Language-dependent adaptation of the signal processing of hearing systems

Abstract

The invention relates to a method and an associated device (30) for adjusting the signal processing of hearing systems (2), wherein each hearing system (2) has a respective at least one hearing instrument (4) which can be parameterised by a plurality of signal processing parameters (pi). In the hearing systems (2) of a test group (42) of hearing systems (2), a respective initial setting (Pin) of the signal processing parameters (pi) is carried out, from which a final setting (Pend) of the signal processing parameters (pi) is derived by changing the parameter setting once or several times. The difference between the final setting (Pend) and the initial setting (Pin) of the signal processing parameters (pi) is used to determine a change vector (<D>) in the parameter range covered by the signal processing parameters (pi), with each change vector (D) being assigned a language specification characterising a specific language. From change vectors (D) that are assigned to the same language or a language group containing said language, a centroid of these change vectors (D) is determined in the parameter range as the change centroid (<D>). This change centroid (<D>) is used for a later adjustment of one of the hearing systems (2) of the test group (42) or another hearing system (2).

Description

[0001] Description

[0002] Language-dependent adaptation of the signal processing of hearing systems

[0003] The invention relates to a method for adjusting the signal processing of hearing systems, each of which has at least one hearing instrument. The invention further relates to an associated device for adjusting such hearing systems.

[0004] A "hearing instrument" generally refers to an electronic device that supports the hearing of a person wearing the hearing instrument (hereinafter referred to as the "wearer" or "user"). In particular, the invention relates to hearing instruments designed to fully or partially compensate for the hearing loss of a hearing-impaired user. Such a hearing instrument is also referred to as a "hearing aid." There are also hearing instruments that protect or improve the hearing of users with normal hearing, for example, enabling improved speech understanding in complex listening situations. Such hearing instruments are also referred to as Personal Sound Amplification Products (PSAPs).

[0005] Hearing instruments in general, and hearing aids in particular, are usually designed to be worn on the head and in particular in or on one ear of the user, particularly as behind-the-ear devices (also known as BTE devices) or in-the-ear devices (also known as ITE devices).

[0006] In terms of their internal structure, hearing instruments usually have at least one (acousto-electrical) input transducer, a signal processing unit (signal processor), and an output transducer. During operation of the hearing instrument, the or each input transducer receives airborne sound from the hearing instrument's environment and converts this airborne sound into an input audio signal (i.e., an electrical signal conveying information about the ambient sound). In the signal processing unit, the or each input audio signal is processed (i.e., modified with regard to its sound information) in order to support the user's hearing, in particular to compensate for any hearing loss the user may have. The signal processing unit outputs an appropriately processed audio signal to the output transducer.Additionally, in some applications, the signal processing outputs the input audio signal in its original or modified form to an external electronic device (peripheral, e.g., another hearing instrument or the user's smartphone) and / or receives another input audio signal from the peripheral. Data transmission between the hearing instrument and the peripheral is usually wireless in modern hearing instruments, e.g., using Bluetooth technology. Modern hearing instruments therefore often also include an antenna for wirelessly transmitting and receiving data.

[0007] In most cases, the output transducer is designed as an electro-acoustic transducer, which converts the (electrical) output audio signal back into airborne sound, which is then emitted into the user's ear canal after being modified from the ambient sound. Such electro-acoustic transducers are also referred to as "receivers."

[0008] A “hearing system” comprises at least one hearing instrument and, optionally, at least one further functional unit (in particular at least one further electronic device and / or a software module) with which the hearing instrument interacts during operation. In the simplest case, the hearing system is therefore made up exclusively of the one hearing instrument itself. Alternatively, in addition to the one hearing instrument, the hearing system may, for example, comprise another hearing instrument to supply the user’s other ear; in this case, it is also referred to as a “binaural hearing system”. In addition to or as an alternative to the additional hearing instrument, the hearing system may, for example, comprise a remote control, an external programming device, an external data transmission interface for facilitating wireless data exchange between different data transmission standards, or an external microphone.Additionally or alternatively, the hearing system comprises a software application (operating app) associated with the hearing instrument, which is or can be installed on a mobile device (e.g., a smartphone) or computer of the user and which provides, for example, functions for remote control and / or programming of the hearing instrument. In the latter case, the user's mobile device or computer on which the operating app is or can be installed is generally not itself a component of the hearing system. In particular, the mobile device or computer is generally manufactured and distributed independently of the hearing system and is used by the hearing system solely as a resource for storage, computing services, and, where applicable, communication services.

[0009] Modern hearing systems generally feature signal processing that can be adjusted using a variety of (signal processing) parameters. These signal processing parameters typically include frequency-selective amplification factors, which, in the case of hearing aids, can fully or partially compensate for an individual user's hearing loss. Other parameters determine, for example, the degree of directional sound pickup (beamforming), dynamic compression, spectral compression or frequency shifting, noise or feedback suppression, so-called "sound smoothing" (i.e., "acoustic anti-aliasing" that removes noise peaks from the acoustic signal), possibly binaural signal processing, etc. These parameters are typically individually adjusted for each user in order to adapt the hearing system's signal processing to the user's specific needs.This adjustment is also called “fitting” or “fit”.

[0010] The fitting of a hearing system generally begins with the measurement of an audiogram (i.e. a spectrally resolved analysis of the user's hearing ability) and a subsequent standardized initial setting ("First Fit"), which calculates the frequency-dependent gain factors based on the audiogram. A selection of standardized calculation formulas is available for this purpose (e.g. DSL I / O, NAL-NL2, etc.). For the First Fit, a fitting station, i.e. a computer system with an installed fitting program, is usually connected to the hearing instrument or each hearing instrument of the hearing system, which adjusts the hearing instrument or each hearing instrument to the calculated gain factors. Standard initial settings are often also made for other signal processing parameters (particularly those that influence compression behavior).

[0011] However, the fitting process is usually not yet complete. Rather, an experienced acoustician regularly performs fine-tuning or adjustments, taking into account various characteristics of the individual user, such as the ability to understand speech, which they relate to, for example, the hearing loss of tones (i.e., the pure tone audiogram). If the user wishes to understand speech better, the acoustician can, for example, readjust the amplification and / or compression behavior (type, onset point, strength) for individual frequency ranges. Furthermore, they have the option of activating or modifying additional signal processing algorithms that cannot be determined using standard calculation formulas (e.g., the strength and dynamic adaptation behavior of beamforming).

[0012] Such readjustments are often already prepared by the manufacturers of hearing systems and are available as “listening programs” for generally known, typical listening situations: office, television, car, outdoors, etc.

[0013] After the user has received the hearing instrument from the acoustician for the first time, they will undergo an adjustment phase of daily use to get used to its handling and the new auditory impressions. In most cases, there will be three or four readjustment appointments during this time, during which the acoustician will fine-tune the parameter settings. During the readjustment sessions, the acoustician will help not only with questions regarding the handling of the hearing instrument, but also with problems relating to speech understanding. Using the extensive range of signal processing options, they can then make appropriate settings, which can then be further fine-tuned in subsequent sessions if necessary. The object of the invention is to provide an effective, yet simplified method and an associated device for adjusting the signal processing (fitting) of hearing systems.

[0014] This object is achieved according to the invention with regard to the method by the features of claim 1. With regard to the device, the above object is achieved according to the invention by the features of claim 11. Advantageous embodiments of the invention are the subject of the dependent claims.

[0015] The invention is based on a number of hearing systems, each of which has at least one hearing instrument. Each of the hearing systems and the associated hearing instrument are designed, in particular, according to one of the above-described designs or configurations.

[0016] For example, each hearing instrument of a hearing system is assigned a parameter set P of signal processing parameters pi (where i = 1, 2, 3, ..., n). These signal processing parameters pi determine the type and function of the signal processing of the respective hearing instrument (and thus of the hearing system as a whole). The number n of signal processing parameters pi is, for example, more than one hundred (n > 100), in particular more than two hundred (n > 200).

[0017] During the parameter setting of each hearing system, a specific value is generally assigned to each signal processing parameter pi of the or each associated hearing instrument. The parameter set P thus determined contains the respective assigned value for each signal processing parameter pi, e.g., an integer value or a floating-point value within a specified range, a binary value, etc. Each parameter set P therefore mathematically represents a vector (P = {pi} = {pi, p2, ps, ... , p n}) in an n-dimensional parameter space spanned by the signal processing parameters pi. The invention is based on the consideration that it is advantageous, when adjusting a hearing system, in particular during the initial setting (first fit), to take into account, in addition to other criteria such as in particular the audiogram of the user (or comparable variables characterizing the hearing ability of the user), a specific language (in particular the native language of the user or the language spoken at the user's location) and thus to vary the parameter adjustment depending on this language.

[0018] The aforementioned initial setting Pin describes the initial / first definition of the parameter set P, which is usually performed by the hearing care professional when handing over the hearing system to the user. The initial setting Pin is thus a specific characteristic of the parameter set P, and thus also a vector in the parameter space of the signal processing parameters pi (P -> Pin).

[0019] An initial setting that already takes a specific language into account (namely, the language the user speaks or hears) has been recognized as having the potential to achieve particularly good speech intelligibility in the signal emitted by the hearing system right from the start of use (and not only after several subsequent adjustments). This is because different languages ​​differ considerably in their pronunciation with regard to their respective sound and frequency distribution. Therefore, different languages ​​can be perceived better or worse with a given hearing loss – regardless of the user's existing language skills.

[0020] For example, there are tonal languages ​​such as Chinese, where a change in the pitch or pitch progression in a syllable is usually accompanied by a change in the meaning of the corresponding word (or morpheme), and non-tonal languages ​​such as German, where the pitch or pitch progression in a syllable has no influence on the meaning of the word. Even non-tonal languages ​​have very different characteristics. For example, there are languages ​​rich in vowels and languages ​​rich in consonants. Even a single language such as Portuguese can be articulated phonetically in one country, Brazil, and also with the many vowels of the language, while in another country, Portugal, the pronunciation is "buried" and vowels are often "swallowed", so that this language appears (especially to outsiders) to be a language rich in consonants. Slavic languages ​​such asPolish is often richer in sibilants than, for example, German. Arabic, on the other hand, contains many voiceless fricatives (i.e., fricative sounds caused by the narrowing of the articulation point and the air flowing through this narrowing), which are relevant for speech comprehension. For these reasons, the spectral center of gravity (i.e., the mean of the sound frequencies occurring in a language) can sometimes vary considerably between different languages. For example, the spectral center of gravity in Italian is, on average, about a third higher than in German. Likewise, the spectral center of gravity is higher in Eastern European languages ​​than in Western European languages.

[0021] It is known that it is beneficial for speech understanding when the signal processing settings of a hearing system take into account the language spoken or heard by the user, as this allows the frequency ranges or frequency distributions important for speech understanding in the respective language to be better considered. Furthermore, differences in compression behavior can also influence speech understanding. In particular, signal processing settings that better emphasize or less distort certain phonemes of a language can significantly improve speech understanding for the hearing system user.

[0022] However, it has proven difficult to date to precisely calculate or predict the effects of modified signal processing on speech intelligibility for specific languages. The invention therefore provides a method and an associated device which enable automatic and model-free (i.e. not based on basic assumptions or scientific models) consideration of language dependency when adapting the signal processing of a hearing system. According to the method according to the invention, an initial adjustment Pin (also known as initial adjustment or “first fit”) of the signal processing parameters is carried out for each hearing system in a test group. The initial adjustment of the hearing systems in the test group is carried out in particular in a conventional manner, for example using a known, standardized method such as DSL I / O, NAL-NL2 or a modification thereof.The initial settings made for the hearing systems in the test group are, in particular, language-independent, meaning they are not specifically designed for a particular language or at least are not varied for different languages. The test group preferably consists of identical hearing systems or at least of hearing systems that have identical or comparable hearing instruments, in particular hearing instruments with the same set of signal processing parameters. The test group preferably includes a statistically meaningful set of hearing systems, e.g., several hundred, thousands, or tens of thousands of hearing systems.

[0023] In a period following the initial adjustment, a final setting Pend of the signal processing parameters is derived for each of the hearing systems in the test group by making one or more changes (fine or subsequent adjustments) to the initial setting - provided this is also still done in the conventional way. The aforementioned final setting describes the terminal / last determination of the parameter set P during the fitting process, which is generally made by the acoustician at the end of the hearing system's familiarization phase and which the user then maintains permanently or at least over an extended period during further use of the hearing system. The final setting Pend is therefore also a specific characteristic of the parameter set P, and thus also a vector in the parameter space of the signal processing parameters Pi (P -> Pend).

[0024] From the difference between the final setting and the initial setting of the signal processing parameters, a change vector D is then determined for each hearing system in the test group in the parameter space spanned by the signal processing parameters. Mathematically speaking, this change vector D is also a vector in the parameter space of the signal processing parameters pi. The final setting Pend is calculated from the vector sum of the initial setting Pin and the change vector D (Pend = Pin + D).

[0025] Each change vector D is assigned a language identifier that characterizes a specific language. The language identifier can specify a language directly (e.g., "German," "English," "Mandarin," etc.). Alternatively, it can also characterize the language by specifying a country, a part of a country, or a region (e.g., "Germany," "Catalonia," "Bavaria," etc.). Different languages ​​can also be characterized by numbers or other identification codes. Different dialects ("Upper Bavarian") or variants of a language (e.g., "Spanish Portuguese" and "Brazilian Portuguese") can, within the scope of the invention—depending on their implementation—be interpreted either as different "languages" or as one and the same language. The language identifier is provided, for example, by the user or an acoustician, for example, via an operating app for the hearing system or an acoustician's fitting station.Alternatively, within the scope of the invention, the language information can also be assigned automatically, for example based on the geographical location where the initial setting of the hearing system is carried out.

[0026] Preferably, the language information is assigned in such a way that it is characteristic of the language that the user predominantly hears or speaks.

[0027] In a first variant of the method according to the invention, a center of gravity of these change vectors in the parameter space is determined from change vectors D that are assigned to the same language according to the associated language information. This center of gravity is also referred to below as the "change center of gravity <d>The center of gravity of the change vectors, referred to as ", is therefore regularly determined separately for each language. In a further variant of the method according to the invention, however, change vectors for several similar languages ​​are taken into account for the calculation of the center of gravity of change, which are assigned to a common language group for the purpose of this analysis. For example, change vectors from hearing systems of Italian and Spanish users can be evaluated together to determine a common center of gravity of change.

[0028] As a (change) focus <d>is a multidimensional mean of the considered change vectors D, e.g. according to where the counting index j (with j = 1, 2, 3, ... , m) runs over the change vectors Dj considered for the calculation of the change center and m denotes the number of these change vectors Dj considered. The change center <d>is mathematically a vector in the parameter space of the signal processing parameters pi.

[0029] The change center of gravity determined for a specific language is finally used for a later adjustment of at least one of the hearing systems of the test group or at least one other hearing system, in particular by adjusting a previous parameter set P by the change center of gravity <d>is changed (Pait + <d>Pneu). The respective hearing system is thus adapted to the specific language.

[0030] In a preferred application, the change focus is taken into account during the initial setting of a new hearing system by changing a non-language-specific initial setting of the hearing system (e.g. according to DSL I / O, NAL-NL2 or a modification thereof) by the change focus that was previously determined for the language that is selected (e.g. by the user or an acoustician) for the hearing system in question.

[0031] In another application, the change focus is taken into account within the framework of a language-specific hearing program, which can be activated and deactivated automatically during operation of the hearing system or reversibly by the user. For example, a "speech" hearing program—which is also frequently found in conventional hearing systems—is adapted to a specific language, e.g., English, within the scope of the invention by applying the relevant change vector. The user can activate this adapted "speech" hearing program, particularly in alternation with other hearing programs such as "music," "office," etc.

[0032] Optionally, the hearing system can contain several hearing programs, each assigned to a different language, which can be activated automatically during operation of the hearing system or alternately (against each other or against other hearing programs) by the user. For example, the user can choose between the hearing programs "Speech (German)", "Speech (English), "Music", "Office", etc. In another embodiment, the hearing system automatically activates a hearing program based on automatic speech recognition or automatic (geographical) location detection that is specifically adapted to the language spoken in the user's environment. Optionally, such language-specific hearing programs can also be dynamically generated during operation of the hearing system based on a language-independent basic setting and the change focus specific to the respective language.

[0033] In an advantageous embodiment of the invention, only those change vectors based on identical or similar initial settings of the signal processing parameters are considered when calculating the change center of gravity. In this embodiment of the invention, the change center of gravity is thus determined not only specifically for a particular language, but also specifically for a specific type of initial setting (and thus, in the case of hearing aids, specifically for a specific type of hearing loss).

[0034] In a further advantageous embodiment of the invention, each change vector—in addition to the language information—is assigned at least one piece of information about a user characteristic. For the calculation of the change center, preferably only those change vectors are considered that match the assigned user characteristic. The change center is thus additionally determined specifically for users with at least one common user characteristic.

[0035] The at least one user property is in particular

[0036] - an indication of the user's age,

[0037] - an indication of the gender of the user, and / or

[0038] - information about the user's musical background / experience.

[0039] The reason for specifying the user's musical background is that musicians or music lovers usually have a trained ear and therefore—even with the same physical hearing ability—perceive tones and sounds differently than users without musical background / experience. Musicians or music lovers therefore often require special parameter settings for a satisfactory listening experience.

[0040] When applying the invention to hearing systems for hearing-impaired users, a user characteristic that is preferably considered (if necessary alternatively or in addition to at least one other user characteristic) is an indication of the period of unaided hearing loss of the user before using a hearing system (unaided period). The indication of the unaided period is based on the fact that unaided hearing loss (i.e., not fully or partially compensated by a hearing system) leads to a gradual degeneration of speech comprehension over time because the brain unlearns how to understand speech. This loss of speech comprehension is often not reversible, or only to a limited extent, even after the start of treatment (i.e., the use of a hearing system).Therefore, users with a longer untreated period of hearing loss generally require different parameter settings for satisfactory speech understanding than users with a shorter untreated period, even with the same physical hearing loss. Preferably, the or each focal point of change is only used in a subsequent adjustment of one of the hearing systems in the test group or another hearing system if it meets at least one specific quality criterion.

[0041] According to a first quality criterion, the magnitude of the change in the signal processing parameters mediated by the change center is checked, in particular whether the change center does not fall below a predetermined minimum threshold. In a preferred embodiment of the invention, the or each change center is only used in this case during a subsequent adjustment of one of the hearing systems in the test group or of another hearing system. In this way, parameter changes are avoided that would be irrelevant or insignificant due to their insignificance. In different variants of the invention, the at least one minimum threshold can be defined with respect to just one signal processing parameter, with respect to a selected group of signal processing parameters, or with respect to all signal processing parameters.For example, a weighted amount of the change center of gravity is determined and the change center of gravity is only used in a later adjustment of one of the hearing systems in the test group or another hearing system if the weighted amount exceeds a specified minimum value.

[0042] The weighted amount | <D)| w the focus of change <d>For example, according to where di are the components of the vectorial center of change (D = {di, d2, ds, ..., dn}) and Xi are empirically determined weighting or scaling factors. According to a second quality criterion, a scatter (in particular a statistical variance) of the change vectors considered for determining the center of change is calculated. The center of change is only used for a later adjustment of one of the hearing systems in the test group or of another hearing system if the scatter does not exceed a predetermined maximum value. In this way, parameter changes are avoided that would be irrelevant or insignificant due to excessive scatter of the individual change vectors.

[0043] Preferably, the respective initial setting Pin, the respective final setting Pend, and the voice input for all hearing systems in the test group are stored in a central database, in particular in a cloud storage (data cloud). The change vector D assigned to each hearing system in the test group is preferably calculated in a cloud server. In an alternative embodiment of the invention, the change vectors are determined within the respective hearing system or in an acoustician's fitting station and fed into a central database.

[0044] The final setting (Pend) of the signal processing parameters assigned to each hearing system in the test group is preferably determined automatically based on the time course of the successive parameter adjustment. In particular, the current setting of the signal processing parameters is adopted as the final setting for each hearing system in the test group if this current setting is not changed for a specified period of time (e.g., two months).

[0045] To prevent the analysis of the change vectors from being corrupted by unsuccessful parameter adjustments, the adoption of the final setting, the calculation of the change vector, or the consideration of the change vector in the calculation of the change center point for each data set assigned to a hearing system are optionally made dependent on a quality control. The final setting is only adopted, or the change vector is only calculated, or only considered in the calculation of the change center point, if at least one quality criterion is met, in particular:

[0046] • if at least a specified minimum number of parameter adjustments (fine adjustments) have been made between the initial setting and the final setting, and / or

[0047] • if a speech audiometric test was carried out after the initial adjustment and after the final adjustment and these tests show a significant improvement in speech understanding (in particular exceeding a specified threshold).

[0048] The speech audiometric test is preferably based on an established standard test for assessing speech comprehension, e.g., the so-called "Oldenburg Sentence Test" for German. This test determines, for example, the threshold sound level at which the user is just able to understand a specified proportion (e.g., 50%) of spoken words or sentences. In this case, the improvement in speech comprehension is determined by the reduction in the threshold sound level between the test after initial adjustment and the test after final adjustment. Speech audiometric tests can be performed fully automatically or under the guidance of an acoustician. In the latter case, the values ​​for the threshold sound level can also be entered manually by the acoustician and thus assigned to the data set of the respective hearing system.

[0049] Alternatively or in addition to conducting a scientifically standardized test (e.g., the aforementioned Oldenburg Sentence Test), other options for assessing speech comprehension under everyday conditions are also possible. For example, questionnaires can be used within the scope of the invention that provide the HG wearer with questions and selectable answers (e.g., indicating the degree of speech comprehension in radio news, particularly differentiated by male and female voices; indicating the degree of effort required to understand radio news; indicating the degree of speech comprehension and / or speech effort during conversations in quiet rooms with few speakers; etc.).

[0050] The results of such questionnaires may exhibit greater variance than the results of a scientific test, since environmental conditions in everyday life vary more than defined laboratory conditions, which are typically used for scientific tests. Nevertheless, questionnaires can also produce results with acceptable variance / dispersion, which can be advantageously used within the scope of the invention.

[0051] Furthermore, speech comprehension tests can also be carried out automatically within the scope of the invention using a peripheral device of the hearing aid or an associated control app. For example, audio files can be made available on the cloud-based server, which the user downloads via the control app and plays back via the hearing instrument. The control app can, for example, adjust the volume to a level suitable for the user, or even vary the level and thus output speech at a defined volume. The implementation of the speech comprehension test controlled via the control app can, for example, be modeled on the Oldenburg Sentence Test. For example, the user is first instructed to find a quiet environment to administer the test. After the test has started, the control app plays audio files at a volume that is easily audible to the HG wearer.The HG wearer repeats the sentences heard. The repeated sentences are recorded by the hearing instrument or a smartphone microphone and analyzed either offline in the app or online on the cloud server using speech recognition software. Based on this analysis, the proportion of correctly reproduced words is determined. If the proportion of correctly reproduced words is high (in particular, exceeds a predefined threshold), the sound level is slightly reduced for the next run. If, on the other hand, the proportion of correctly reproduced words is low (in particular, falls below another threshold), the sound level is slightly increased for the next run. The set sound level converges on a value to be determined as the test result.

[0052] One advantage of app-controlled language comprehension testing is that it allows for large numbers of language comprehension tests to be administered with relatively little effort. This also makes it easier to capture temporal changes in language comprehension (e.g., due to increasing aging, progressive illness, or medication use) by repeating the language comprehension tests multiple times for the same user.

[0053] The device according to the invention for adjusting the signal processing of hearing systems comprises, in a narrower sense, at least one (first) adjustment unit which is configured to carry out the above-described initial adjustment of the signal processing parameters for the hearing systems of the above-described test group of hearing systems. The adjustment unit is further configured to carry out the above-described final adjustment of the signal processing parameters for each hearing system in the test group, wherein the final setting results from the initial setting by changing the signal processing parameters of the respective hearing system one or more times. Within the scope of the invention, the adjustment unit can be implemented as part of each hearing system, e.g., as a function of an operating app, via which the user can carry out the initial adjustment of the hearing system and at least one subsequent adjustment of this initial setting themselves.In this case, the fitting unit is thus formed from a plurality of sub-units distributed among the individual hearing systems. Preferably, however, the device comprises one or more fitting stations, each operated by an acoustician, i.e., computers with fitting software installed on them, to which the hearing instruments of the hearing systems can be connected to adjust the signal processing parameters. These fitting stations are located, for example, at acousticians' offices, whom the users of the hearing systems in the test group visit to adjust the signal processing parameters. In a further preferred embodiment, the fitting unit is implemented as a central unit, e.g., in the form of a cloud server, which is connected to all hearing systems in the test group for data transmission purposes, e.g., via the Internet.

[0054] The device also comprises an evaluation unit for determining the change vectors described above and for determining the language-specific change focus described above. The evaluation unit preferably comprises a central server, particularly one implemented in a data cloud, with evaluation software installed thereon. Parts of the evaluation unit, e.g., software modules for determining the change vectors, can, in certain embodiments of the invention, also be implemented in the hearing systems (e.g., as a function of an operating app) or in the fitting stations.

[0055] In a broader sense, the hearing systems of the test group also form part of the facility.

[0056] Finally, the or each first fitting unit or a further (second) fitting unit of the facility is designed to use the focus of change in a later adjustment of one of the hearing systems of the test group or of another hearing system.

[0057] In general, the device according to the invention is configured to automatically carry out the method according to the invention. Embodiments of the method therefore correspond to corresponding embodiments of the device, so that the above explanations regarding embodiments of the method can also be applied to the device, and vice versa.

[0058] In summary, the invention is based on the idea

[0059] • to demonstrate a procedure that can be used to check whether hearing aid fine adjustments made in practice show language-specific deviations from a standard initial setting, and

[0060] • to use such detected language-specific deviations, if necessary after a successful test for quality or significance, for automated adjustments of the signal processing parameters, in particular initial settings (first fits).

[0061] In one embodiment for implementing the method according to the invention, a central database is provided, e.g., as a cloud-based solution, in which a large number of initial settings for hearing systems are stored, each with an associated voice prompt. Alternatively and equivalently, the central database stores hearing loss measurement data (in particular, audiogram data) of the respective user for each hearing system, along with information on the standard procedure for creating the initial setting, from which the initial setting values ​​are reconstructed.

[0062] Subsequently, parameter values ​​that were changed during refitting are also saved in the central database. If the data set assigned to a hearing system remains unchanged for a specified period of time (e.g., two months) after several refittings, this is considered an indicator that the refitting is complete; i.e., the last updated parameter settings stored in the central database are considered the final settings of the hearing system.

[0063] Once the initial setting, final setting and speech information for a larger test group of hearing systems are available in the central database, the data is analyzed by a cloud server that acts as an evaluation unit.

[0064] For each data set, the associated change vector is calculated from the difference between the final setting and the initial setting. Visually speaking, the change vectors assigned to the same language represent a point cloud (i.e., a "cluster" in the n-dimensional parameter space of the signal processing parameters). The change center, i.e., the cluster center, is calculated from the change vectors.

[0065] The points of the cluster, i.e. the change vectors considered, are further examined for homogeneity, ie the 'cluster quality' is calculated. In the example shown here, the calculation of the cluster quality includes

[0066] • calculating a weighted amount of the center of change and comparing this amount with a predetermined minimum value to ensure that the center of change is not too small, ie represents a perceptible effect at all, and

[0067] • calculating the dispersion (namely the statistical variance) of the change vectors considered around the center of change and comparing this dispersion with a predetermined maximum value to ensure that the dispersion is not too large compared to the magnitude of the center of change; if the dispersion is too large, the center of change is not meaningful for further use.

[0068] If at least one of the cluster quality criteria is met—that is, if the weighted magnitude of the change center exceeds the minimum value and / or the variance falls below the maximum value—the previously determined change center will be used for future initial adjustments of new hearing aids whose users share the same language. Alternatively, the previously determined change center will only be used for future initial adjustments of new hearing aids (with the same assigned language or language group) if both cluster quality criteria are met.

[0069] In concrete terms, taking the change center into account means that after the standardized initial setting (such as DSL I / O, NAL-NL2, or a manufacturer-specific standard initial setting), the resulting signal processing parameter values ​​are adjusted by the change center values. This shortens the path the user typically takes to complete the fine-tuning process.

[0070] In the following, exemplary embodiments of the invention are explained in more detail with reference to a drawing. In the drawings:

[0071] Fig. 1 shows a schematic representation of a hearing system with a hearing instrument that can be parameterized by a plurality of signal processing parameters, and with an associated software application for remote control and configuration of the hearing instrument (operating app) that is connected to the hearing instrument by data transmission technology and is installed on a smartphone of the user of the hearing system, and

[0072] Fig. 2 shows a schematic representation of a device for adjusting the signal processing of hearing systems.

[0073] Corresponding parts and sizes are always provided with the same reference symbols in all figures.

[0074] Fig. 1 shows a roughly schematic representation of a hearing system 2, which is formed from a hearing instrument 4 and an associated software application (operating app 6).

[0075] The hearing instrument 2 is, for example, a hearing aid, i.e., a hearing instrument designed to support the hearing of a hearing-impaired user. In the embodiment shown here, the hearing instrument 2 is designed as a BTE device. It accordingly comprises a housing 8, which is intended to be worn behind a user's ear, and an earpiece (not shown in detail) that is intended to be inserted into the user's ear canal. The hearing instrument 2 further comprises a sound tube (also not explicitly shown) that mechanically connects the housing 8 to the earpiece. Within a housing 8, the hearing instrument 4 has the following components:

[0076] • at least one microphone 10 (in the example shown two microphones 10) as an acousto-electrical input transducer,

[0077] • a loudspeaker, also known as earpiece 12, as an electro-acoustic output transducer

[0078] • a (particularly digital) signal processor 14, and

[0079] • a battery 16.

[0080] The hearing instrument 4 further comprises a wireless communication device (not shown in detail) for exchanging data with the operating app 6, e.g. in the form of a Bluetooth transceiver.

[0081] During normal operation of the hearing instrument 4, the microphones 10 each record airborne sound from the environment of the hearing instrument 4. The microphones 10 convert the sound into an (input) audio signal I, i.e., into an electrical signal containing information about the recorded sound. The respective input audio signal I is fed within the hearing instrument 4 to the signal processor 14, which modifies this input audio signal I to support the user's hearing ability, in particular by frequency-selectively amplifying it to compensate for a hearing loss of the user. The function of the signal processor 14 can be variably configured using a plurality of signal processing parameters pi.

[0082] The signal processor 14 outputs the modified audio signal (output audio signal 0) to the receiver 12. The output audio signal O is an electrical signal containing information about the processed and thus modified sound. The receiver 12 converts the output audio signal O into airborne sound, which is conducted via a sound channel 18 to a tip 20 of the housing 8. From there, the airborne sound emitted by the receiver 12 is conveyed to the earpiece via the sound tube and emitted into the user's ear canal. The signal processor 14 and all other electrical or electronic components of the hearing instrument 4 are supplied with a direct electrical voltage, referred to as the operating voltage U, from the battery 16.

[0083] The operating app 6 is used in particular for remote control and programming of the hearing instrument 4. During operation of the hearing system 2, the operating app 6 is installed and running on a (particularly mobile) computer. In the example shown, this computer is a smartphone 22 of the user. The computer, in particular the smartphone 22, is not itself a component of the hearing system 2, but is used by the operating app 6 only as an external resource for computing power, storage space, and communication services. In particular, the operating app 6 accesses a wireless communication device (not shown in detail) (in particular a Bluetooth transceiver) of the smartphone 22 in order to exchange data with the hearing instrument 4 via a wireless data communication connection 24. The operating app 6 also serves - using the smartphone 22 - as an intermediary for data exchange between the hearing instrument 4 and the mobile network and the Internet, e.g.to use the hearing instrument 4 to make telephone calls and to load updates to the firmware of the hearing instrument 4 or configuration data (in particular values ​​for the signal processing parameters pi) and to upload them to the hearing instrument 4.

[0084] Fig. 2 shows a device 30 for adjusting the signal processing of hearing systems 2, which are, for example, hearing systems 2 of the type shown in Fig. 1.

[0085] The device 30 comprises a first adaptation unit 32, a (settings) database 34, an evaluation unit 36, a (speech adaptation) database 38, and a second adaptation unit 40. In a preferred implementation, the adaptation units 32 and 40 as well as the evaluation unit 36 ​​are implemented as cloud servers, and the databases 34 and 38 are implemented as cloud storage. The adaptation units 32 and 40 are connected to the hearing systems 2 to be adapted via the internet (in particular via a mobile radio connection) for data transmission purposes.

[0086] The first fitting unit 26 functions as a conventional fitting station, which performs individual adjustment of the hearing systems 2 of a test group 42 to the needs of the respective user, independent of the user's (spoken or typically heard) language. The test group 42 preferably comprises a statistically relevant number of hearing systems 2, e.g., several hundred, thousand, or ten thousand hearing systems 2.

[0087] During the fitting process performed by the fitting unit 32, the fitting unit 32 creates a first parameter set of the signal processing parameters pi for each of the hearing systems 2 of the test group 42 based on audiogram data that characterizes the hearing ability of the user of the respective hearing system 2, which is referred to below as the initial setting Pin. For this purpose, the fitting unit 32 uses, in particular, a conventional fitting model, e.g., DSL I / O or NAL-NL2. Following creation, the fitting unit 32 transmits the initial setting Pin to the respective hearing system 2, which configures its signal processing accordingly.

[0088] The user thus begins using the respective hearing system 2 of test group 42 with the initial setting Pin specified by the adjustment unit 32. During an acclimatization phase following the start of use, the initial setting Pin is changed by the adjustment unit 32 in one or more fine-tuning steps based on user feedback regarding problems with the hearing impression and / or based on hearing comprehension tests. The changes can either be specified manually by an acoustician operating the adjustment unit 32 or determined automatically by the adjustment unit 32. In both cases, the adjustment unit 32 creates a parameter set (intermediate setting) containing the changed signal processing parameters pi in each fine-tuning step and transmits this to the respective hearing system 2, which in turn configures its signal processing accordingly.The adjustment unit 32 stores the number of fine adjustment steps for each hearing system 2 of the test group 42.

[0089] The adjustment unit 32 ends the familiarization phase when a predetermined condition occurs, e.g. when a predetermined command is entered by the user or acoustician, when a predetermined period of time has elapsed after use has begun, or after a predetermined number of fine-adjustment steps. Furthermore, it can be provided that the adjustment unit 32 automatically ends the familiarization phase if no further fine-adjustment is carried out for a predetermined period of time (e.g., 2 months). In all cases, at the end of the familiarization phase, the adjustment unit 32 stores the last created intermediate setting for each hearing system 2 of the test group 42 as the final setting Pend in the settings database 34, together with the associated initial setting Pin, the number of fine-adjustment steps, and a language entry. The language entry preferably characterizes the language that the user of the respective hearing system 2 predominantly speaks or hears.Preferably, a hearing comprehension test (e.g., an "Oldenburg sentence test") is conducted for each hearing system 2 by means of the adjustment unit 32 upon commencement of use and at the end of the familiarization phase. In this case, the adjustment unit 32 determines the improvement in the user's hearing comprehension during the familiarization phase based on the comparison of the test results and also stores this information in the settings database 34.

[0090] Optionally, the fitting unit 32 additionally records one or more of the following information for each hearing system 2 of the test group 42:

[0091] • Age of the user

[0092] • Gender of the user

[0093] • Type and / or severity of hearing loss

[0094] • Untreated period (i.e. period of untreated hearing loss before starting to use the hearing system 2),

[0095] • musical background (and thus implicitly ear training). Based on this data stored in the settings database 34, the evaluation unit 36 ​​determines language-specific change priorities in a language-specific evaluation. <d>the signal processing parameter pi, i.e. average changes in the signal processing parameter pi for individual languages ​​or for language groups of similar languages. This language-specific evaluation, described in more detail below, is preferably carried out for the first time by the evaluation unit 36 ​​when a predetermined minimum number of (e.g. 100) final settings Pend are stored in the settings database 34. The language-specific evaluation is preferably repeated according to predetermined criteria (e.g. at regular time intervals or after a predetermined number of further final settings Pend have been stored in the settings database 34). The language-specific evaluation is preferably always carried out for all languages ​​or language groups at each repetition. Alternatively, this evaluation is carried out and repeated separately for different languages ​​or different language groups.

[0096] The evaluation unit 36 ​​determines a change vector D for all stored final settings Pend (and thus for each hearing system 2 of the test group 42) - or at least for each stored final setting Pend of the corresponding language or language group - which represents the difference between the corresponding values ​​of the final setting Pend and the initial setting Pin for each signal processing parameter pi. The change vector D, which mathematically results from the vector subtraction of the final setting Pend and the initial setting Pin (D = P end - Pin ), thus represents the change that the signal processing parameters pi of the respective hearing system 2 experienced during the familiarization phase. The evaluation module 36 preferably stores the determined change vectors D in the settings database 34 so that these change vectors D do not have to be determined again when the language-specific evaluation is carried out repeatedly.

[0097] In a subsequent step of the language-specific evaluation, the evaluation unit 36 ​​determines the change focal points by averaging the change vectors D determined for a specific language or language group. <d>for the respective language or language group. In certain embodiments, it may be provided that the evaluation unit 36 ​​for determining the change focus <d>all change vectors D stored for the respective language or language group. In order to ensure the significance of the change focal points <d>To improve the analysis, the evaluation unit 36 ​​checks the change vectors D in question in one of the determination of the change focal points <d>upstream quality control on at least the following quality criteria:

[0098] • The number of fine adjustments assigned to the respective change vector D exceeds a specified limit (e.g. one fine adjustment).

[0099] • The improvement in the user's hearing comprehension during the familiarization phase, assigned to the respective change vector D and possibly determined by comparing the test results of hearing comprehension tests, exceeds a specified limit value.

[0100] In this case, the evaluation unit 36 ​​takes into account when determining the focus of change <d>only those change vectors D that meet at least one of the tested quality criteria.

[0101] Also about the significance of the change priorities <d>To improve the performance of the language, the evaluation unit 36 ​​filters the change vectors D that are relevant for a specific language or language group according to at least one further filter criterion, e.g.

[0102] • according to the position of the final setting Pend or initial setting Pin in the parameter space spanned by the signal processing parameters pi,

[0103] • the age of the user,

[0104] • the gender of the user, the type and / or severity of the hearing loss, the length of time the user has not been treated, and / or the user’s musical background.

[0105] In this case, the evaluation unit 36 ​​takes into account when determining the focus of change <d>only those change vectors D that correspond to at least one filter criterion, e.g.

[0106] • with regard to the position of the final setting Pend or initial setting Pin in a specific region of the parameter space,

[0107] • are associated with a specific age range, gender, level of musical proficiency and / or type and / or severity of hearing loss, and / or

[0108] • be within a specified range with regard to the user's unused period.

[0109] In this case, the evaluation unit 36 ​​preferably determines different change priorities for one and the same language or language group <d>which are assigned to different filter criteria, in particular different regions of the parameter space, age groups, etc.

[0110] Also about the significance of the change priorities <d>To improve the quality of life, the evaluation unit 36 ​​examines the identified key points of change <d>- in addition to or alternatively to the upstream quality control described above - in a downstream quality control,

[0111] • whether the size of the respective change focus <d>does not fall below at least a predetermined minimum threshold, in particular by exceeding the weighted amount described above | <d>| W of the respective change focus <d>calculated and compared with a given threshold, and / or whether the dispersion (namely the statistical variance) of the values ​​used to determine the respective change center <d>considered change vectors D does not exceed a specified threshold value.

[0112] All identified key change areas <d>that do not meet the quality criterion being tested (or at least one of both quality criteria being tested) are discarded by the evaluation unit 36 ​​(and thus are not further stored or used). The remaining key changes <d>which meet the or each tested quality criterion in a subsequent quality check are stored by the evaluation unit 36 ​​in the language adaptation database 38.

[0113] The second adjustment unit 40 is fundamentally similar to the first adjustment unit 32 and, like the latter, serves to individually adapt the signal processing of new hearing systems 2 to the needs of the respective user. In doing so, it first creates—basically like the adjustment unit 32—an initial adjustment Pin of the signal processing parameters pi of the respective hearing system 2, which it then modifies in one or more fine-tuning steps to a final adjustment Pend.

[0114] In contrast to the adaptation unit 32, the adaptation unit 40 takes into account the language spoken or heard by the user by determining a suitable change focus based on a language specification and optionally further information about the user (such as age, gender, etc.) from the language adaptation database 38. <d>selects, if one is available. The adaptation unit 40 creates the initial setting Pin by determining a language-independent basic setting for the signal processing parameters pi in a manner analogous to the adaptation unit 32, in particular using the same conventional adaptation model, and then adjusting this basic setting by the selected change focus <d>changed. The initial setting Pin transmitted by the adaptation unit 40 to the associated hearing instruments 2 is thus already specifically adapted to the language spoken or heard by the user, thereby achieving improved hearing comprehension for the user right from the start of use. The adaptation units 32 and 40 as well as the evaluation unit 36 ​​are each configured as separate software and, if applicable,

[0115] Hardware unit. In further implementations of the invention, the adaptation units 32 and 40 can also be combined (together with or without the evaluation unit 36) in a software and, if applicable, hardware unit. In particular, the adaptation units 32 and 40 can be implemented as two different operating modes of the same software and, if applicable, hardware unit. Furthermore, the databases 34 and 38 can optionally be implemented as separate units—as shown in Fig. 2—or combined into one unit.

[0116] Alternatively, the fitting units 32 and / or 40 can also be divided into several subunits as distributed systems. For example, such subunits of the fitting units 32 and / or 40 can be implemented in fitting stations that are operated remotely by different acousticians. Alternatively or additionally, subunits of the fitting units 32 and / or 40 can also be implemented as components of the respective operating app 6 of the hearing systems 2. In this case, the initial fitting Pin and the final fitting Pend are thus determined by the respective hearing system 2 itself. In addition, the database 34 or the database 38 can also be further divided into subunits.

[0117] In a further embodiment, the adjustment unit 40 determines the language-adapted final setting Pend in the form of a hearing program that can be reversibly activated (and deactivated accordingly) during operation of the hearing system 2 by user action or by an automatic classification of hearing situations. The respective language-specific change focus <d>influences the signal processing of hearing system 2 only during the activation of the respective hearing program. In this case, hearing system 2 can optionally also contain several hearing programs for different languages, between which it is possible to switch while hearing system 2 is in operation. The adjustment of the signal processing of hearing system 2 to a specific language (or to several languages ​​alternating) using the respective adjustment focus <d>This can also be done in the hearing system 2 itself while it is in operation.

[0118] The invention is particularly clear from the exemplary embodiments described above, but is by no means limited thereto. Rather, further embodiments of the invention can be derived from the claims and the above description.

[0119] List of reference symbols

[0120] 2 hearing system

[0121] 4 Hearing instrument

[0122] 6 Operating app

[0123] 8 housings

[0124] 10 Microphone

[0125] 12 listeners

[0126] 14 Signal processor

[0127] 16 Battery

[0128] 18 sound channels

[0129] 20 lace

[0130] 22 smartphones

[0131] 24 Data communication connection

[0132] 30 furnishings

[0133] 32 (first) adjustment unit

[0134] 34 Settings database

[0135] 36 Evaluation unit

[0136] 38 Language adaptation database

[0137] 40 (second) adjustment unit

[0138] 42 test group

[0139] <d>Focus of change

[0140] Pi signal processing parameters

[0141] D change vector

[0142] I (Input) audio signal

[0143] 0 (output) audio signal

[0144] Pend end setting

[0145] Pin initial setting

[0146] II Operating voltage< / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d> < / d>

Claims

Claims 1. Method for adjusting the signal processing of hearing systems (2), wherein each hearing system (2) has at least one hearing instrument (4) which can be parameterized by a plurality of signal processing parameters (pi), - wherein an initial setting (Pin) of the signal processing parameters (pi) is carried out for each hearing system (2) of a test group (42) of hearing systems (2), - wherein for each of the hearing systems (2) of the test group (42) a final setting (Pend) of the signal processing parameters (pi) is derived by changing the parameter setting once or several times, - whereby a change vector (D) in the parameter space spanned by the signal processing parameters (pi) is determined from the difference between the final setting (Pend) and the initial setting (Pin) of the signal processing parameters (pi), - each change vector (D) is assigned a language specification characterising a particular language, - wherein from change vectors (D) which are assigned to the same language or to a language group containing this language according to the associated language specification, a center of gravity of these change vectors (D) is determined in the parameter space as the center of gravity of change «D», and - whereby this change focus «D» is used for a later adjustment of one of the hearing systems (2) of the test group (42) or of another hearing system (2).

2. Method according to claim 1, wherein the change center of gravity «D» is taken into account during the initial setting (Pin) of a new hearing system (2).

3. Method according to claim 1 or 2, whereby the change focus «D» is taken into account within the framework of a language-specific hearing program which can be activated automatically during operation of the hearing system (2) or reversibly upon user request.

4. Method according to one of claims 1 to 3, wherein a plurality of hearing programs, each assigned to different languages, are stored in the hearing system (2), which programs can be activated alternately during operation of the hearing system (2) upon user request or automatically, and within the framework of which a change focus «D» specific for the respective language is taken into account.

5. Method according to one of claims 1 to 4, wherein for the calculation of the change center of gravity «D» only those change vectors (D) are taken into account which are based on the same or similar initial settings (Pin) or final settings (Pend) of the signal processing parameters (pi).

6. Method according to one of claims 1 to 5, wherein each change vector «D» is assigned at least one indication of a user property, in particular - an indication of the period of untreated hearing loss of the user before using a hearing system (2), - an indication of the user's age, - an indication of the gender of the user, and / or - an indication of the user's musical background is assigned, and for the calculation of the change center «D» only those change vectors (D) are taken into account which agree in the or each assigned user characteristic.

7. Method according to one of claims 1 to 6, wherein the change center «D» is only used in a later adjustment of one of the hearing systems (2) of the test group (42) or of another hearing system (2) if it does not exceed a minimum threshold falls below, in particular if its amount does not fall below a specified minimum value.

8. Method according to one of claims 1 to 7, wherein a scatter of the change vectors (D) from which the change center of gravity «D» was calculated is determined, and wherein the change center of gravity «D» is only used in a later adjustment of one of the hearing systems (2) of the test group (42) or of a further hearing system (2) if the scatter does not exceed a predetermined maximum value.

9. Method according to one of claims 1 to 8, wherein for all hearing systems (2) of the test group (42) the respective initial setting (Pin), the respective final setting (Pend) and the voice information are created in a central database (34), in particular in a data cloud, and wherein the change vector (D) is calculated in a cloud server (36) for each hearing system (2) of the test group (42).

10. Method according to one of claims 1 to 9, wherein for each hearing system (2) of the test group (42) a current setting of the signal processing parameters (pi) is adopted as the final setting (Pend) if the current setting of the signal processing parameters (pi) is no longer changed over a predetermined period of time.

11. Device (30) for adjusting the signal processing of hearing systems (2), each having at least one hearing instrument (4) that can be parameterized by a plurality of signal processing parameters (pi), - with at least one adjustment unit (32) for making an initial setting (Pin) of the signal processing parameters (pi) of each hearing system (2) of a test group (42) of hearing systems (2), and for making a final setting (Pend) of the signal processing parameters (pi) of each hearing system (2) of the test group (42) by changing the signal processing parameters (ps) one or more times, - with an evaluation unit (36) for determining a change vector (D) in the parameter space spanned by the signal processing parameters (pi) from the difference between the final setting (Pend) and the initial setting (Pin) of the signal processing parameters (ps), each change vector (D) being assigned a language information is assigned, - wherein the evaluation unit (36) is designed to determine a center of gravity of these change vectors (D) in the parameter space as the center of gravity of change «D» from change vectors (D) which are assigned to the same language or to a language group containing this language according to the associated language specification, and - wherein the adaptation unit or a further adaptation unit (40) of the device (30) is designed to use this change center of gravity «D» in a later adjustment of one of the hearing systems (2) of the test group (42) or of a further hearing system (2).

Images