Abstract
A hearing device, e.g. a hearing aid or a headset, configured to be worn at or in an ear of a user, the hearing device comprising at least one input transducer for converting a sound in an environment of the hearing device to at least one electric input signal representing said sound; an own voice detector configured to estimate whether or not, or with what probability, said sound originates from the voice of the user, and to provide an own voice control signal indicative thereof, a face mask detector configured to estimate whether or not, or with what probability, said user wears a face mask while speaking, and to provide face mask control signal indicative thereof. A method of operating a hearing device is further disclosed. Thereby an improved hearing aid or headset may be provided.
Claims
1. A hearing device configured to be worn at or in an ear of a user, the hearing device comprising: at least one input transducer for converting a sound in an environment of the hearing device to at least one electric input signal representing said sound; an own voice detector configured to estimate whether or not, or with what probability, said sound originates from the voice of the user, and to provide an own voice control signal indicative thereof, wherein the hearing device further comprises: a mouth wear detector configured to estimate whether or not, or with what probability, said user wears a mouth wear while speaking, and to provide mouth wear control signal indicative thereof.
2. A hearing device according to claim 1 comprising a feature extractor configured to identify acoustic features in the at least one electric input signals indicative of the user's own voice.
3. A hearing device according to claim 2 comprising a memory wherein reference values of said acoustic features extracted from said at least one electric input signal when the user wears the hearing device and speaks, while not wearing a mouth wear, are stored.
4. A hearing device according to claim 2 comprising a memory wherein differences in reference values of said acoustic features extracted from said at least one electric input signal when the user wears the hearing device and speaks, while wearing and while not wearing a mouth wear, are stored.
5. A hearing device according to claim 1 comprising a signal processor for processing said at least one electric input signal, or one or more signals based thereon, and to provide a processed signal.
6. A hearing device according to claim 1 comprising an output transducer for converting an electric output signal to stimuli perceivable by the user as sound.
7. A hearing device according to claim 5 wherein said signal processor is configured to control processing of said at least one electric input signal, or one or more signals based thereon in dependence of said mouth wear control signal.
8. A hearing device according to claim 1 comprising at least two input transducers providing at least two electric input signals.
9. A hearing device according to claim 8 comprising an own voice beamformer configured to provide an estimate of the voice of the user in dependence of said at least two electric input signals and configurable beamformer weights of the own voice beamformer.
10. A hearing device according to claim 9 wherein the hearing device comprises a signal processor configured to process said estimate of the voice of the user in dependence of said mouth wear control signal, and to provide an improved estimate of the voice of the user.
11. A hearing device according to claim 10 wherein the signal processor is configured to modify the frequency shape of the user's own voice in dependence of said mouth wear control signal and to provide said improved estimate of the voice of the user.
12. A hearing device according to claim 1 comprising a transceiver configured to transmit and/or receive audio signals from another device or system.
13. A hearing device according to claim 1 comprising a keyword detector configured to identify a specific keyword of key phrase in the at least one electric input signal or a signal derived therefrom in dependence of said own voice control signal and said mouth wear control signal.
14. A hearing device according to claim 13 comprising a voice control interface configured to control functionality of the hearing device by predefined spoken commands, when detected by said keyword detector.
15. A hearing device according to claim 1 comprising or being connectable to a user interface allowing the user to indicate a specific kind of mouth wear that the user may occasionally wear.
16. A hearing device according to claim 1 configured to identify a current location or receive information about a current location from another device and configured to trigger a reminder regarding whether or not a user is currently wearing a mouth wear based on the mouth wear control signal.
17. A hearing device according to claim 1 wherein the own voice detector and/or the mouth wear detector is fully or partially implemented using a learning algorithm.
18. A hearing device according to claim 1 being constituted by or comprising a headset, an air-conduction type hearing aid, a bone-conduction type hearing aid, a cochlear implant type hearing aid, or a combination thereof.
19. A hearing device according to claim 1, wherein the mouth wear comprises a mouthpiece, a face mask, or other face or mouth covering device or item.
20. A method of operating a hearing device configured to be worn at or in an ear of a user, the method comprising: converting a sound in an environment of the hearing device to at least one electric input signal representing said sound; estimating whether or not, or with what probability, said sound originates from the voice of the user, and providing an own voice control signal indicative thereof; and estimating whether or not, or with what probability, said user wears a mouth wear while speaking, and to providing a mouth wear control signal indicative thereof.
21. A non-transitory computer readable medium storing an application, termed an APP, comprising executable instructions configured to be executed on an auxiliary device to implement a user interface for a hearing device according to claim 1, the APP being configured to exchange data with said hearing device and to allow the user to indicate a kind of mouth wear that the user might wear, said kind of mouth wear being selectable among a multitude of different types of mouth wears, and to communicate information related the selected mouth wear to the hearing device.
Description
BRIEF DESCRIPTION OF DRAWINGS
[0095] The aspects of the disclosure may be best understood from the following detailed description taken in conjunction with the accompanying figures. The figures are schematic and simplified for clarity, and they just show details to improve the understanding of the claims, while other details are left out. Throughout, the same reference numerals are used for identical or corresponding parts. The individual features of each aspect may each be combined with any or all features of the other aspects. These and other aspects, features and/or technical effect will be apparent from and elucidated with reference to the illustrations described hereinafter in which:
[0096] FIG. 1A shows a user speaking while wearing a binaural hearing aid system comprising first and second hearing devices; and
[0097] FIG. 1B shows the user of FIG. 1A, while simultaneously wearing a face mask,
[0098] FIG. 2A shows a part of a hearing device comprising an own voice detector according to a first embodiment of the present disclosure:
[0099] FIG. 2B shows a part of a hearing device comprising an own voice detector according to a second embodiment of the present disclosure,
[0100] FIG. 2C shows an own voice processor according to the present disclosure implemented as a neural network,
[0101] FIG. 2D shows an own voice detector according to the present disclosure implemented as a neural network, and
[0102] FIG. 2E schematically illustrates different feature layers in an implementation of an own voice processor or own voice detector based on a neural network according to the present disclosure,
[0103] FIG. 3 shows a measurement of the difference between sound pressure level recorded without and with a face mask,
[0104] FIG. 4 shows a part of a hearing aid comprising an own voice detector and a face mask detector according to an embodiment of the present disclosure,
[0105] FIG. 5 shows an embodiment of an own voice processor according to the present disclosure,
[0106] FIG. 6 shows a hearing device according to an embodiment of the present disclosure comprising an own voice processor comprising an own voice detector and a face mask detector,
[0107] FIG. 7A shows a hearing system comprising a hearing aid and an auxiliary device in communication with each other, and
[0108] FIG. 7B shows the auxiliary device of FIG. 7A configured to implement a user interface for the hearing aid by running an application program from which a mode of operation of the hearing aid can be selected, and
[0109] FIG. 8 shows an embodiment of a headset or a hearing aid comprising own voice estimation and the option of transmitting the own voice estimate to another device, and to receive sound from another device for presentation to the user via a loudspeaker, e.g. mixed with sound from the environment of the user according to the present disclosure.
[0110] The figures are schematic and simplified for clarity, and they just show details which are essential to the understanding of the disclosure, while other details are left out. Throughout, the same reference signs are used for identical or corresponding parts.
[0111] Further scope of applicability of the present disclosure will become apparent from the detailed description given hereinafter. However, it should be understood that the detailed description and specific examples, while indicating preferred embodiments of the disclosure, are given by way of illustration only. Other embodiments may become apparent to those skilled in the art from the following detailed description.
DETAILED DESCRIPTION OF EMBODIMENTS
[0112] The detailed description set forth below in connection with the appended drawings is intended as a description of various configurations. The detailed description includes specific details for the purpose of providing a thorough understanding of various concepts. However, it will be apparent to those skilled in the art that these concepts may be practiced without these specific details. Several aspects of the apparatus and methods are described by various blocks, functional units, modules, components, circuits, steps, processes, algorithms, etc. (collectively referred to as “elements”). Depending upon particular application, design constraints or other reasons, these elements may be implemented using electronic hardware, computer program. or any combination thereof.
[0113] The electronic hardware may include micro-electronic-mechanical systems (MEMS), integrated circuits (e.g. application specific), microprocessors. microcontrollers, digital signal processors (DSPs), field programmable gate arrays (FPGAs), programmable logic devices (PLDs), gated logic, discrete hardware circuits, printed circuit boards (PCB) (e.g. flexible PCBs), and other suitable hardware configured to perform the various functionality described throughout this disclosure, e.g. sensors, e.g. for sensing and/or registering physical properties of the environment, the device, the user, etc. Computer program shall be construed broadly to mean instructions, instruction sets, code, code segments. program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures. functions, etc. whether referred to as software, firmware, middleware, microcode, hardware description language. or otherwise.
[0114] The present application relates to the field of hearing devices, e.g. hearing aids or headsets. The application deals with handling acoustic effects of a user's application of a mouthwear, e.g. a mouthpiece or a face mask (such as a surgical mask) on the detection and/or estimation of the user's own voice in a hearing device, such as a hearing aid or a headset. The present application deals in particular with detection of a user's own voice, and specifically with detection of the user's own voice while wearing a face mask or other face protection device or means. The present application is further focused on identifying and/or compensating acoustic changes due to such face mask or face protection.
[0115] When detecting or estimating a user's own voice, it is important to distinguish between when the hearing instrument user is talking with and without face mask (or other face or mouth covering device or item). FIG. 1A shows a user (User) speaking while wearing a binaural hearing aid system comprising first and second hearing devices (HD1, HD2). The fact that the user is speaking is indicated by solid arrows from the user's mouth (Mouth) to the right and left ears of the user (User), and thus to the first and second hearing devices (HD1, HD2), each comprising at least one input transducer for converting a sound in the environment of the hearing device to an electric input signal representing said sound, possibly including the user's own voice.
[0116] FIG. 1B shows the user of FIG. 1A, while simultaneously wearing a face mask (FM), e.g. a surgical mask.
[0117] A proposed solution is sketched in FIG. 2A. 2B. The solutions of FIGS. 2A and 2B may fully or partially be implemented using a learning algorithm, e.g. a trained neural network, e.g. a deep neural network as indicated in FIG. 2C, 2D.
[0118] FIG. 2A shows a part of a hearing device. e.g. a hearing aid, comprising an own voice processor detector (OVP) according to an embodiment of the present disclosure. The hearing device comprises a multitude M of input transducers. IT.sub.m, m=1, 2, . . . , M, here microphones. Other input transducers than microphones may be used e.g. vibration sensors. e.g. one or more accelerometers. Each microphone is configured to convert sound around the hearing device to an electric input signal x.sub.m. The input transducer. IT.sub.m, m=1.2, . . . , M. may comprise an analogue to digital converter for converting an analogue electric signal from a microphone to a digitized signal (x.sub.m, m=1, 2, . . . , M) comprising a stream of digitized samples. The input transducer (IT.sub.m) may comprise further circuitry for processing the input signal. such as e.g. an analysis filter bank to provide the electric input signal (x.sub.m) in a time frequency representation x.sub.m(k,n) as the case may be (k, n being frequency and time-frame indices, respectively). The exemplary own voice processor (OVP) of FIGS. 2A (and FIG. 2B) yields three output probabilities or binary values, denoted; No OV (‘No own voice’). OVxFM (‘own voice without face mask) and OV+FM (sown voice with face mask’). The confidence level of the output probabilities (or binary values) of a given hearing device may e.g. be further improved by comparison (e.g. combination) with a corresponding value from a contralateral device of a binaural hearing system (e.g. HD1, HD2 of FIG. 1A, 1B). The output probabilities (or binary values) may be further processed into decisions in other parts of the hearing device (e.g. in relation to estimating a user's own voice. e.g. in connection with a communication mode or to a voice control interface mode of operation of the hearing device, see e.g. FIG. 6, 7B). The transition between own voice and no own voice will in general change more frequently than the transition between mask and no mask. It is hence desirable that the OV/no OV decision can change/fluctuate more rapidly compared to the face mask/no face mask decision.
[0119] The own voice detection may be based on different features such as acoustic features (F.sub.1, F.sub.2, . . . . , F.sub.NA). This is illustrated in the (part of an) embodiment of a hearing device of FIG. 2B. comprising the same elements as the embodiment of FIG. 2A. Additionally, the embodiment of FIG. 2B comprises a feature extractor (FEX) for extracting features of the electric input signals (x.sub.1, x.sub.2, . . . , x.sub.M) and providing a number NA of acoustic features (F.sub.1. F.sub.2, . . . , F.sub.NA). The acoustic features (F.sub.1, F.sub.2, . . . , F.sub.NA) may e.g. be or comprise or relate to the microphone signals (x.sub.m) captured by the hearing device or signals derived from the microphone signals. such as: [0120] Magnitude or power spectrum of the microphone signals, or of one or more signals derived therefrom, [0121] Phase difference between the microphone signals. [0122] Relative transfer functions between the microphones (e.g. both magnitude and phase), [0123] Beamformed signals (such as signals provided by own voice cancelling beamformers. e.g. derived with and without face mask), or from the beamformer derived control signals (as the adaptive coefficient beta in the generalized sidelobe canceller)
[0124] The acoustic features (F.sub.1, F.sub.2, . . . , F.sub.NA) may further be influenced by one or more other input signals (O-INP), e.g. one or more signals from sensors or detectors, e.g. related to the acoustic environment, or to the user's present condition (movement/no movement, mental state. etc.).
[0125] The features (F.sub.1. F.sub.2, . . . , F.sub.NA) extracted by the feature extractor (FEX) are fed to an own voice detector (OVD). The own voice detector provides the three output probabilities or binary values of the own voice processor (OVP): No OV (‘No own voice’), OVxFM (‘own voice without face mask’) and OV+FM (‘own voice with face mask’).
[0126] Features derived from microphone signals in a binaural setup may as well be applied. In an embodiment at least one microphone is located in the ear canal. In addition to acoustic features, other features may as well be applied. E.g. vibrations picked up by an accelerometer located inside the hearing device or outside the hearing device. e.g. near the ear canal, may be used to distinguish between ‘OV’ or ‘No OV’ (not ‘FM’ or ‘no FM’). The own voice processor (FIG. 2C) or the own voice detector (FIG. 2D) or the feature extractor may be fully or partly based on a neural network trained on the different classes (e.g. own voice with and without a (possibly specific) mask or different masks, in different signal to noise environments, etc.). The weights of the neural network may be selected based on the type of used face mask (scarf, surgical mask, face visor, material used for the face mask, acoustic attenuation of the face mask, etc.). A number N.sub.FM of different sets of optimized parameters for neural networks may thus be provided, each corresponding to a specific type of face mask or face protection product (cf. e.g. FIG. 5).
[0127] FIG. 2E schematically illustrates different feature layers (‘Feature layer#q’, q=1, 2, 3, 4, . . . , N.sub.F, where N.sub.F is the number of feature layers) in an implementation of an own voice processor (OVP or own voice detector (OVD) based on a neural network (DNN) according to the present disclosure. The different feature layers may be provided by distinct functional blocks, e.g.: [0128] Analysis filter bank (FBA) providing the M electric (time domain) signals (x.sub.m, m=1, . . . , M) in a time-frequency representation (as M frequency domain signals X.sub.m, m=1, . . . , M). [0129] Beamformer filtering unit (BFU) provided a number of beamformed signals (or beamformers) BFp. p=1, . . . , N.sub.BF based on combinations of the electric input signals X.sub.m.
[0130] The different feature layers may, however additionally or alternatively be provided by outputs of different layers of a neural network, e.g. a deep neural network (DNN) comprising an input layer (‘IN-L’ receiving beamformed signals (or beamformers) BFp as inputs and providing features of ‘Feature layer#3’ as output), a number of intermediate (hidden) layers (‘INT-L’, ‘, . . . ’ providing’ ‘Feature layer#4’, . . . , ‘Feature layer#NF’), and an output layer (‘OUT-L’ providing functional outputs, here ‘No OV,’OV, ‘No FM’, ‘FM’, see e.g. also FIGS. 2A-2D and FIG. 4). The neural network (DNN) may e.g. include the beamformer filtering unit (BFU). The beamformer filtering unit (BFU) may thus form part of or constitute the feature extraction unit (FEX) in FIG. 2B or 4. But the feature extraction unit may also be considered as forming part of a neural network implementation of the functional feature (her an own voice processor or own voice detector or face mask detector according to the present disclosure).
[0131] FIG. 3 shows a measurement of the difference between sound pressure level recorded without and with a face mask. The two curves show the difference in level recorded at a microphone located at a hearing device mounted at the left and the right ear, respectively. Both graphs show the difference between no face mask and face mask (at ‘Left’ and ‘Right’ sides, respectively). At low frequencies (below a threshold frequency fth. e.g. below 4 or 5 kHz), the sound seems to be reflected from the face mask resulting in a relatively higher level (received at the ears while using face mask) at low frequencies (≤ f.sub.th) compared to the higher frequencies (>f.sub.th). The difference between the left and right ears at higher frequencies illustrated by FIG. 3 may e.g. be due to or at least influenced by minor facial face mask mounting asymmetries. This change of spectral tilt may be used as a feature to distinguish between whether the person is wearing a face mask. The hearing aid may comprise a memory wherein reference data for own voice reception at the microphones of the hearing aid are stored (cf. e.g. FIG. 5), e.g. focused on frequencies below the threshold frequency fth. Such data may e.g. include data as shown in FIG. 3. or equivalent, recorded while the user (or other person, or model) speaks with and without a face mask (or similar, e.g. visor).
[0132] At frequencies above the threshold frequency f.sub.th, the user's own voice is attenuated. The effect of the face mask on the user's voice (e.g. as received at the ears of the user) may hence be equal to that of a low-pass filter. At frequencies above a 3 dB cut-off frequency of the low-pass filter, e.g. the threshold frequency fun, the user's own voice is attenuated.
[0133] Due to the frequency tilt of the user's own voice, when wearing a face mask, it may be easier to detect own voice, if the user wears a face mask.
[0134] It may be advantageous to focus on frequencies below the threshold frequency f.sub.th, when detecting own voice.
[0135] The user's own voice may be detected using a (trained) neural network.
[0136] FIG. 3 is focused on differences in magnitude (level). Differences in phase may also be used to detect the wearing or non-wearing of a face mask.
[0137] FIG. 4 shows. The own voice processor (OVP) of the hearing aid comprises an own voice detector (OVD) as well as a face mask detector (FMD). FIG. 4 shows an implementation wherein the own voice detector (OVD) and the face mask detector (FMD) are implemented as two different detectors. This may be advantageous as the OVD and the FMD may have different input features. The two detectors may have same, different, or partly overlapping input features. For example, the own voice detector may depend on both acoustic features and vibration-related features, where the face mask detector mainly relies on differences in acoustic features. In the example of FIG. 4. feature F.sub.2 may represent a vibration-related feature, which is only fed to the own voice detector (but not to the face mask detector). In an embodiment the face mask detector (FMD) is only updated when own voice is detected (cf. input OV from the OVD). Both the FMD and the OVD may be implemented by use of a trained neural network (cf. e.g. FIG. 2C, 2D).
[0138] Depending on the detection of a face mask, different actions can be taken. As the acoustic properties change when the user is wearing a face mask (as illustrated in FIG. 3), the frequency shape of the user's own voice may be modified in order to provide a more natural own voice-both for the user and for hands-free telephony.
[0139] An own voice enhancing beamformer may as well take advantage of a face mask detector, as the transfer function between the different microphones may change depending on the face mask. The beamformer may be implemented as an MVDR beamformer relying either on a relative own voice transfer function with or without a face mask. The relative own voice transfer functions may as well be estimated during use—either when the user is talking without face mask or when the user is talking while wearing a face mask.
[0140] In a keyword spotting system where a keyword or a wake word is detected while the user is talking. the presence or absence of a face mask may as well be taking into account e.g. by compensating for the spectral shape of the input signal to a keyword detector such that the spectral properties are similar both for an own voice signal both in presence and in absence of a face mask. Alternatively, training the detector using signals both with and without the person wearing a face mask.
[0141] A face mask detector may as well be used to trigger a reminder. E.g. if the user is not wearing a face mask at places, where is advantageous or required to wear a face mask, the user could be reminded, e.g. via audio feedback played through the hearing device or via a smart phone, smart watch or similar. The reminder could be enabled based on the user's location, e.g. outside the user's home, in public transportation or in shopping areas.
[0142] Wearing a face mask may be an indication that other people as well are wearing a face mask. It may thus be advantageous to adjust the settings of the hearing instruments such that more help is provided in difficult situations (in terms of increased noise reduction or improved speech clarity) as other people wearing a face mask may result in increased mumbling as well as the lack of lip-reading cues.
[0143] FIG. 5 shows an embodiment of an own voice processor (OVP according to the present disclosure. The own voice processor (OVP) comprises a feature extractor (FEX) for extracting features of the electric input signals (x.sub.1, x.sub.2, . . . , x.sub.M) and providing a number NA of acoustic features (F.sub.1, F.sub.2, . . . , F.sub.NA) as described in connection with FIG. 2B. In the example of FIG. 5, the acoustic property focused on is power spectral density (PSD). Values of current power spectral density (denoted PSD(n), where n is a time index) are provided by the feature extractor (FEX). PSD(n) may represent the current power spectral density of a single one of the electric input signals, or of some or all of the electric input signals (x.sub.1, x.sub.2, . . . , x.sub.M), or of a dedicated own voice signal estimate (e.g. the output of an own voice beamformer, cf. e.g. user own voice signal ‘UOV’ in FIG. 8). The own voice processor (OVP) further comprises a memory (MEM) wherein reference data for own voice reception at the input transducers (e.g. microphones) of the hearing device (e.g. hearing aid) are stored (cf. data PSD* in block MEM of FIG. 5). The reference data (PSD*) may e.g. include data as shown in FIG. 3, or equivalent, recorded while the user (or other person, or model) speaks with (PSD*(FMj)) and without (PSD*(OV)) a face mask. The reference data (PSD*) are typically frequency dependent representing acoustic properties (‘acoustic features’) related to the user's own voice. The frequency dependency is indicated by parameters [f.sub.1, f.sub.2, . . . , f.sub.k], where f is a frequency (index) and K is the number of frequencies (e.g. frequency bands) considered. Data PSD*(FMj), j=1, 2, . . . , N.sub.FM), where N.sub.FM is the number of different kinds of face masks considered, represent reference values for N.sub.FM different face masks (e.g. standardized masks or otherwise characterized face masks, optionally ‘home made’ (or other uncharacterized) face masks for which own voice data are available) recorded while the user (or a model of the user) wears the face mask in question while speaking. The reference data (PSD*) may e.g. further or alternatively include difference data ΔPSD*(FMj), j=1.2, . . . , N.sub.FM) representing the acoustic distortion of the different types of face masks, in other words ΔPSD*(FMj)=PSD*(OV)−PSD*(FMj) [dB]. j=1, 2, . . . , N.sub.FM, when using a logarithmic representation of the values.
[0144] The own voice processor (OVP) further comprises a comparator (COMP) for comparing the current value of the acoustic property (PSD(n)) with the stored reference values (PSD*(OV), PSD*(FMj), ΔPSD*(FMj)) and based thereon to provide a degree of similarity of the comparison (cf. signal CMP) to a controller (OVD-FMD-CNT) for providing the own voice control signal(s) No OV, OV and face mask control signals (No FM, FM) as described in connection with FIG. 2A, 2B, 2C, 2D, 4.
[0145] Other acoustic features than the power spectral density used in the example of FIG. 5 may be used in the same principle way.
[0146] FIG. 6 shows a hearing device according to an embodiment of the present disclosure comprising an own voice processor comprising an own voice detector and a face mask detector. FIG. 6 shows an embodiment of a hearing device (HD) comprising an own voice processor (OVP) (comprising an own voice detector (OVD) in combination with a face mask detector (FMD)) and a voice control interface (VCT) according to the present disclosure. The hearing device (HD) of FIG. 6. e.g. a hearing aid or a headset, comprises first and second microphones (Mic1, Mic2) providing respective first and second electric (e.g. digitized) input signals (IN1, IN2) representative of sound in the environment of the hearing device. The hearing device is configured to be worn at or in an ear of a user. The hearing device comprises a forward path comprising the two microphones, first and second analysis filter banks (FB-A1, FB-A2) for converting the first and second (possibly feedback corrected) time domain input signals (IN1. IN2) to first and second frequency sub-band signals (X1, X2), respectively. The frequency sub-band signals of the forward path are indicated by bold line arrows in FIG. 5. The forward path further comprises a beamformer filtering unit (BFU) for providing a spatially filtered signal Y.sub.BF in dependence of the first and second input signals (X1, X2). The beamformer filtering unit (BFU) may e.g. be configured to substantially leave signals from a target direction unattenuated while attenuating signals from other directions, e.g. adaptively attenuating noise sources around the user wearing the hearing device. The forward path further comprises a processor (HAG) for applying one or more processing algorithms to the beamformed signal Y.sub.BF, (or a signal derived therefrom), e.g. a compressive amplification algorithm for applying a frequency and level dependent compression (or amplification) to a signal of the forward path according to a user's needs (e.g. a hearing impairment). The processor (HAG) provides a processed signal (Y.sub.G) to a synthesis filter bank (FB-S) for converting a frequency sub-band signal (Y.sub.G) to a time domain output signal (OUT). The forward path further comprises a loudspeaker (SP) for converting the electric output signal (OUT) to an output sound intended for being propagated to the user's ear drum. The first and second feedback corrected frequency sub-band signals (X.sub.1, X.sub.2) are (in addition to the beamformer filtering unit (BFU)) fed to the own voice detector (OVD) provides an own voice control signal (OV) indicative of whether or not or with what probability the electric input signals comprise speech of the user at a given point in time. The own voice detector (OVD) may e.g. operate on one or more of the first and second (possibly feedback corrected) electric input signals (X.sub.1, X2) and/or on a spatially filtered signal (e.g. from an own voice beamformer, Y.sub.OV). The own voice detector (OVD) may be configured to influence its indication (of OV or not, or p(OV)) by a signal from one or more sensors or detectors. Likewise, the face mask detector (FMD) provides a face mask control signal (FM) indicative of whether or not or with what probability the user wears a face mask at a given point in time. The own voice and face mask control signals (OV, FM) are fed to a keyword detector (KWD) for detecting whether specific words or commands are spoken by the user at a given point in time.
[0147] The keyword detector (KWD is e.g. configured to determine whether or not (or with what probability p(KWx)) the current electric input signals (X.sub.1, X.sub.2) or a signal from an own voice beamformer Y.sub.OV comprise a particular one (KWx) of a number Q (e.g. <20) of predefined keywords or key phrases. In an embodiment, a decision regarding whether or not or with what probability the current electric input signals comprises a particular keyword (or key phrase) AND is spoken by the user of the hearing device is determined as a combination of simultaneous outputs of a KWD-algorithm (e.g. a neural network) and an own voice detector (OVD, e.g. as an AND operation of binary outputs or as a product of probabilities of a probabilistic output).
[0148] The result (e.g. a key word KWx) of the keyword detector (KWD) at a given point in time is fed to a voice control interface (VCT) configured to convert a given detected keyword (or key phrase) to a command (BFctr, Pctr, Xcmd) for controlling a function of the hearing device (HD), e.g. the beamformer filtering unit (BFU, cf. command BFctr), the processor (HAG, cf. command Pctr) and/or of another device or system (cf. command Xcmd forwarded to transceiver Tx/Rx for being transmitted to another device or system). One of the keywords (BFctr) may relate to controlling the beamformer filtering unit (BFU) of the hearing device (HD), e.g. an omni- or DIR mode (e.g. ‘DIR-back’, or ‘DIR-right’, to give a currently preferred direction of the beamformer, other than a default direction, e.g. a look direction), cf. signal BFctr. The same or another one of the keywords may relate to controlling the gain of the processor (HAG) of the hearing device (HD). e.g. ‘VOLUME-down’ or ‘VOLUME-up’ to control a current volume of the hearing device), cf. signal Getr. The same or another one of the keywords may relate to controlling an external device or system, cf. signal Xcmd. Other functions of the hearing device may be influenced via the voice control interface (and/or via the detectors, e.g. the own voice detector), e.g. the feedback control system, e.g. whether an update of filter coefficients should be activated or disabled, and/or whether the adaptation rate of the adaptive algorithm should be changed (e.g. increased or decreased)). A command (Xcmd) may be transmitted to another device or system via appropriate transmitter (Tx) and antenna (ANT) circuitry in the hearing device. Further, in a telephone (or headset) mode, wherein a user's own voice is picked up by a dedicated own-voice beamformer and transmitted to a telephone, and an audio signal (Xaud) is received by appropriate antenna and receiver circuitry (ANT, Rx) from the telephone and presented to the user via an output unit (e.g. a loudspeaker, here SP) of the hearing device (cf. e.g. FIG. 8), may be entered (or left) using a command spoken by the user (e.g. ‘TELEPHONE’ to take (or close) a telephone call). Preferably, the keyword detector of the hearing device is capable of identifying a limited number of keywords to provide voice control of essential features of the hearing device, e.g. program shift, volume control, mode control, etc., based on local processing power (without relying on access to a server or another device in communication with the hearing device). In an embodiment, activation of a ‘personal assistant’ (such as ‘Siri’ of Apple devices or ‘Genie’ of Android based devices or ‘Google Now’ or ‘OK Google’ for Google applications or ‘Alexa’ for Amazon applications) on another device, e.g. a smartphone or similar (e.g. via an API of the other device), may be enabled via the voice control interface of the hearing device. The keyword detector of the hearing device may be configured to detect the wake-word (e.g. ‘Genie’) as one of the keywords. and when it is detected to transmit it (or another command. or the following words or sentences spoken by the user, or a communication partner) to the smartphone (e.g. to an APP, e.g. an APP for controlling the hearing device), from which the personal assistant or a translation service (e.g. initiated by another subsequent keyword, e.g. ‘TRANSLATE’) may thereby be activated. In all cases a valid detection of the user's own voice is of importance. Hence a compensation for any distortion of the user's own voice that might lower the confidence of the own voice control detector from the own voice detector a user's voice is of interest. Such compensation may be provided by the own voice processor (OVP) according to the present disclosure, e.g. by the face mask control signal (FM) indicative of whether or not the user wears a face mask.
[0149] In case a face mask (FM) is detected, a compensation for the change of the input spectrum due to the own voice modified by a face mask may be provided by the hearing device. By compensating for the spectral change due to a face mask, the input feature to the keyword detector (KWD) may be more similar to the own voice without face mask.
[0150] Alternatively, the keyword detector (KWD) may be trained on data recorded with and without a face mask.
[0151] FIGS. 7A and 7B together illustrate an exemplary application scenario of an embodiment of a hearing system (HD1, HD2, AD) according to the present disclosure.
[0152] FIG. 7A shows a hearing system comprising a hearing device (HD1, HD2), e.g. a hearing aid, and an auxiliary device (AD) in communication with each other. FIG. 7A shows an embodiment of a head-worn binaural hearing system comprising left and right hearing devices (HD1, HD2) in communication with each other and with a portable (handheld) auxiliary device (AD) functioning as a user interface (U1) for the binaural hearing aid system (see FIG. 7B). The binaural hearing system may comprise the auxiliary device AD (and the user interface UI). The binaural hearing system may comprise the left and right hearing devices (HD1, HD2) and be connectable to (but not include) the auxiliary device (AD). In the embodiment of FIG. 7A, the hearing devices (HD1, HD2) and the auxiliary device (AD) are configured to establish wireless links (WL-RF) between them. e.g. in the form of digital transmission links according to the Bluetooth standard (e.g. Bluetooth Low Energy, or equivalent technology). The links may alternatively be implemented in any other convenient wireless and/or wired manner, and according to any appropriate modulation type or transmission standard, possibly different for different audio sources.
[0153] The hearing devices (HD1. HD2) are shown in FIG. 7A as devices mounted at the ear (behind the ear) of a user (U). Other styles may be used. e.g. located completely in the ear (e.g. in the ear canal), fully or partly implanted in the head. etc. As indicated in FIG. 7A, each of the hearing devices may comprise a wireless transceiver to establish an interaural wireless link (IA-WL) between the hearing devices. e.g. based on inductive communication or RF communication (e.g. Bluetooth technology). Each of the hearing devices further comprises a transceiver for establishing a wireless link (WL-RF, e.g. based on radiated fields (RF)) to the auxiliary device (AD), at least for receiving and/or transmitting signals. e.g. control signals, e.g. information signals. e.g. including audio signals. The transceivers are indicated by RF-IA-Rx/Tx-1 and RF-IA-Rx/Tx-2 in the right (HD2) and left (HD1) hearing devices, respectively. The remote control-APP may be configured to interact with a single hearing device (instead of with a binaural hearing system, as illustrated in FIG. 7A).
[0154] The auxiliary device (AD) is adapted to run an application program, termed an APP, comprising executable instructions configured to be executed on the auxiliary device (e.g. a smartphone) to implement a user interface for the hearing device (or hearing system). The APP is configured to exchange data with the hearing device(s). FIG. 7B shows the auxiliary device (AD) of FIG. 7A configured to implement a user interface for the hearing device(s) (HD1, HD2) by running an application program from which a mode of operation of the hearing aid can be selected and via which selectable options for the user, and/or current status information can be displayed.
[0155] FIG. 7B illustrates the auxiliary device running an APP for configuring own voice detection features. An exemplary (configuration) screen of the user interface UI of the auxiliary device AD is shown in FIG. 7B. The user interface (UI) comprises a display (e.g. a touch sensitive display) displaying guidance to the user to configure features of the hearing system related to own voice detection. The user interface (UI) is implemented as an APP on the auxiliary device (AD, e.g. a smartphone). The APP is denoted ‘Own voice detection APP’. Via the display of the user interface, the user (U) is instructed to select one or more of ‘Detect face mask’. ‘Activate Voice control’, and ‘Activate telephone mode’. The Voice control interface may be configured via activation of one or more selectable features ‘Change mode’. ‘Change volume’, ‘Change program’. Other features (e.g. ‘Activate wake-word detection for PDA’ to allow detection of a wake-word in the hearing device(s) for a personal digital assistant of the auxiliary device, e.g. a smartphone, e.g. ‘Hey Siri, of an Apple smartphone, or the like) may be added or selectable instead. The activation of a given feature is selected by pressing the ‘button’ in question, which when selected is indicated in bold face and a filled square (.square-solid.) in front of the activated feature(s). In the exemplary ‘Configuration’ screen of the ‘Own voice detection APP’, the features ‘Detect face mask’ and ‘Activated Voice control’(specifically ‘Change volume’) are selected (activated). In the lower part of the screen information to the user of the current status of the bearing device(s) regarding the selected features can be displayed, here a symbol and corresponding text ‘face mask detected’ are provided, thereby the user is informed that the system has detected that the user wears a face mask. In this field of the screen of the user interface, information to the user that he or she should contemplate wearing a face mask in the current environment can be displayed (e.g. in addition to or as an alternative to an acoustic reminder via the output transducer(s) of the hearing device(s)). The current environment may be detected by the hearing device(s) and/or by the auxiliary device (e.g. using acoustic features extracted from the electric input signals of the hearing device(s), and/or GPS functionality of the auxiliary device).
[0156] Further screens (e.g. a ‘Select type of face mask’ screen) of the APP may allow the user to indicate a kind of face mask that the user might wear. The kind of face mask is selectable among a multitude of different types of face masks. The different types of face masks may be characterized in having different acoustic propagation properties of the user's own voice. The hearing device or the auxiliary device may contain a memory wherein such (typically frequency dependent) acoustic properties (‘acoustic features’) of the different types of face masks are stored (cf. e.g. FIG. 5). The APP may be configured to communicate information related the selected face mask (e.g. its kind, e.g. EN14683, N95. KN95, etc., and/or its acoustic properties) to the hearing device(s).
[0157] Switching between different screens of the APP may be achieved via left and right arrows in the bottom of the auxiliary device, or via ‘soft buttons’ integrated in the display of the user interface (UI).
[0158] In the embodiment of FIG. 7A,/B, the auxiliary device (AD) is described as a smartphone. The auxiliary device may, however, be embodied in other portable electronic devices, e.g. an FM-transmitter, a dedicated remote control-device, a smartwatch, a tablet computer, etc. FIG. 8 shows an embodiment of a headset or a hearing aid comprising own voice estimation and the option of transmitting the own voice estimate to another device, and to receive sound from another device for presentation to the user via a loudspeaker, e.g. mixed with sound from the environment of the user according to the present disclosure. The hearing device (HD) comprises two microphones (M1, M2) to provide electric input signals IN1, IN2 representing sound in the environment of a user wearing the hearing device. The hearing device further comprises spatial filters DIR and Own Voice DIR, each providing a spatially filtered signal (ENV and OV respectively) based on the electric input signals (IN1, IN2). The spatial filter DIR may e.g. implement a target maintaining, noise cancelling, beamformer. The spatial filter Own Voice DIR implements spatial filter configured to pick up the user's own voice. The spatial filter Own Voice DIR implements an own voice beamformer directed at the mouth of the user. The activation and control of the Own Voice DIR is controlled by an own voice processor (OVP) according to the present disclosure. The own voice processor provides control signals (OV, FM) indicative of the presence of the user's own voice (OV) and of whether the user wears a face mask (FM), respectively. In a specific telephone mode of operation. the user's own voice is picked up by the microphones M1, M2 and spatially filtered by the own voice beamformer of spatial filter ‘Own Voice DIR’ providing an estimate of the user's own voice (signal UOV). The signal UOV may be used by the own voice processor as inputs to determine the own voice and/or face mask control signals (OV, FM) as indicated by dashed arrow from the ‘own Voice DIR-’ to the ‘OVP’-block. The bearing device further comprise an own voice signal processor (OV-PRO) configured to improve the estimate of the user's own voice and provide a modified own voice signal (UOVOUT) in dependence of the face mask control signal (FM). The own voice signal processor may be configured to modify the frequency shape of the user's own voice in dependence of the face mask control signal (FM). Thereby the frequency shaping of the user's own voice performed by the face mask can be compensated for. The modified (improved) own voice signal (UOVOUT) is fed to transmitter Tx and transmitted (by cable or wireless link to another device or system (e.g. a telephone. cf dashed arrow denoted ‘To phone’ and telephone symbol). In the specific telephone mode of operation. signal PHIN may be received by (wired or wireless) receiver Rx from another device or system (e.g. a telephone, as indicated by telephone symbol and dashed arrow denoted ‘From Phone’). When a far-end talker is active, signal PHIN contains speech from the far-end talker, e.g. transmitted via a telephone line (e.g. fully or partially wirelessly, but typically at least partially cable-borne). The ‘far-end’ telephone signal PHIN may be selected or mixed with the environment signal ENV from the spatial filter DIR in a combination unit (here selector/mixer SEL-MIX), and the selected or mixed signal PHENV is fed to output transducer SPK (e.g. a loudspeaker or a vibrator of a bone conduction hearing device) for presentation to the user as sound. Optionally, as shown in FIG. 8, the selected or mixed signal PHENV may be fed to processor PRO for applying one or more processing algorithms to the selected or mixed signal PHENV to provide processed signal OUT, which is then fed to the output transducer SPK. The embodiment of FIG. 8 may represent a headset, in which case the received signal PHIN may be selected for presentation to the user without mixing with an environment signal. The embodiment of FIG. 8 may represent a hearing aid, in which case the received signal PHIN may be mixed with an environment signal before presentation to the user (to allow a user to maintain a sensation of the surrounding environment; the same may of course be relevant for a headset application, depending on the use-case). Further, in a hearing aid, the processor (PRO) may be configured to compensate for a hearing impairment of the user of the hearing device (hearing aid).
[0159] It is intended that the structural features of the devices described above, either in the detailed description and/or in the claims, may be combined with steps of the method, when appropriately substituted by a corresponding process.
[0160] As used, the singular forms “a.” “an.” and “the” are intended to include the plural forms as well (i.e. to have the meaning “at least one”), unless expressly stated otherwise. It will be further understood that the terms “includes,” “comprises,” “including,” and/or “comprising,” when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components. and/or groups thereof. It will also be understood that when an element is referred to as being “connected” or “coupled” to another element, it can be directly connected or coupled to the other element but an intervening element may also be present, unless expressly stated otherwise. Furthermore, “connected” or “coupled” as used herein may include wirelessly connected or coupled. As used herein, the term “and/or” includes any and all combinations of one or more of the associated listed items. The steps of any disclosed method is not limited to the exact order stated herein, unless expressly stated otherwise.
[0161] It should be appreciated that reference throughout this specification to “one embodiment or “an embodiment” or “an aspect” or features included as “may” means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the disclosure. Furthermore, the particular features, structures or characteristics may be combined as suitable in one or more embodiments of the disclosure.
[0162] The previous description is provided to enable any person skilled in the art to practice the various aspects described herein. Various modifications to these aspects will be readily apparent to those skilled in the an, and the generic principles defined herein may be applied to other aspects.
[0163] The claims are not intended to be limited to the aspects shown herein but are to be accorded the full scope consistent with the language of the claims, wherein reference to an element in the singular is not intended to mean “one and only one” unless specifically so stated, but rather “one or more.” Unless specifically stated otherwise, the term “some” refers to one or more.
REFERENCES
[0164] P-2019-009EP, when published>Oct. 17, 2020 [0165] EP3709115A1 (Oticon) 16.09.2020 [0166] EP3588981A1 (Oticon) 01.01.2020
