing aid 1 is output. The signal processor 4 additionally also comprises a classifier, by means of which the ambient noise contained in the microphone signals S _M is used to infer different listening situations (e.g. speech, music, vehicle and the like). Depending on the hearing situation detected, at least one specifically assigned parameter of the signal processing algorithm is changed. As a result, the signal processing of the hearing device 1 can be adapted to different listening situations, and the sound output of the loudspeaker 5 can be improved with regard to the spectral properties of the ambient noise that characterizes the respective hearing situation.

In order to be able to recognize a communication situation, ie a situation in which the hearing aid wearer is talking to a third party (ie speaking to this third party) and to be able to adapt the signal processing thereupon, one is used Fig. 2 Operating method described in more detail below in the hearing aid 1, specifically carried out by the signal processor 4. The signal processor 4 comprises a voice activity detection module (also: "voice activity detector" or "speech recognition VAD") for the detection of speech components in the ambient noise, as well as a native language detection module (also: "own voice detector" or "own language recognition OVD") for the detection of one's own Language of the hearing aid wearer.

In a first method step 10, a characteristic speech measure T is determined by means of the speech recognition VAD for a speech component contained in the ambient noise. The language measure T represents a probability value that reflects a probability for the presence of the language component. In a second method step 20, a characteristic native language dimension O is determined by means of the native language recognition OVD for the presence of the own language of the hearing device wearer (also short: HIW for "hearing instrument wearer"). Analogous to language measure T, this own language measure O also represents a probability value. In addition, the own language measure O is characteristic of an activity of the hearing device wearer, namely that the hearing aid user is speaking himself.

In an alternative embodiment, the language measure represents a binary variable that can only assume two values or states. These two values are representative of the presence or absence of the speech component in the ambient noise. For example, the speech measure T assumes the value "1" if a speech component is detected in the ambient noise. Conversely, the speech measure T assumes the value "0" if no speech component is detected in the ambient noise. The same applies to the native language O.

In a further method step 30, the language measure T and the native language measure O are brought together. If at least the language measure T has a (probability) value representative of the presence of the language components, a relationship between the language measure T and the native language measure O is evaluated. As part of this evaluation, a temporal behavior of the language measure T and the native language measure O is considered. In order to recognize a communication situation, that is to say whether the hearing aid wearer is speaking to a third party, the criterion is checked as to whether, within a period of time (referred to as "time window Z"), the native language dimension O is representative of the presence of the hearing aid wearer's own language Assumes value. In this case, it is also checked as a further criterion whether a duration D _O , for which the native language dimension O assumes the value representative of the existence of one's own language, of a duration D _T , for which the language dimension T represents the presence of the language components Value assumes, deviates, specifically whether the duration D _{O is} shorter than the duration D _T (cf. Fig. 3 ). This situation indicates that both the hearing aid wearer's own language - cf. Diagram in Fig. 3 , "Line" HIW - and the language of a third party - cf. Diagram in Fig. 3 , "Line" OT (short for "other talker") - (specifically for different periods). Because it is recognized that the language measure T indicates both the presence of the third part's speech components and the presence of the hearing aid wearer's own language. As an additional criterion, it is also checked whether within the time window Z the native language measure O or (optional and) the language measure T, at least twice that for assuming the presence of your own language or the language portion representative value (cf. Fig. 3 ). This indicates an "exchange conversation" between the hearing aid wearer and the third party, so that a conversation, ie the communication situation, is comparatively high. As a further criterion, it is also checked whether the duration D _O and the duration for which only the language measure T has the value representative of the presence of the language component are longer than 2 seconds. Such a long utterance by the hearing aid wearer and / or the third party also indicates a conversation between these people.

In a decision step 40, it is subsequently checked whether the criteria mentioned above are met. For each criterion that is met, a probability value for the existence of the communication situation is increased (increased). If the probability value exceeds a predetermined threshold - d. H. If a sufficient probability for the existence of the communication situation is determined, the signal processing algorithm is changed in a subsequent method step 50 as a function of the current value of the probability. Specifically, at least one parameter of the signal processing algorithm aimed at speech intelligibility is changed, so that the speech reproduction of the hearing device 1 is improved. Alternatively, the threshold value comparison can also be omitted, the parameter or parameters being changed as a function of this over the entire value range of the probability value.

As an alternative to method step 50, ie if no probability has been assigned for the existence of the communication situation and the criteria described above have not been met, in an alternative method step 60 the signal processing algorithm is not changed as a function of the probability value for the communication situation, but only if necessary Dependence on probability values for alternative listening situations.

In an alternative exemplary embodiment (not shown in more detail), instead of the native language recognition OVD, a rotation sensor of the hearing device 1 is used, by means of which an additional dimension characteristic of an activity, specifically a head rotation of the hearing device wearer, is determined. The speech measure T and the additional measure are evaluated to determine whether the hearing aid wearer is pointing his head in the direction of the source of the speech components in the ambient noise. This is interpreted to mean that the hearing aid wearer looks at his interlocutor. The probability value is then increased to a value representative of the existence of the communication situation.

In Fig. 4 An alternative embodiment of the hearing device 1 is shown. This hearing aid 1 has an input interface 62 which is set up to receive an audio signal S _T provided by an external source and transmitted as a data signal. The audio signal S _T is relayed upon receipt of the signal processor 4 and by this - indicates a user-specific activation of a hearing situation called "Zuhörsituation" - used to generate the output signal S _A. Specifically, the microphone signal S _{M is} almost suppressed, so that only the audio signal S _{T is} contained in the output signal S _A if possible. In such a listening situation, during the reproduction of the audio signal S _T in the form of the output signal S _A in method step 50 (i.e. when the communication situation is recognized), the microphone signal S _{M is} mixed with the audio signal S _T in the output signal S _A , so that the perceptibility and intelligibility of the Microphone signal S _{M is} increased. In addition, the audio signal S _{A is} withdrawn. As a result, an active change in the signal processing algorithm of the hearing device 1 by the hearing device wearer (“switching” to a corresponding conversation program) is not necessary.

In a further embodiment (not shown in more detail), the signal processing settings of the listening situation (and thus the prioritized reproduction of the external audio signal S _A ) are switched back to when the probability value for the existence of the communication situation has been reduced in such a way that there is no communication situation, and is reduced even after a predetermined period of time, for example 10 seconds.

The object of the invention is not limited to the exemplary embodiments described above. Rather, other embodiments of the invention can be derived by the person skilled in the art from the above description. In particular, the individual features of the invention described with the aid of the various exemplary embodiments and their design variants can also be combined with one another in another way. The scope of the present invention is defined by the appended claims.

Claims (18)

1. Method for operating a hearing apparatus (1) that comprises a microphone (3) for converting ambient sounds into a microphone signal (S_M), a signal processor (4) for processing the microphone signal (S_M) into an output signal (S_A) and an output transducer (5) for outputting the output signal (S_A) to the ear of a wearer of the hearing apparatus (1), wherein the method involves

   - a characteristic voice measure (T) being ascertained for a voice component contained in the ambient sounds,

   - a supplementary measure (O) characteristic of an activity of the wearer of the hearing apparatus (1) being ascertained,

   - a correlation between the voice measure (T) and the supplementary measure (O) being evaluated,

   - this evaluation being used to increase a probability value for the presence of a communication situation between the wearer of the hearing apparatus (1) and a third party if the voice measure (T) and the supplementary measure (O), under at least one prescribed criterion, assume a value representative of the presence of the voice component and of the activity of the wearer, and

   - the probability value for the presence of the communication situation being taken as a basis for altering at least one signal processing algorithm that is executed to process the microphone signal (S_M),

   wherein the prescribed criterion used is a temporal correlation between the voice measure (T) and the supplementary measure (O).
2. Method according to Claim 1,
   wherein the supplementary measure ascertained is an own voice measure (O) characteristic of a presence of the own voice of a wearer of the hearing apparatus (1), and wherein the probability value for the presence of the communication situation is determined by virtue of a correlation between the voice measure (T) and the own voice measure (O) being evaluated.
3. Method according to Claim 2,
   wherein the probability value for the presence of the communication situation is increased if a temporal difference between the voice measure (T) and the own voice measure (O) is ascertained.
4. Method according to Claim 3,
   wherein the probability value for the presence of the communication situation is increased if within a prescribed period (Z) the voice measure (T) and the own voice measure (O) assume a value representative of the presence of the voice component or of the own voice of the wearer for a respective different length of time (D_T, D_O).
5. Method according to Claim 4,
   wherein within the prescribed period (Z) phases in which the voice measure (T) does not assume a value representative of the presence of the voice component are ignored.
6. Method according to Claim 4 or 5,
   wherein the prescribed criterion used is that within the prescribed period (Z) the voice measure (T) and/or the own voice measure (O) repeatedly assume the value representative of the presence of the voice component or of the own voice of the wearer.
7. Method according to one of Claims 1 to 6,
   wherein the hearing apparatus (1) comprises a classifier for detecting different hearing situations, and wherein the or the respective signal processing algorithm is altered with a higher priority in comparison with other hearing situations detected in parallel on the basis of the probability value for the presence of the communication situation.
8. Method according to one of Claims 4 to 7,
   wherein the criterion used is that within the prescribed period (Z) the length of time (D_T) for which the voice measure (T) assumes the value representative of the presence of the voice component approaches twice the length of time (D_O) for which the own voice measure (O) assumes the value representative of the presence of the own voice of the wearer.
9. Method according to one of Claims 1 to 8,
   wherein the criterion used is that within a past period of time of prescribed length the probability value for the presence of the communication situation has had a prescribed value.
10. Method according to one of Claims 1 to 9,
    wherein a wearer-specific inclination toward verbal communication is taken into consideration.
11. Method according to one of Claims 1 to 10,
    wherein the presence of the communication situation is precluded if only the voice measure (T) assumes the value representative of the presence of the voice component.
12. Method according to one of Claims 2 to 11,
    wherein the presence of the communication situation is precluded if the voice measure (T) and the own voice measure (O) substantially concurrently assume the value representative of the presence of the voice component and of the own voice of the wearer.
13. Method according to one of Claims 1 to 12,
    wherein a memory unit of the hearing apparatus (1) is used to store a frequency index for the alteration of the signal processing algorithm on the basis of the probability value for the presence of the communication situation.
14. Method according to one of Claims 1 to 13,
    wherein, during a reproduction of an external audio signal (S_T), the output signal (S_A) is produced by virtue of a mix ratio between the microphone signal (S_M) and the audio signal (S_T) being altered on the basis of the probability value for the presence of the communication situation.
15. Method according to Claim 14,
    wherein, when the probability value for the presence of the communication situation has been increased, the output signal (S_A) is produced by virtue of the proportion of the microphone signal (S_M) being increased in comparison with the audio signal (S_T) .
16. Method according to one of Claims 1 to 15,
    wherein when the probability value for the presence of the communication situation is decreased, in particular after additional expiry of a prescribed waiting time, the alteration of the or of the respective signal processing algorithm, in particular the altered mix ratio between the microphone signal (S_M) and the audio signal (S_T), is reset.
17. Hearing apparatus (1) that comprises a microphone (3) for converting ambient sounds into a microphone signal (S_M), a signal processor (4) for processing the microphone signal (S_M) into an output signal (S_A) and an output transducer (5) for outputting the output signal (S_A) to the ear of a wearer of the hearing apparatus (1), wherein the signal processor (4) is set up to perform the method according to one of Claims 1 to 16.
18. Hearing apparatus (1) according to Claim 17,
    having an input interface (62) that is set up for receiving an external audio signal (S_T).

Images