Hearing device system and method of operation thereof

Abstract

The invention relates to a hearing device system and a method for operating the same. The hearing device system (2) comprises a hearing aid device (4) and a peripheral device (6) providing a digital aid device (16) for voice control, wherein a microphone signal (S) is checked by the hearing aid device (4) for a voice component of a wearer of the hearing aid device (4). M ), wherein, when the self-speech component is recognized, the hearing aid device (4) checks the microphone signal (S) for a predefined activation instruction of the digital assistant device (16) M ), wherein, when the activation instruction is recognized, the hearing aid device (4) performs an operation on the microphone signal (S) according to the algorithm (A) M ) represents the activation instruction and the subsequent speech sequence (S T ) is processed for speech recognition by a digital assistant device (16), and wherein the microphone signal (S M ) after processing (S b ) is transmitted to the peripheral device (6).

Description

Technical Field

[0001] The invention relates to a method for operating a hearing device system. Furthermore, the invention relates to a hearing device system. Background Art

[0002] Hearing aids are used to provide hearing-impaired or hearing-impaired people who usually wear such devices all the time or most of the time to compensate for hearing defects. Here, hearing aids can be provided in different structural forms to meet the individual needs of the wearer. For example, the hearing aid is constructed as a behind-the-ear hearing device ( HdO,behind the ear,“BTE”), hearing devices with external earpieces (reciever in canal,“RIC”), or in-the-ear hearing devices ( “IdO”).

[0003] The hearing aid device has at least one electroacoustic converter, in particular a microphone, which collects (air) sound from the wearer's surroundings. The corresponding signal output by the converter is processed, in particular amplified, by a signal processing unit according to the wearer's needs. The processed signal is guided to the earpiece (speaker) and output by the earpiece as sound to the ear canal or the wearer's eardrum.

[0004] Here, the hearing aid can be coupled to a digital (virtual) assistant of a peripheral device, which is configured to provide (assistance) services. To this end, speech recognition is performed and the recognized speech or its content is interpreted by the digital assistant.

[0005] For example, DE 10 2018 209 719 A1 discloses a method for operating a hearing device system and identifying a user of a hearing device, the hearing device system having a hearing device and a virtual assistant, the user sending a voice command to at least one virtual assistant, wherein the content of the voice command is received by the at least one virtual assistant, and the hearing device transmits an information signal containing information about the presence of a release condition to the at least one virtual assistant. In this case, the content of the voice command is executed by the at least one virtual assistant as a function of the information about the presence of the release condition.

[0006] DE 10 2018 209 822 A1 describes a method for controlling data transmission between at least one hearing device and a peripheral device of a hearing device system, wherein a self-speech component of a hearing device wearer is automatically checked in a microphone signal captured by the hearing device and wherein, when the self-speech component is identified, a subsequent audio sequence of the microphone signal is transmitted from the hearing device to the peripheral device. Summary of the invention

[0007] The technical problem to be solved by the present invention is to specify a particularly suitable method for operating a hearing instrument system having a hearing aid and a peripheral device of a digital assistant providing voice control. In particular, an incorrect recognition and / or an incorrect interpretation of voice sequences by the digital assistant is to be avoided. Furthermore, such a hearing instrument system and a corresponding hearing aid are to be specified.

[0008] In terms of the method, the above technical problem is solved by the features of the present invention. In terms of the hearing aid device, the above technical problem is solved by the features of the present invention, and in terms of the hearing device system, the above technical problem is solved by the features of the present invention. Advantageous expansion schemes and design schemes are the subject of the following description. Here, the description in conjunction with the method also applies to the hearing device system, and vice versa.

[0009] The method is used to operate a hearing device system having a hearing aid and a peripheral device. The peripheral device, which is designed, for example, as a smartphone or a so-called smart speaker, can be coupled to the hearing aid by signal transmission technology and is also appropriately coupled. In this case, the peripheral device provides a voice-controlled digital assistant, which is also called a virtual assistant. For example, the voice-controlled digital assistant is Apple's "Siri", Amazon's "Alexa" or the Google Apps "Google Now".

[0010] According to the method, the microphone signal from the hearing aid device is preferably automatically checked for a self-speech component of the wearer of the hearing aid device. In other words, the microphone signal is analyzed for the presence of the wearer's own speech. Here, the microphone signal represents the ambient sound collected by the microphone of the hearing aid device, which is converted into the microphone signal by the microphone.

[0011] In this case, the self-speech recognition (English: "own voice detection") is performed, for example, by means of a level comparison of the signal of an (ambient) microphone worn on or behind the pinna and a (ear canal) microphone worn in the ear canal of the wearer. Alternatively or additionally, in order to recognize the wearer's own voice, a spatial filter, in particular monaural or binaural beamforming, is applied to the signal of the microphone of the hearing aid device, so that the sound from the wearer's mouth can be determined.

[0012] Furthermore, according to the method, when the self-speech component is recognized, the hearing aid device preferably automatically checks the microphone signal for a predefined activation command for the digital assistant. In this case, the activation command is preferably stored in a memory of the hearing aid device. That is, the hearing aid device performs speech recognition (word recognition) on the microphone signal. In particular, the activation command is used to activate the speech interface of the digital assistant.

[0013] If an activation command is recognized, the hearing aid device processes the portion of the microphone signal that represents the activation command and the subsequent speech sequence of the wearer's words for speech recognition of the digital assistant device using an algorithm. For this purpose, the hearing aid device has, in particular, a signal processing unit. In the following, this portion of the microphone signal that represents the activation command and the subsequent speech sequence of the wearer is also referred to as speech signal.

[0014] The processed part of the microphone signal is subsequently transmitted to the peripheral device and expediently forwarded further to the digital assistant for speech recognition by the digital assistant.

[0015] The transmission is preferably carried out in a (live) "streaming" manner. When an activation command is recognized, the activation command and the subsequent voice sequence are recorded and transmitted directly to the peripheral device, in particular as an audio signal, ie as a signal in an audio format such as "wma" or "mp3".

[0016] Thus, based on the processing performed by the hearing aid, it is advantageously possible to improve the recognition of speech by the speech recognition of the digital assistant. In particular, here, instead of processing the signal to obtain better intelligibility for a person, or instead of processing the signal to compensate for the hearing loss of the wearer of the hearing aid, for example using different algorithms, filters or parameters, for which the processed speech signal or the corresponding output sound is uncomfortable and / or unaccustomed to the person, but is more suitable for speech recognition of the digital assistant in terms of recognition rate.

[0017] The speech signal is preferably processed so that the recognition rate of the activation command or the subsequent speech sequence or its words by speech recognition is as high as possible and as correct as possible. In other words, the speech signal is adjusted for speech recognition of the digital assistant.

[0018] According to a preferred extension, the algorithm is adaptive for this purpose. In other words, the algorithm is changeable, in particular adjustable. The algorithm is preferably adjustable and the algorithm is preferably adjusted so that the recognition rate of the speech recognition of the digital assistant is improved. For this purpose, for example, the parameters, coefficients and / or functions of the algorithm are adjusted. Here, for example, the adjustment is performed according to an optimization method. For example, the algorithm includes a neural network or includes machine learning.

[0019] According to a preferred embodiment, the algorithm is adjusted or adapted and / or adapted by means of training, preferably for the purpose of improving the recognition rate of speech recognition of the digital assistant device. For example, such training is performed during a general fitting process by an acoustician, in a training operating mode of the hearing aid device provided for this purpose, or during normal operation of the hearing aid device.

[0020] The algorithm is preferably adjusted when the recognition rate is below a preset threshold value. To determine the recognition rate, for example, a response signal is output from the digital assistant device or the peripheral device to the hearing assistant device, wherein the response signal contains information about whether the activation command and / or the subsequent speech sequence was recognized by the speech recognition of the digital assistant device or was recognized with sufficient reliability.

[0021] Overall, it is particularly preferred that the algorithm is adapted and / or can be adapted for speech recognition of the digital assistant, in particular for the highest possible recognition rate of an activation command and / or a subsequent speech sequence.

[0022] According to a suitable design, the algorithm includes a spatial filter, in particular a monaural or binaural beamforming. Thus, when processing the portion of the microphone signal representing the activation command and the speech sequence that follows it, the spatial filter is applied to this portion of the microphone signal. Thus, noise / speech sources that are spatially spaced apart from one another can be amplified or attenuated accordingly.

[0023] Interference sources are preferably suppressed by means of beamforming, i.e. their signal strength is reduced. Compared to beamforming for processing signals for output to the wearer, relatively "aggressive" filters can be used here, since the processed speech signal is not output to the wearer and is perceived as comfortable by the wearer. Instead, only the recognition rate of the speech recognition of the digital assistant is to be increased. It is therefore particularly preferred to reduce the components of the speech signal that do not originate from the wearer's mouth to zero and / or to use a relatively small (spatial) area as the wearer's mouth.

[0024] For example, a spatial filter for self-speech recognition is used here. However, it is preferred to use another spatial filter which does not change the self-speech component of the speech signal but reduces, preferably to zero, other parts of the speech signal.

[0025] According to a suitable design, the algorithm includes a noise filter. The noise filter reduces the noise in the speech signal, i.e. the part of the microphone signal representing the activation command and the subsequent speech sequence of the wearer, according to the signal-to-noise ratio of the speech signal. In general, when processing the speech signal, the noise filter is applied to the speech signal.

[0026] In particular, the intensity of the noise reduction, i.e. the amount of noise attenuation, has already been adjusted or regulated by means of training that has been carried out beforehand, for example in the course of a general adaptation of the hearing aid device at an acoustician. Additionally or alternatively, the intensity of the noise reduction can be adjusted by means of (further) training.

[0027] According to a suitable embodiment of the method, the algorithm includes a spectral filter, which in particular selectively modifies the spectral components of the speech signal. The spectral filter reduces the components of the speech signal in a suitable manner in the frequency range of low frequencies, in particular frequencies less than 500 Hz or less than 200 Hz. In this case, these components can also be reduced to zero, so that these components are removed from the speech signal. As an alternative or in addition, the components of the speech signal in the frequency range of high frequencies, in particular frequencies greater than 4 kHz, i.e. in the range of fricatives and consonants, are amplified in a suitable manner by means of the spectral filter.

[0028] The spectral filter can be used to improve the recognizability of the activation command and / or the speech sequence that follows it, in particular in the case of a relatively low signal-to-noise ratio.

[0029] For example, during the training process, the frequency range for reducing the low-frequency components of the speech signal and / or for amplifying the high-frequency components of the speech signal is adjusted. Additionally or alternatively, the amount of attenuation or amplification can be adjusted during the training process.

[0030] The spatial filter, the noise filter and / or the spectral filter preferably processes the speech signal only in the frequency space.

[0031] According to a suitable embodiment of the method, during the processing of the speech signal, speech recognition is performed by the hearing aid device, in particular by its speech recognition unit. In this case, according to an expedient manner, after each recognized activation command and / or word of the speech sequence, a (word) separation signal, in particular a zero signal, is inserted at the corresponding point in the speech signal. In other words, the activation command and / or the words of the speech sequence are thus separated from each other.

[0032] In this way, the recognition of the activation command and / or the words of the speech sequence by the speech recognition of the digital assistant is facilitated even in the case of unclear pronunciation.

[0033] According to a suitable embodiment of the method, if the speech recognition of the peripheral device or the digital assistant does not recognize or does not recognize sufficiently reliably the activation command and / or the speech sequence following it, the peripheral device, in particular the digital assistant, transmits an error signal to the hearing aid. In this case, the speech signal is processed again with the aid of an algorithm, wherein the algorithm was changed beforehand. In particular, the parameters, coefficients and / or functions of the algorithm are adjusted in a manner similar to training. In this case, the adjustment is performed, for example, according to an optimization method for the recognition rate.

[0034] Furthermore, it is provided that the reprocessed speech signal is transmitted to the peripheral device and there to the digital assistant for speech recognition, thereby advantageously enabling the activation command and the subsequent speech sequence to be analyzed again by the speech recognition of the digital assistant without the wearer of the hearing device having to repeat the activation command or the speech sequence.

[0035] In this case, the reprocessing with the modified algorithm and the retransmission preferably take place automatically. In any case, the comfort of the wearer of the hearing aid device is advantageously increased.

[0036] According to the invention, the hearing aid of the hearing device system has at least one microphone, by means of which sounds from the environment of the hearing aid are picked up and a microphone signal corresponding to the sounds can be generated.

[0037] Furthermore, the hearing aid device has not only a self-speech recognition unit (which is coupled to the microphone in signal transmission technology) for checking the microphone signal for a self-speech component of the wearer of the hearing aid device, but also a first speech recognition unit for checking the microphone signal for predefined activation instructions for a voice-controlled digital assistant device of an associated peripheral device.

[0038] Furthermore, the hearing aid device comprises a signal processing unit which provides an algorithm for processing the speech signal. Here, the signal processing unit is configured to process the part of the microphone signal which represents the activation instruction and the speech sequence which follows it according to the algorithm, i.e., to process the speech signal according to the algorithm when a predefined activation instruction of the digital assistant is recognized. In particular, the algorithm is adapted and / or can be adapted to the digital assistant or the speech recognition of the digital assistant. Suitably, the signal processing unit is coupled to the self-speech recognition unit and / or to the first speech recognition unit via a signal transmission technique.

[0039] In order to transmit the processed part of the microphone signal to the peripheral device, the hearing aid device has a first transmission unit. The first transmission unit is preferably configured for wireless transmission or radio transmission, for example by means of Bluetooth or Bluetooth LE (Bluetooth Low Energy).

[0040] According to the invention, a hearing device system that operates or can be operated, in particular according to the method in one of the above-described variants, has at least one hearing aid device. For example, the hearing aid device system has a second hearing aid device for binaural listening of the wearer, wherein the two hearing aid devices can be coupled to each other in signal transmission technology and are also coupled to each other.

[0041] In addition, the hearing device system comprises a peripheral device with a voice-controlled digital assistant. The peripheral device comprises a second transmission unit for receiving a portion of the microphone signal transmitted by the hearing assistant. Additionally or alternatively, the second transmission unit is used to transmit a response signal, for example an audio signal for outputting a sound by a receiver (loudspeaker) of the hearing assistant, and / or an error signal (if the activation command and / or a subsequent voice sequence is not or not sufficiently reliably recognized by means of voice recognition of the peripheral device), to the hearing assistant. For voice recognition of the activation command and / or the voice sequence, the peripheral device has a second voice recognition unit, which is coupled to the digital assistant, in particular controlled and / or used by the digital assistant. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0043] Figure 1 A hearing device system is schematically shown having a hearing aid device and a peripheral device providing a voice-controlled digital aid, and

[0044] Figure 2The method flow for operating a hearing device system is shown, wherein a portion of a microphone signal representing an activation command for a digital assistant and a subsequent speech sequence is processed by the hearing assistant and wherein the processed portion is transmitted to a digital assistant of a peripheral device.

[0045] Parts and variables that correspond to one another are always provided with the same reference symbols in all the figures. DETAILED DESCRIPTION

[0046] Figure 1 A hearing device system 2 is shown, which has a hearing aid 4 and a peripheral device 6. The hearing aid comprises two ambient microphones 8, by means of which sounds can be picked up from the surroundings of a wearer of the hearing aid 4. Furthermore, the hearing aid 4 has an ear canal microphone 10 for picking up sounds in the ear canal of the wearer.

[0047] The ambient microphone 8 and the ear canal microphone 10 are electroacoustic converters. They convert the collected sound in the wearer's surrounding environment (also referred to as ambient sound) or the collected sound in the ear canal (also referred to as ear canal sound) into a microphone electrical signal S M , S M '、S M ”.

[0048] Microphone signal S M , S M '、S M " can be output to the self-speech recognition unit. For this purpose, the microphones 8 and 10 are connected to the self-speech recognition unit 12 by means of signal transmission technology.

[0049] Here, the self-speech recognition unit 12 is configured to perform self-speech recognition. In other words, the microphone signal S is evaluated for the presence of the wearer's own voice by means of the self-speech recognition unit 12, for example by means of a filter suitable for this or by means of a level comparison. M , S M '、S M " or a signal derived therefrom for analysis.

[0050] exist Figure 1 And hereinafter, for the purpose of better understanding, only reference is made to the microphone signal S M The following description also applies in a similar manner to the microphone signal S M ' and S M ” or a signal derived therefrom.

[0051] Furthermore, the hearing aid device 4 comprises a first (word recognition unit) speech recognition unit 14 coupled to the speech recognition unit 12 by means of a signal transmission technique for receiving a predefined activation command for a voice-controlled digital aid 16 for the associated peripheral device 6, and for receiving the microphone signal S M A check (analysis) is performed. In this case, the activation command is stored in a memory of the hearing aid device 4 (not shown in any more detail).

[0052] In order to process the microphone signal S M or part of it T The hearing aid device 4 includes a signal processing unit 18 coupled to the speech recognition unit 12. The signal processing unit 18 provides a signal processing unit for processing the microphone signal S M or part of it T Here, the signal processing unit 18 is configured to process the microphone signal S according to the algorithm A when a predefined activation instruction of the digital assistant device 16 is recognized. M The part S in the activation command and the speech sequence following it T . This part of the microphone signal S T Also referred to below as speech signal S T .

[0053] The hearing aid device has a first transmission unit 20 for transmitting the processed speech signal S T,b Wireless transmission to the peripheral device 6. For this purpose, the first transmission unit is coupled to the signal processing unit 18.

[0054] In addition, the first transmission unit 20 and the self-speech recognition unit 12 are coupled to the amplification unit 22 by signal transmission technology. The amplification unit 22 processes and amplifies the audio signal or the microphone signal S received by the first transmission unit 20 according to the hearing impairment of the wearer. M Or the microphone signal S processed by the self-speech recognition unit 12 for self-speech recognition M , and the corresponding earpiece signal S H The output is to an earpiece 24 (here a micro-speaker), which is also introduced into the ear canal of the wearer. The earpiece 24 converts the earpiece signal into sound and outputs the sound into the ear canal of the wearer. According to an alternative solution not shown in detail, the amplifier unit 22 is integrated in the signal processing unit 18.

[0055] also, Figure 1 The hearing device system 2 shown has a peripheral device 6. The peripheral device 6 comprises a second transmission unit 26 for receiving the processed speech signal S transmitted by the hearing aid device 4. T,b , and / or for transmitting the response signal S A, for example, an audio signal or an error signal is transmitted to the hearing aid device 4 .

[0056] The peripheral device provides a voice-controlled digital assistant 16. The digital assistant 16 controls and / or uses a second speech recognition unit 28 of the peripheral device 6 coupled to the second transmission unit 26 in signal transmission technology, and an analysis unit 30 coupled to the second speech recognition unit 28, the second speech recognition unit 28 being used to process the transmitted processed speech signal S T,b Speech recognition is performed and the analysis unit 30 is used to analyze the words or recognized word sequences recognized by the second speech recognition unit 28 .

[0057] Figure 2 A method for operating a hearing instrument system 2 is shown. In a first step I., sound is picked up by an ambient microphone 8 and converted into a microphone signal S representing the sound. M . Microphone signal S M is output to the self-speech recognition unit 12. In the second step II., the self-speech recognition unit 12 processes the microphone signal S M Performs own voice detection.

[0058] If the wearer's own voice of the hearing aid 4 is recognized in step II., then in a third step III., the microphone signal S is processed by means of the first voice recognition unit 14 of the hearing aid 4 for the activation command of the digital assistant 16. M The analysis is performed and the activation instructions are stored in the memory of the hearing aid device. Here, the words or word sequences determined as a result of the analysis are compared with the stored activation instructions.

[0059] If an activation command is recognized, ie the determined word or the determined word sequence corresponds to a stored activation command, then in a fourth step IV. the microphone signal S is processed according to the algorithm A. M , more precisely processing the microphone signal S M The part S in the activation instruction and the speech sequence (word sequence) immediately following it T .

[0060] Here, algorithm A includes a spatial filter F R , spatial filter F R It is designed as a monaural or binaural beamforming in order to separate the own speech of the wearer of the hearing aid device 4 from other noise or speech sources. R The recognized speech signal S T The self-speech component in the speech signal S T The other components in are reduced to zero.

[0061] Additionally or alternatively, algorithm A includes a noise filter F N , noise filter F N Reduce the speech signal S T The strength of the reduction depends on the speech signal S T The smaller the signal-to-noise ratio SNR, the greater the reduction.

[0062] Additionally or alternatively, the algorithm includes a spectral filter F S , spectral filter F S Reduce (attenuate) the speech signal S T The components corresponding to frequencies less than 5 kHz are also obtained by using the spectral filter F S To amplify the speech signal S T The component corresponding to frequencies greater than 10kHz.

[0063] In short, when processing the speech signal S T When the spatial filter F R , Noise filter F N and / or spectral filter F S Applied to this T.

[0064] As filter F R 、F N 、F S In addition or replacement, when processing the speech signal S T When the first speech recognition unit 14 is used to recognize the speech signal S T Perform voice recognition. For each recognized activation command and / or voice signal S T After the word of the speech sequence, in the speech signal S T The separation signal S0 is inserted at the corresponding point of the . Here, the separation signal S0 represents silence, i.e., no sound, for example. Therefore, the words of the activation instruction and / or voice sequence are separated from each other by means of the separation signal S0.

[0065] With the aid of the training carried out in advance in time, the algorithm A, in particular its filter F, is trained for the highest possible recognition rate of the speech recognition of the digital assistant 16. R 、F N 、F S And / or the separated signal S0, for example its (signal) length, is adjusted. This is done, for example, in the course of an adaptation during the training of the hearing instrument or in a training operating mode provided for training. In addition, if the digital assistant 16 processes the processed speech signal S T,bIf the speech recognition and / or analysis is unsuccessful, then when the hearing device system 2 is running, in particular also when the hearing device system 2 is running normally, according to the response signal S A (See Figure 2 ) to train algorithm A.

[0066] In summary, the algorithm is adaptive and, with the aid of training, is adapted to the speech recognition of the digital assistant device 16 .

[0067] For example, during the training process, in the noise filter F N In the training process, the spectral filter F S For example, the frequency range for reducing a component of the speech signal and / or for amplifying a component of the speech signal is adjusted in the training, and / or the amount of reduction or amplification is adjusted. In addition, during the training process, for example, the amount of reduction of interference noise is adjusted in the spatial filter, wherein interference noise is understood to be noise or speech that does not originate from the area associated with the wearer's mouth. Alternatively or additionally, the area associated with the mouth can be adjusted.

[0068] Subsequently, in a fifth step V., the speech signal S processed in this way is transmitted by means of the first transmission unit 20 and the second transmission unit 26. T,b , that is, the microphone signal S M The processed part S of the activation command and the speech sequence immediately following it T,b , transmitted from the hearing assistance device 4 to the peripheral device 6.

[0069] In step VI, the processed speech signal S transmitted to the peripheral device 6 is processed according to the speech recognition of the digital assistant device 16. T,b The digital assistant 16 uses the second speech recognition unit 28 for this purpose. Furthermore, in step VI., the digital assistant 16 analyzes and interprets the content of the word sequence thus recognized (which contains the activation command and the subsequent speech sequence) with the aid of the analysis unit 30. If, after the activation command, a further command is recognized in the speech sequence, this further command is executed with the aid of the digital assistant 16. In this case, the audio signal is optionally transmitted as the response signal S in step VII. A The audio signal is transmitted to the hearing aid device 4 , wherein the audio signal is processed by the amplification unit 22 and output as sound via the earpiece.

[0070] However, if the digital assistant does not recognize the activation command and / or the subsequent voice sequence by voice recognition, or if the evaluation unit cannot determine a further command by interpretation, then in step VII. an error signal is output as response signal S Ato the hearing aid device 4. Figure 2 In this case, method steps VII. and VIII. are shown in dashed lines. In this case, the previously processed speech signal S transmitted to the peripheral device is automatically reprocessed in step VIII. T , wherein the algorithm used for processing has been changed. Here, the algorithm is adjusted and trained by means of an optimization method, wherein the optimization method is configured to improve the speech recognition of the digital assistant for the speech signal S T In addition, in step VIII., the reprocessed signal S T,b It is retransmitted to the peripheral device 6 for speech recognition and analysis by the digital assistant 16 .

[0071] The present invention is not limited to the embodiments described above. On the contrary, those skilled in the art can also derive other variations of the present invention therefrom without departing from the subject matter of the present invention. In particular, all individual features described in conjunction with the embodiments can also be combined with each other in other ways without departing from the subject matter of the present invention.

[0072] Reference numerals list

[0073] 1 Hearing Device System

[0074] 4 Hearing Assistive Devices

[0075] 6 Peripheral devices

[0076] 8 (ambient) microphones

[0077] 10 (Ear canal) microphone

[0078] 12 Self-speech recognition unit

[0079] 14 Speech recognition unit for hearing aids

[0080] 16 Digital Assistance Devices

[0081] 18 Signal processing unit for hearing aid equipment

[0082] 20 Transmission unit for hearing aids

[0083] 26 Transmission unit for peripheral devices

[0084] 28 Speech recognition unit for peripheral devices

[0085] 24 Receiver

[0086] 22 Amplification unit

[0087] 30 Analysis Units

[0088] A Algorithm

[0089] F N Noise Filter

[0090] F R Spatial filter

[0091] F S Spectral filter

[0092] S A Response signal

[0093] S b Processed part of the microphone signal

[0094] S H Handset signal

[0095] S M Microphone signal

[0096] S T Part of the microphone signal

[0097] S T,b Processed part of the microphone signal

[0098] S0 separation signal

[0099] SNR Signal to Noise Ratio

[0100] I. Collect sound and generate microphone signal

[0101] II. Performing self-speech recognition on microphone signals

[0102] III. Speech Recognition

[0103] IV. Part that processes microphone signals

[0104] V. Transmit the processed portion of the microphone signal

[0105] VI. Speech recognition and analysis via digital assistants

[0106] VII. Transmitting a response signal to the hearing aid device

[0107] VIII. Reprocessing and transmitting microphone signals

Claims

1. A method for operating a hearing device system (2) having a hearing aid device (4) and a peripheral device (6) providing a voice-controlled digital aid (16), -in, For the presence of the wearer's own voice of the hearing aid device (4), the microphone signal (S) is checked by the hearing aid device (4) for the wearer's own voice component. M ), wherein, when the self-speech component is identified, the hearing aid device (4) checks the microphone signal (S) for a predefined activation instruction of the digital assistant device (16) M ), - wherein, when the activation instruction is recognized, the hearing aid device (4) uses an algorithm (A) to process the microphone signal (S M ) represents the activation instruction and the subsequent speech sequence (S T ) for speech recognition by the digital assistant device (16), and - Among them, the microphone signal (S M ) after processing (S T,b ) is transmitted to the peripheral device (6), It is characterized in that If the voice recognition of the peripheral device (6) does not recognize the activation command and / or the voice sequence, the peripheral device (6) constructs a response signal (S) as an error signal A ) is transmitted to the hearing aid device, - Reuse the modified algorithm (A) to process the microphone signal (S M ) part (S T ),as well as - Transmit the microphone signal (S M ) of the reprocessed part (S T ) is transmitted to the peripheral device (6).

2. The method according to claim 1, It is characterized in that The algorithm (A) is adaptive.

3. The method according to claim 1, It is characterized in that The algorithm (A) is adapted and / or adjusted by means of training.

4. The method according to claim 1, It is characterized in that The algorithm (A) comprises a spatial filter (F R ).

5. The method according to claim 4, It is characterized in that The components of the speech signal not coming from the wearer's mouth are reduced to zero, wherein the speech signal (S T ) is defined as the portion of the microphone signal that represents the activation command and the wearer's speech sequence that follows it.

6. The method according to claim 1, It is characterized in that The algorithm (A) includes a noise filter (F N ), the noise filter according to the part (S T ) reduces the signal-to-noise ratio (SNR) of the microphone signal (S M ) of the portion (S T ) noise.

7. The method according to claim 1, It is characterized in that The algorithm (A) comprises a spectral filter (F S ), the spectral filter modifies the microphone signal (S M ) part (S T ) , where in particular the attenuated microphone signal (S M ) in the low frequency range (S T ), and / or amplify the microphone signal (S M ) in the high frequency range (S T ).

8. The method according to claim 7, It is characterized in that Attenuate microphone signal (S M ) less than 500 Hz, and / or amplifying the microphone signal (S M ) with a frequency greater than 4kHz.

9. The method according to any one of claims 1 to 8, It is characterized in that In processing the microphone signal (S M ) represents the activation instruction and the subsequent speech sequence (S T ), speech recognition is performed with the aid of the hearing aid (4), wherein, after the recognized activation command and / or the word of the speech sequence, the microphone signal (S M ) part (S T ) is inserted into the separation signal (S0) at the corresponding point of the .

10. A hearing aid device (4) for use in a hearing device system (2), comprising: - at least one microphone (8) for generating a microphone signal (S M ), a self-speech recognition unit (12) for checking the microphone signal (S) for a self-speech component of the wearer of the hearing aid device (4); M ), a first speech recognition unit (14) for checking a microphone signal (S) for a predefined activation command of a digital auxiliary device (16) for voice control of a peripheral device (6) M ), a signal processing unit (18) which provides an algorithm (A) and is configured to, when a predefined activation instruction is recognized, process the microphone signal (S) with the aid of the algorithm (A) M ) for speech recognition by the digital assistant device (16), and - a first transmission unit (20), for transmitting a microphone signal (S M ) is transmitted to the peripheral device (6), It is characterized in that The signal processing unit (18) is configured to, if the peripheral device (6) sends a response signal (S A ) is transmitted to the hearing aid device, and the microphone signal (S) is processed again using the changed algorithm (A) M ) part (S T ).

11. A hearing device system (2), in particular operated according to the method according to any one of claims 1 to 9, comprising a hearing aid device (4) according to claim 10 and a peripheral device (6) having a voice-controlled digital aid (16), wherein: The peripheral device (6) has a second transmission unit (26) and a second speech recognition unit (28), wherein the second transmission unit is used to receive the microphone signal (S) transmitted and processed by the hearing aid device (4). M ) part (S T,b ), and / or for transmitting a response signal (S A ) is transmitted to the hearing aid device (4), and the second speech recognition unit is used to perform speech recognition on the activation instruction and / or speech sequence.

Images