WO2018154143A1

WO2018154143A1 - Measurement-based adjusting of a device such as a hearing aid or a cochlear implant

Info

Publication number: WO2018154143A1
Application number: PCT/EP2018/054830
Authority: WO
Inventors: Michael BOEDTS
Original assignee: Tympres Bvba
Priority date: 2017-02-27
Filing date: 2018-02-27
Publication date: 2018-08-30

Abstract

The current invention relates to a method for adjusting an acoustic device, said acoustic device adapted to be worn by a human listener in an ear canal, at an ear or in the ear of the listener and suitable for being connected to a signal processing circuit adapted to process one or more input audio signals to provide a processed audio signal suitable for transmission by said acoustic device as an audible signal to the listener; said method comprising a plurality of steps involving a transfer function measurement associated with a tympanic membrane belonging to said ear of said listener, and a correlation value indicative of an intentional focus.

Description

MEASUREMENT-BASED ADJUSTING OF A DEVICE SUCH AS A

HEARING AID OR A COCHLEAR IMPLANT

Technical field

The invention pertains to the technical field of acoustic devices. More specifically, the invention provides a method and system/device for adjusting acoustic devices based on transfer function measurements of the tympanic membrane for capturing auditory attention.

Background

A human ear is comprised of an outer ear, a middle ear and an inner ear. The outer ear comprises an ear canal, the middle ear comprises an eardrum (or tympanic membrane), and the inner ear comprises a cochlea. As for acoustic devices facilitating hearing, the outer ear relates to hearing aids. Hearing aids have been widely used to compensate hearing losses of human ears. Depending on individual needs, people often use different types of hearing aids. The types of hearing aids include in-ear aids, behind-ear aids, and canal aids. The methods known in the state of the art largely focus on the properties of the incoming pressure wave (i.e. the sound) to adjust the hearing aids. Middle ear implants and cochlear implants, on the other hand, relate to the middle ear and the inner ear, respectively.

To focus on a specific sound source in situations where this sound source is heard between competing sound sources, the so-called cocktail party problem, the human hearing system is faced with sound that reaches the tympanic membrane as a single integral acoustic input, comprising content coming from several sound sources. In the case of monaural hearing, this acoustic input is to be processed so that one or more auditory streams of interest can be attended to at the expense of other auditory streams. In the case of binaural hearing, two acoustic inputs corresponding to two tympanic membranes are to be processed. Hereby, sound of all auditory streams is processed up to some degree, preferentially foregrounding sound originating from an auditory stream of interest.

In the processing, a prime concern is the way in which an acoustic input is processed. As humans, we use a diverse approach taking into account physical properties of the sound waves (bottom-up) combined with an intentional focus to obtain information from one or more sources of interest, thereby attempting to filter out all but the auditory streams of interest. Hereby, the top-down process of selecting acoustic info from one sound source at the expense of information from other sources can be guided by external elements or by a centrally mediated process. The former relies on a bottom-up approach, whereas the latter concerns a top-down approach.

To date, methods and devices for processing an acoustic input comprising different auditory streams are typically limited by their strict reliance on the physical properties of the sound waves, according to a purely bottom-up approach. The few methods allowing to preferentially select one or more specific sound sources in a top-down fashion are crude and impractical. US 8,199,945 describes a method with use of directional microphones, which is however complex in design as well as operation.

US 7,715,577 describes how in state-of-the-art hearing devices the acoustic reflectance is determined as a function of the sound frequency, whereby the magnitude, phase and latency of the acoustic reflectance are defined. Furthermore, US 7,715,577 provides a method and system for automatically adjusting acoustic devices based on acoustic reflectance. However, also the method according to US 7,715,577 lacks a top-down approach, not allowing to adjust the hearing aid according to the needs of a specific listener. Furthermore, US 7,715,577 does not provide a solution for binaural hearing, despite its importance and potential for practical applications.

EP 2560 412 discloses a related method but disregards acoustic reflectance.

EP 3 139 638, US 2012/076313, US 201303941 and WO 2009/023633 relate to acoustic processing but are not directed at the cocktail party problem.

There remains a need in the art for improved methods and devices to process sound according to the intentional focus of the listener, which do not necessitate the use of multiple directional microphones.

The present invention aims to resolve at least some of the problems mentioned above. Summary of the invention

The present invention relates to adjusting of a device such as a hearing aid, a middle ear implant, a cochlear implant or any device that transmits acoustic information toward the brain of the listener. Hereby the aim is to process sound according to the intentional focus of the listener. Alternatively or in addition to using multiple directional microphones, the method bases itself on transfer function measurements with respect to at least a tympanic membrane, taking into account at least the amplitude and/or a phase-related measurement such as the group delay or the phase. In a first aspect, the invention provides a method according to claim 1.

Hereby, the first tensioning and the second tensioning of the tympanic membrane relate to two different tension states of the tympanic membrane. Hereby, each tensioning, or, equivalently, tensioning state, is characteristic of the attention of the listener to steer his or her tympanic membrane while processing the audible signal. Given the complexity of the TRRS, low intentional focus cannot simply be mapped to low tensioning of the membrane, and likewise, high intentional focus cannot simply be mapped to high tensioning of the membrane. Rather, each tensioning relates to a three-dimensional surface and a two-dimensional tension function as defined by the tympanic membrane in its current state. The transfer function measurement cannot sample this directly, but rather uses the phase of the reflected acoustic energy, and preferably also the amplitude of the reflected acoustic energy, as a characteristic "fingerprint" of a given tensioning state of the tympanic membrane. The tensioning state hereby is indicative of spectral and/or spatial features of the incoming sound provided by the audible signal. This should not be confused with a basic approach whereby only simple amplitude values indicative of the pressure on the tympanic membrane would be measured, lacking phase-related measurement, which would not be suitable to discriminate phenomena relating to the cocktail party effect. Furthermore, note that in view of the important role of the intentional focus of the listener in the method, said at least two pre-determined profiles relate to a corresponding calibration method with one or more calibration sessions, wherein a second listener preferably equaling the listener is asked to direct its attention according to attention-related instructions.

The advantage of said method is the measurement-based adjusting of said device, directed at detection of cocktail-party-related changes to the tensioning of the tympanic membrane. In contrast to prior art methods, existing profiles relating to auditory attention are uniquely combined with transfer function measurements comprising a phase-related measurement. While amplitude measurement is a primary indicator of frequency bands calling for attention; it are phase-related measurements such as group delay measurement and phase measurement that are crucially indicative of several aspects of the attention of the listener, such as the attention level of the listener (see "Detailed description" below). By performing phase-related measurements, the invention is much better adapted to quantify aspects of the attention of the listener than is possible with prior art devices and methods. To characterize the attention level, the invention advantageously incorporates at least two profiles that serve as a reference. Opposed to a method according to the state of the art, said measurements are not directed toward external sound as it arrives at the tympanic membrane of a listener, but rather to the reflected pressure or conducted pressure of said external sound via the tympanic membrane of the listener.

In a preferred embodiment, said at least two pre-determined profiles comprise said at least one profile associated with said first tensioning of the membrane and low intentional focus with respect to a first auditory stream and said at least one profile associated with said second tensioning of the membrane and high intentional focus with respect to said first auditory stream; said at least one profile associated with said first tensioning of the membrane preferably further associated with high intentional focus with respect to said second auditory stream; and said second tensioning of the membrane preferably further associated with low intentional focus with respect to said first auditory stream. Such an embodiment is advantageous in that it allows to capture further aspects of the listener's attention relating to the cocktail party effect, whereby the listener actively steers the tensioning so that said first auditory stream can be attended to at the expense of other auditory information present in the audible signal. In an alternative embodiment, said high intentional focus and said low intentional focus are not related to any specific auditory stream, and are merely indicative of the listener paying attention or not paying attention.

In a related preferred embodiment, said at least two pre-determined profiles comprise said at least one profile associated with said first tensioning of the membrane and said low intentional focus and at least two profiles associated with a third tensioning of the membrane and high intentional focus to a first auditory stream and a fourth tensioning of the membrane and high intentional focus to a second auditory stream different from said first auditory stream; said at least one profile associated with said first tensioning of the membrane preferably further associated with low intentional focus with respect to both said first and said second auditory stream. Such an embodiment is advantageous since it allows to discriminate between the listener's attention being directed to a first auditory stream, to a second auditory stream, or to none of both streams and/or to none of the auditory information present in the audible signal. This advantageously applies phase-related measurements to reflected acoustic energy to reveal the cocktail-party-effect-relating information present in the tensioning state of the tympanic membrane.

In a preferred embodiment, said step of repeatedly altering said processing comprises mixing a test audio signal, preferably provided by a test audio signal generator comprised in said signal processing circuit, into said processed audio signal for enabling determining of said intentional focus level of said tympanic membrane, said mixing of said test audio signal preferably triggered by a schedule and/or by a detection of an event present in said one or more input signals. In an alternative embodiment, no test audio signal is applied, and the processing is performed purely on the natural external sound as provided by said one or more input audio signals.

An advantage of such a method is that playing back a base signal, or, equivalently, a test audio signal, at sufficient power level ensures that the transfer function measurements can be measured with sufficient precision. Hereby, in a preferred embodiment, the base signal may be any typical signal used in listening tests such as LiSN-S, LiSN-S PGA or similar, and may, e.g., concern chirps, interrupted or continuous sine waves, or noise. In a more preferred embodiment, the test audio signal comprises human speech, i.e. the test audio signal relates to a recording or a synthetically generated sample of one or more words spoken out loud by a human or human-like voice. Hereby, the obtained measurements are sufficiently accurate in view of the sufficient power level of the test audio signal, and comprise useful information regarding the current environment of the listener and the influence of the current environment on the listener, since they comprise a registration of the conditioning of the TRRS as it was attained just before the test audio signal was provided, e.g. at the end of the operational period preceding/during the test period concerned. Additionally, the measurements may be impacted by the characteristics of the ear canal of the listener. However, if the base signal is known, such characteristics may be filtered out of the measurements, allowing a more accurate characterization of the conditioning of the TRRS. Preferably, the playback of the base signal takes place just after a part of an operational period that is non-overlapping with a test period in the time and/or frequency domain. In this way, again, the measurements of the TRRS conditioning primarily reflect the response to the listener's current environment, with limited impact from the transient effects caused during the adjusting of the device. Alternatively, the effects may be aimed at frequency ranges that do not include the range of the base signal.

Another advantage of such a method is that a standard base signal may be used for which listening test results are available before actual operation. Hence, the device can be adjusted relative to those listening test results, allowing higher accuracy in the measurement of the conditioning of the TRRS. This is opposed to a method according to the state of the art, for which adjusting of the device during operation is done solely on the basis of said bottom-up approach, e.g . with simple amplification of sound from the current environment of the listener. Due to this limitation, and the fact that such sound is variable and unpredictable, a method according to the state of the art is prone to errors in the adjustment process. In a second aspect, the invention offers a signal processing circuit according to claim 14.

In a third aspect, the invention provides an acoustic device according to claim 15.

In a preferred embodiment, said device is a hearing aid, a middle ear implant, a cochlear implant, a bone-anchored hearing aid and/or a bone-conducting microphone. In another preferred embodiment, said device is configured to carry out a method such as the method according to the present invention.

According to another aspect, the invention offers a use according to claim 18.

Further preferred embodiments are discussed in the detailed description and the claims.

Description of figures

Figure 1 shows a first example embodiment of a device according to the present invention.

Figure 2 shows a second example embodiment of a device according to the present invention.

Figure 3 shows a third example embodiment of a device according to the present invention.

Figure 4 shows a fourth example embodiment of a device according to the present invention. Figure 5 illustrates example aspects of acoustic processing according to the present invention.

Detailed description

As this invention pertains to the adjusting of an acoustic device, it is clear to the person skilled in the art that one or more steps of the method according to the present invention may be executed by the signal processing circuit or the device according to the present invention. Likewise, the device and the signal processing circuit may be configured to execute steps of the method. Hence, in this document, preferred embodiments may apply to any aspect of the invention.

As mentioned, in view of the important role of the intentional focus of the listener in the method, said at least two pre-determined profiles relate to a corresponding calibration method with one or more calibration sessions, wherein a second listener preferably equaling the listener is asked to direct its attention according to attention- related instructions. Hence, according to a further aspect of the invention, the invention relates to a calibration method for determining said at least two predetermined profiles, said at least two pre-determined profiles preferably adapted for use in the method according to claims 1-13 and/ in the signal processing circuit according to claim 14 and/or the acoustic device according to claim 15. Hereby, said calibration method may relate to hearing tests and listening tests used when adjusting hearing aids, which are subjective tests: listeners hear a tone, a word or sentence or text, and have to repeat or write down or indicate what they hear. The most commonly used are audiometry, speech audiometry, speech audiometry in noise, etc. These are used in adjusting the hearing aid to estimate the hearing loss. Lis-n test is an example of a subjective test.

Such tests may for instance allow listeners to listen to a text or tone, whereby the understanding of the listener is of secondary importance. Rather, the action performed by the listener by means of his eardrum is of prime importance. The main aim of the test is the measurement of the transfer function, during different conditions of auditory attention.

Measuring auditory attention objectively is known when relating to EEG measurements, but not when adjusting or adjusting hearing aids. EEG measurements, like the transfer function measurements, are indicative of the degree of auditory attention. This should be distinguished from related pupil size measurements, which are mainly indicative of the effort the person spends to create auditory attention. The present method measures the degree (intensity) and direction of auditory attention.

In one example embodiment of the calibration method, the listener has to relax, or listen to a word or text fragment (e.g. words by 1 speaker, e.g. a mix of 2 voices) or to a composite tone, or a mix of composite tones. There may be an odd-ball system in which the listener must notice a change or exception. Or he may be asked to write down or indicate what he understands, etc.

In an example embodiment of the calibration method, the test is set up in such a way that listeners sometimes have no auditory attention (e.g. by pointing their attention at something else, e.g. by having them read a text, letting them do a mathematical exercise without writing, counting from back to front, or to have them perform a task that requires visual attention); or may be attentive.

In an example embodiment of the calibration method, if auditory attention is "ON", the test may e.g. offer two composite tones / voices / text fragments / ... that differ only in spectral area. When the person has to zoom in on one of these two, it may be deducted that he does this on the basis of spectral differences. The example test may subsequently offer two composite tones / voices / ... that differ from each other only because a different delay between left and right ear is set for those two. If the listener then zooms in on one of those two, it may be deducted that he does so on the basis of spatial characteristics; and then only on that part of spatial distinction that is based on the time difference between left and right ear, not on that part that is based on intensity difference. In a preferred embodiment, another option is to ask the test subject to NOT actively listen to voices, sounds, etc. During all these conditions transfer function measurements are preferably performed for both ears.

In a preferred embodiment, said transfer function measurement indicating low and/or high intentional focus may trigger a feedback signal for perceptual feedback to said listener, to enable a neurofeedback/biofeedback functionality. Since a persistently active auditory attention is indicative of alertness, and therefore may be indicative of stress or anxiety, the invention may advantageously be used as a biofeedback means for use in yoga, meditation, or in general daily life, to relax. In a preferred embodiment, the intentional focus being low relates to a feedback signal comprising a first indication and the intentional focus being high relates to a feedback signal comprising a second indication. In a more preferred embodiment, a plurality of gradual indications between high and low is provided. The signal may relate to any visual indication, such as a colored led whereby the first indication relates to a first color and the second indication relates to a second color. Alternatively or additionally, the signal may relate to any acoustic indication, such as a beep or jingle whereby the first indication relates to a low frequency and the second indication relates to a high frequency. This advantageously allows the listener to try to attain the first indication as often/much as possible, by avoiding focusing his ears as much as possible and coming to rest. Moreover, in a preferred embodiment, the relative share of the first indication over a certain period of time, e.g. a day, may be notified to the listener, e.g. by means of a display, allowing the listener to verify whether certain aims with respect to e.g. relaxing have been achieved during the given time period.

In a preferred embodiment, said signal processing circuit is comprised in said device. Such an embodiment is advantageous since it leads to a compact and user-friendly device. In an alternative embodiment, the signal processing circuit may be provided in the form of some external module integrated in an external device, such as an external controller, a mobile device, a smartphone, a smartwatch, a tablet, a laptop, etc. In such cases, the hearing device is connected to the external module by means of conducting wire and/or by means of a digital wireless communication protocol such as Bluetooth, Zigbee, NFC or IEEE 802.11 (WiFi). Such an embodiment may be particularly advantageous where the device belongs to e.g. a binaural headphone comprising two hearing devices, one per ear, which are both connected with a wire to a single controller outside of said two hearing devices but comprised in said binaural headphone.

In a preferred embodiment, said method comprises the further step of:

- evaluating at least one of said correlation values with respect to at least one pre-determined correlation threshold, for detecting whether a predetermined focus-related criterion is met; preferably for detecting whether said intentional focus level changes from high to low or from low to high;

- if said criterion is met, perform an action relating to said signal processing circuit, preferably entering a passive processing mode if said intentional focus level changes from high to low and/or entering an active processing mode if said intentional focus level changes from low to high; alternatively entering an active processing mode if said intentional focus level changes from high to low and/or entering a passive processing mode if said intentional focus level changes from low to high.

In a preferred embodiment, said active processing mode relates to an increased difference between the processed audio signal and the one or more input signals that are processed, whereas the passive processing mode may relate to limited difference or even no difference. This is advantageous since it allows a selective processing rather than a permanent processing, which may be more suitable in practical use and hence more user-friendly. Alternatively or in addition hereto, active processing mode may relate to a mode wherein the processing is performed at full complexity and hence corresponds to full power requirements, while passive processing mode may relate to lowered power consumption. This is advantageous since, in an embodiment where the device is at least partially battery-powered, it may allow an increased battery life, whereby electric power is consumed selectively depending on the needs.

In another preferred embodiment, said method comprises the further steps of: - evaluating an increment of said correlation of at least one of said correlation values with respect to a previous correlation value for detecting an evolution of said correlation value;

- adapting a control parameter relating to said signal processing circuit (7, 68, 57, 58, 19) in function of said correlation increment, preferably manipulating a control parameter relating to said processing of said one or more input audio signals.

Such an embodiment since it allows incremental adaptation and adjustment of the hearing device. In another preferred embodiment, said determining of said correlation value with respect to each of said at least at least two pre-determined profiles essentially relates to a measurement frequency band with an upper measurement frequency, said upper measurement frequency smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz. Such an embodiment surprisingly provides for reliable measurements, mainly by relying on subharmonics of the reflected signal. Apparently, the tensioning of the tympanic membrane may advantageously be detected at such low frequencies. Furthermore, such detecting at low frequencies may be advantageously combined with a further preferred embodiment with a test sample audio signal, wherein said test sample audio signal essentially relates to a steering frequency band; wherein said determining of said correlation value with respect to each of said at least at least two pre-determined profiles essentially relates to a measurement frequency band with an upper measurement frequency B, said upper measurement frequency B preferably smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz; wherein at least 50% of said steering frequency band is situated above said upper measurement frequency B, preferably at least 80% of said steering frequency band is situated above said upper measurement frequency B. Such an embodiment advantageously utilizes the detectability of the tensioning of the tympanic membrane at low frequencies with the providing of a test audio signal of higher frequency. Such an embodiment advantageously prevents problems relating to undesirable acoustic feedback loops between the "steering" of the tympanic membrane and the measurement of the tensioning of the membrane.

In another preferred embodiment, at least one of said at least two pre-determined profiles, preferably each of said at least two pre-determined profiles, is based on a prior transfer function measurement associated with a prior-measurement-related tympanic membrane belonging to an ear of a prior-measurement-related listener, said prior-measurement-related tympanic membrane preferably being said tympanic membrane and said prior-measurement-related listener preferably being said listener; wherein said prior transfer function measurement relates to a listening test comprising listening to a human or human-like voice. Hereby, a human-like voice may refer to any synthetically generated voice that is aimed at imitating or resembling a human voice. Such an embodiment advantageously uses the sensitivity of the human ear to human voices, preferably with fragments of human speech. Since a typical use of the hearing device according to the present invention includes the processing of the human voice, such an embodiment also effectively targets improved performance for such application. In another preferred embodiment, said method is applied to a first tympanic membrane belonging to a first listener and a second tympanic membrane belonging to said first listener; wherein said transfer function measurement comprises a first measurement associated with said first tympanic membrane and a second measurement associated with said second tympanic membrane and essentially concurrent with said first measurement, and wherein said repeatedly altering of said processing is performed dependent on correlation values based on each of said first and said second measurement. Such an embodiment is advantageous because of the phase-related inter-ear differences related to binaural hearing, whereby particularly the intentionally focused streams stand out, as discussed further in this document. In a further preferred embodiment, said repeatedly determining said correlation value comprises calculating a difference between said first measurement and said second measurement, and wherein said correlation value relates to a binaural correlation of said difference with respect to at least two pre-determined profiles being binaural profiles, said correlation value indicative of an intentional focus level of said first and said second tympanic membrane; wherein said at least two pre-determined binaural profiles comprise at least one binaural profile associated with a first tensioning of the first and the second membrane and low intentional focus and at least one binaural profile associated with a second tensioning of the first and the second membrane and high intentional focus. Such an embodiment is particularly advantageous in that it allows quantifying the inter-ear phase differences indicative of intentional focus related to, e.g., certain frequency bands.

As mentioned in the summary section, opposed to a method according to the state of the art, said measurements are not directed toward external sound as it arrives at the tympanic membrane of a listener, but rather to the reflected pressure or conducted pressure of said external sound via the tympanic membrane of the listener. Hereby, the tympanic membrane, and in fact the entire tympanic resonance regulating system (TRRS) of the listener is conditioned by the engaging of the listener in the interpretation of the sound (and related incident pressure). This conditioning is realized in part by slight deformations of the tympanic membrane relating to variations of the membrane's stiffness, but also other parts of the tympanic resonance regulating system are involved (see "Detailed description" below). A typical example wherein this conditioning is crucial is the cocktail party effect. As a result, various characteristics of the transfer function measurement among which the group delay comprise a registration of said conditioning. Therefore, by measuring the transfer function, preferably during test periods, and adjusting the device accordingly, improved hearing is possible during operational periods as a result of the adjustment. Opposed to a method according to the state of the art, said measuring is not done entirely before the operational period as a form of "initial calibration". Such initial calibration allows adjusting the device to certain properties of the hearing system of a specific listener as well as certain capabilities of the specific listener, but cannot take into account the current environment of the listener, and, more importantly, the intentional focus of said listener. In a preferred embodiment according to the present invention, measurements during test periods are alternated with operational periods, to allow adjusting the device to the current environment of the listener. This alternation allows for real-time adjusting of the device during use, by first measuring the immediate changes in the conditioning of the TRRS, to then adjust the way in which sound (and related incident pressure) is processed. In a preferred embodiment, said adjusting is performed based on the information gained from the combination of transfer function measurements as well as listening tests, preferably spectral and/or spatial listening tests, all relating to said specific tympanic membrane of said listener.

A surprising additional advantage of the present invention is that it also allows listeners to actively adjust the device. Since the transfer function associated with the TRRS is attention-related, a listener may deliberately shift his/her attention to trigger the adjusting of the device. Indeed, as the transfer function measurements are function of the attention given by the listener, he/she may undertake actions such as the deliberate variation of the tympanic membrane's stiffness in order to achieve an associated desired adjusting of the device. This type of deliberate adjusting is enabled by the schedule of operational and testing periods. In a preferred embodiment, the predictability of such a schedule may help a listener to appropriately time the deliberate actions undertaken for adjusting the device. According to an aspect of the present invention, said determining in step (a) comprises detecting whether a pre- defined pattern is present in said transfer function measurement, the presence of said pre-defined pattern being manipulable by the listener, wherein said adjusting in step (b) takes into account whether said pre-defined pattern was detected. Hereby, the pre-defined pattern concerns a deliberate manipulation of the TRRS by the listener according to sequence of one or more steps that are pre-defined. In a preferred embodiment, the pre-defined pattern is associated with a certain basic function that is part of the adjusting, such as turning on or off background noise filters, increasing/decreasing of background noise filters, toggling between omnidirectional processing and directional processing, the tuning of the device between omnidirectional processing and directional processing, the adjusting of volume, etc. This has the advantage that it enhances the user's convenience, not having to press buttons on the device for certain basic functions. Hereby, the listener may actively train the pre-defined pattern, e.g. by means of listening tests, which may include active training of the execution of such pre-defined pattern. In a further preferred embodiment, the pre-defined pattern comprises a short and easily-explainable action, such as rapidly tensioning the tympanic membrane a pre-defined number of times, e.g., two or three times. Such a pattern can be trained beforehand and can be put to practice by a listener whenever it is his/her intention to adjust the device deliberately. In an alternative embodiment, the pre-defined pattern comprises a characterization of the shape of the transfer function curve, such as a flat curve, ripple, swell and crest, as explained below. This is advantageous since it is known from listening tests that listeners have the ability to steer the shape of the curve. In this document, the terms "test audio signal" and "base signal" are used interchangeably. Also the terms "listener" and "user" are interchangeable.

Furthermore, as mentioned, in a preferred embodiment, said schedule comprises at least one operational period which is at least partially non-overlapping with any test period in the time domain and/or the frequency domain. Correspondingly, in the use of the device periods may potentially be identified during which no testing takes place and no adjusting is performed. This ensures that a sufficiently stable conditioning of the TRRS can take place, whereby the conditioning of the TRRS and the measurements thereof are shaped primarily in response to the listener's current environment, with limited impact from the transient effects caused during the adjusting of the device. This is particularly useful in case of hearing aids where all measurement and playback takes place in the outer ear, and adjusted sound may be fed back into the measurements, causing an undesired feedback loop. Such a feedback loop may easily be prevented, however, by means of a short non- operational period relating to a test period. In a preferred embodiment relating for instance to hearing aids, the alteration of operational and non-operational periods may be such that adjustments only impact the start of each word or each distinct sound, e.g. typically for less than a second, preferably less than 100 ms, more preferably less than 10 ms. The remaining portion of each word/sound may then be reproduced with less or no adjustment. Such a scheme is useful since it is known that the listener is able to extract a lot of information from the very first milliseconds of each new sound, which may again reflect in deformations of the tympanic membrane and other changes in the TRRS, which result in changes in the transfer function measurements. A further advantage of such a scheme is that a listener is trained to a certain degree to still "exercise" his/her TRRS. In a related alternative embodiment, said undesired feedback loop is prevented by remaining operational but rather avoid adjusting continuously. In an alternative embodiment relating to a middle ear implant and/or a cochlear implant, said undesired feedback loop does not occur since energy is carried over to the middle ear and inner ear without a sound source in the outer ear, with therefore no risk of interference. Nevertheless, it may be useful also in such an embodiment to alternate between periods of adjusting and plain testing without adjusting, in order to allow the listener to adapt to new incoming sounds without interference by transient effects caused during the adjusting of the device.

In a preferred embodiment of said method, following steps take place during said test period :

- playing back a base signal in direct proximity of said tympanic membrane;

- determining an unprocessed transfer function measurement associated with said tympanic membrane and said base signal;

- processing said unprocessed transfer function measurement taking into account at least said base signal to obtain said transfer function measurement; wherein said transfer function measurement varies with the incident pressure.

In the context of the present invention, a "transfer function" is identified and is associated with a TRRS through at least one tympanic membrane. Hereby, the transfer function describes the relation of acoustic output signals to acoustic input signals. In a preferred embodiment, the transfer function comprises a set of transfer curves, each curve providing the ratio of a parameter at the output and the same parameter at the input, as a function of frequency or time, while all other parameters are kept constant. In its turn, each curve may comprise one or more components, e.g. individual scalar or complex measurement values. Key parameters in the present invention relate to amplitude, phase, delay, frequency and time. Hereby, a phase- related measurement may for instance relate to the group delay calculated for some given frequency band, or to the phase difference of two signals, e.g. the phase difference between an input audio signal and a reflected audio signal in a certain frequency band with respect to the average group delay and/or phase over the full frequency spectrum.

In a preferred embodiment, the transfer function measurements comprise reflectance components comprising an acoustic phase as well as an acoustic amplitude. Hereby, the association of the transfer function to a tympanic membrane is not to be interpreted as limiting the invention in any way. Rather, the transfer function relates to the entire TRRS associated with a single ear, and while the tympanic membrane is obviously a key part of the TRRS, also the other parts of the TRRS are assumed to contribute to each measurement of the transfer function. Furthermore, note that a transfer function inherently varies both with the acoustic input and the attention given to certain aspects by the listener, whereby the attention in itself relates to e.g. the familiarity of a given sound present in the acoustic input.

The present invention concerns a method for adjusting a device such as a hearing aid, a middle ear implant, a cochlear implant, a bone-anchored hearing aid or a bone- conducting microphone to process sound according to the intentional focus of the listener. Hereby, the intentional focus of the listener is reflected in the conditioning of the listener's Tympanic Resonance Regulating System (TRRS) in its attempt to focus on a desired aspect of the incoming sound. The basis for this conditioning is the act of "straining one's ears", according to the auditory attention of the listener. Hereby, "auditory attention" encompasses the process leading to selection of certain auditory information sources over others. This involves not only "attracting" information from one or more sound sources, but also "repelling" information from other sound sources. Attention may relate to the content type, but may also involve spatial aspects and spectral aspects. Hereby, spatial aspects relate to different directions in space, whereas spectral aspects relate to different frequency bands and/or different characteristics in the frequency domain.

In its operation, the TRRS makes use of two antagonistic sets of muscles, innervated by two nerves. These sets of muscles have a different effect on the stiffness of the tympanic membrane, thereby modifying the group delay and phase delay of the tympanic membrane for certain frequency bands. This system enables one to "strain one's ears". By adding delay to certain frequency bands of the sound waves entering the auditory pathway via the tympanic membrane, one can change the phase or time delay of the related frequencies of the sound waves that reach the cochlea; and by adding more or less delay in one of both ears, one can change the difference between the phase or time delay of sound waves in one ear and the phase or time delay in the other ear. One can also vary, in a random or regular fashion, the group delay for certain frequencies or frequency ranges over time. Moreover, changing the stiffness of the tympanic membrane allows one to cause or prevent standing waves on the tympanic membrane, which again will influence group delay and phase. Similarly, one can influence resonances in the middle ear, which influences group delay distortions and phase distortions for certain frequencies or frequency ranges. The modulations of the group delay of the tympanic membrane by the TRRS modify the characteristics of the information transmitted to the central auditory pathways, making possible certain facilities of central processing of sounds. It allows one to "attract" or "favor" some sounds and "repulse" other sounds, depending on spatial or spectral properties of these sounds. An example is the cocktail party effect. Another example is listening in reverberation spaces, relating to the "de-reverberating" capability of the TRRS.

Actions performed by the TRRS involve influencing the absolute values and variations over time of the group delay and/or phase of both tympanic membranes. These actions produce often different results for both tympanic membranes, in a way that is coordinated so that the difference between both tympanic membranes provides extra possibilities for central processing used in hearing and auditory attention.

One way the TRRS may possibly help in directing and modulating auditory attention is by influencing the properties of the critical bands. It is well-known that the cochlea acts as a filter bank, containing a number of "critical bands". These critical bands may shift on the frequency scale, depending on the auditory input. Specifically, an audible tone with a specific frequency will create a critical band around its frequency, masking the frequencies around this center frequency. This filter bank mechanism is modelled in e.g. gamma tone or gamma chirp filters, and is taken advantage of in e.g. the MP3 coding format. When however more than one tone is presented to the cochlea, a mechanism capable of shifting the center frequency of one or more of these critical bands may enable the individual to favor the transmission to the central auditory pathways of certain frequencies over others. These critical bands work on a "first- come first-served" basis, which means that a sound wave that arrives slightly earlier at the cochlea and auditory nerve will decide on the center frequency of a critical band, at the expense of a sound wave arriving slightly later. By adjusting the group delay and/or phase of the signal transmitted via the tympanic membrane to the cochlea, for certain frequencies or frequency bands, the TRRS steers the center frequency of some critical bands. This effect, adhering to a 'first-come first-served' principle, influences the critical bands only for the first few milliseconds of every new sound, and not for the following seconds of continuous acoustic input. Hence, in the human processing of speech, each new word (or even each syllable) may result in alterations of the transfer function of the tympanic membrane, as opposed to continuous background noise, which does not contain any specific cue for alterations. The effect of the filter, governed by the actions of the TRRS, then only influences only the first part of each new syllable or word or sound, and not the remaining part, nor continuous background noise. The different steering of the critical bands for the first few milliseconds of a new sound wave train may provide extra information and better possibilities for directing auditory attention to a preferred sound source, while the following time period of the sound wave train, where the critical bands are not steered in this way, provide a reliable image of the acoustic surroundings. By applying different strategies for both ears, and thus steering both filter banks in a different manner, different inputs are created for the critical bands involved, thus facilitating the creation of two auditory streams for further (human) processing. This creates the possibility for directing auditory attention, based on spectral cues, and the creation of an attended and unattended auditory stream. Hereby, the TRRS is capable of influencing the group delay and/or phase in both ears differently. The resulting time delay for certain frequencies between the two ears offers cues for directional listening, or directing auditory attention based on spatial cues.

Measurements relating to the TRRS involve spectral and/or spatial listening tests. Spectral and spatial listening tests allow to gather information regarding a listener's TRRS. Examples of such tests include LiSN-S and LiSN-S PGA. A requirement for such measurements is the ability to measure group delay and/or phase of sound transmitted through any or any combination of following parts of the TRRS: the tympanic membrane, the tympano-ossicular system, the ossicles, the oval and/or round window, the cochlea, the neural pathways. Hereby, signals can be measured on one side of the chain and recorded on any part inside the chain or on the other side. Any method that measures the transfer function associated with the tympanic membrane and thus the entire TRRS may be used. Preferably, measurements are done by measuring signals that are reflected from or transmitted through / carried over by part of the chain, such as the tympanic membrane. Measurements may be performed monaurally. In a preferred embodiment, measurements are performed binaurally.

In a preferred embodiment, measurements are done in the form of transfer function measurements comprising incident pressure measurement and reflected pressure measurement. Typically, this concerns measurements in the lower frequency range up to 2000 Hz. However, this may also entail the measurement of reflectance components relating to standing waves in the ear canal, which may occur for higher frequencies, e.g. in the frequency range of 2000 Hz and above. Furthermore, preferably, sound emitted into the ear canal and reflected by the tympanic membrane is recorded in the ear canal. In an alternative embodiment, an emitting microphone may be placed in the middle ear and a measuring microphone in the ear canal, or vice versa. In yet another embodiment, both an emitting microphone and measuring microphone may together measure tympanic membrane reflectance from inside the middle ear. Such a setup may be preferably used in cochlear implant or middle ear implant users. In another setup preferable for use with cochlear implants one may measure the group delay and/or phase not by recording sound but rather by measuring vibrations of the middle ear ossicles, or vibrations of the oval or round window, or vibrations measured inside the cochlea. In an alternative embodiment, one may measure signals resulting from brain activity. Another setup may be similar to sonotubometry, whereby an emitting microphone is placed in the nostrils, nose, nasopharynx, pharyngeal recess and/or Eustachian tube, and a recording microphone is positioned in the external ear canal. In a further preferred embodiment, several types of measurement may be combined. For example, transfer function measurements in which emitting and recording microphone are located in the ear canal may be combined with a modified sonotubometry, in which the emitting microphone is situated in the nostril(s) and the recording microphone in the ear canal or preferably in both ear canals.

Stimuli in listening tests can be clicks, chirps, words, sentences, or natural sound reaching the ear of the individual using the method. Stimuli can be emitted monaurally but much more information is obtained with binaural measurements, in which a similar stimulus is presented binaurally. If the stimuli do not consist of the natural sounds reaching the ear of the listener, they can be made inaudible for the listener, e.g. by making them very short, by using ultrasound stimuli or other means.

Measurement encompasses a form of signal processing. In this context, group delay is the time delay of the amplitude envelopes of the various sinusoidal components of a signal through a device under test, and is a function of frequency for each component. Phase delay, in contrast, is the time delay of the phase as opposed to the time delay of the amplitude envelope. Related, in this document, the term "frequency band" refers to a frequency interval comprising all frequencies larger than a first frequency and smaller than a second frequency. Since actual devices processing sound impact a certain finite-width frequency band rather than a single discrete frequency component, the term "frequency band" generically applies to any form of sound processing, whereby the frequency band may span a narrow band but may also span the entire audible spectrum, which, in case of human hearing, is commonly given as 20 to 20 000 Hz.

All frequency components of a signal are delayed when passed through a device such as an amplifier, a loudspeaker, or propagating through space or a medium, such as air. This signal delay will be different for the various frequencies unless the device has the property of being linear phase. The delay variation means that signals consisting of multiple frequency components will suffer distortion because these components are not delayed by the same amount of time at the output of the device. This changes the shape of the signal in addition to any constant delay or scale change. A sufficiently large delay variation can cause problems such as poor fidelity in audio.

Measurements of transfer functions, which may comprise measurements of reflectance components, are obtained in a listening test. In a preferred embodiment, measurements are performed in two subsequent stages: in rest, and then during/after tasks and/or stimuli, and this for both ears. For instance, one may evaluate the group delay and/or phase as a function of the frequency quantitatively, whereby qualitative parameters describe the shape of the curve obtained in the frequency domain. These qualitative parameters correspond to different states of the tympanic membrane. More particularly, several typical appearances of the curve of e.g. the group delay (in ms) as a function of the frequency (in Hz), which are referred to as "flat curve, ripple, swell and crest". Hereby, a "flat curve" refers to a curve in which there is minimal deviation from linearity. A "ripple" refers to a curve with irregular and not very outspoken deviations from linearity. A "swell" is a curve with regularly spaced and formed deviations or distortions, resembling waves and potentially indicating standing waves on the tympanic membrane. A "crest" refers to a solitary, outspoken deviation from linearity, for a specific rather narrow frequency range. The latter is a potential indicator for a resonance in the middle ear cavity. Some quantitative parameters of these distortions are the frequency of valleys, zero-crossings and peaks.

In a preferred embodiment, individual characteristics of the ear canal and TRRS as identified in the listening test results are used in the processing of transfer function measurements and the adjusting of the device. For instance, the same attention- related event can cause a change in group delay and/or phase at different frequency bands for different listeners. Hereby, also the width may vary: where a modified group delay and/or phase may be measured for a first listener at a broad frequency range of e.g. 700 to 1000 Hz, it may be measured at a narrow frequency range around e.g. 1500 Hz for a second listener. Therefore, the adjusting preferably takes into account this information, prioritizing the most relevant frequency bands in the processing.

In a preferred embodiment, the method according to the present invention further comprises following steps taking place during said test period :

(A) measuring an incident pressure in direct proximity of said tympanic membrane during said test period; (B) measuring a reflected pressure in direct proximity of said tympanic membrane during said test period; whereby said determining of said transfer function measurement associated with said tympanic membrane is based on said incident pressure, said reflected pressure and an acoustic frequency.

The advantage of this method is that it allows incorporating information regarding the base signals as well as sound originating from the current environment of the listener. Since the measured transfer function comprises components associated with both sources, improved processing is possible by taking into account both sources. In a further preferred embodiment, the device comprises a first measurement means measuring said incident pressure in the ear canal associated with said tympanic membrane, preferably by a directional microphone capturing air pressure waves, alternatively a bone-conducting microphone. In a further preferred embodiment relating to a hearing aid, the device comprises a second measurement means measuring said reflected pressure also in the ear canal associated with said tympanic membrane, preferably a second directional microphone capturing air pressure waves, alternatively a second bone-conducting microphone. In another embodiment relating to a middle ear implant and/or a cochlear implant, the device comprises a second measuring means measuring said reflected pressure in the middle air and/or the inner ear. In the latter case, said reflected pressure amounts to transduced pressure carried over from the ear canal by the tympanic membrane.

In a further preferred embodiment, the schedule relating to the method comprises at least one test period which is at least partially overlapping with at least one operational period in the time domain and/or the frequency domain. The advantage of such an embodiment is that adjustment of the device may extend beyond test periods and in absence of test audio signals, allowing for more measurements, and related, a higher accuracy in adjusting.

In a further preferred embodiment, the adjusting of said device comprises applying a gain and/or a delay and/or a filter in one or more frequency bands. Hereby, the filter may be any finite or infinite impulse response filter applied to one or more frequency bands, such as a gamma tone or gamma chirp filter bank. For instance, sound may be filtered in a certain frequency band so as to retain only a single frequency in the given band, with the intention of steering the center frequency of some of said critical bands, leading to improved hearing. In a further preferred embodiment, the method according to the present invention is applied binaurally, i.e. it is applied to a first tympanic membrane belonging to a listener and a second tympanic membrane belonging to said listener, whereby said transfer function measurement comprises a first measurement associated with a said first tympanic membrane and a second measurement associated with said second tympanic membrane, whereby said adjusting comprises a first adjustment associated with said first tympanic membrane and a second adjustment associated with said second tympanic membrane, and whereby said adjusting is further based on a difference of at least one component belonging to said first measurement and at least one component belonging to said second measurement. In a possible embodiment, said adjusting takes place for each ear separately, after which the combination is made through the comparison of both. In a preferred embodiment, however, said adjusting takes place in an integrated fashion, whereby the measurements of both ears are essentially synchronized, and the processing of both measurements is done concurrently and jointly. Preferably, this is combined with a step of determining an attention-relating frequency band during said test period, whereby said determining is based on at least one component of the transfer function; whereby said adjusting maximizes an objective function relating to a combination of said first adjustment and said second adjustment with respect to said attention-relating frequency band, preferably by applying a gain and/or a delay and/or a filter in one or more frequency bands, preferably comprising said attention-relating frequency band.

An advantage of such a binaural approach is the enabling of binaural masking and binaural unmasking, as described in, e.g., (B.C. J. Moore, An introduction to the psychology of hearing, Academic Press, 1997). When listening to, e.g., two competing sound sources with similar amplitude, a first source exhibiting an inter-aural phase difference for a relevant frequency band in the lower frequency range is perceived (subjectively) as longer-lasting and/or subjectively louder than a second source that does not exhibit an inter-aural phase difference in this frequency range, with reported subjective improvement of up to 13 dB for a phase difference of about 180 degrees, i.e. a phase inversion. Correspondingly, adjusting a device so as to create an inter- aural difference in (phase) delay in an attention-relating frequency band may help a listener to separate a first source from a second source, leading to improved adjusting. Hereby, note that the first source is not necessarily masking the second source in this case; rather, the given that one of both sources exhibits a phase inversion helps the listener to separate one source from the other. In a preferred embodiment, said separating of said first source relative to said second source is enforced via the maximization of an objective function quantifying the level of separation. In said example of two competing sources with similar amplitude and said first source exhibiting said inter-aural phase difference, the objective function relates to the inter-aural phase difference, and the maximization of the objective function relates to said inter-aural phase difference approaching 180 degrees (phase inversion). Related, in the present invention, one may assume that the listener may consciously or unconsciously position his ears so as to maximize a phase difference with respect to an attention-related frequency band, since maximizing the phase difference for this band makes that it is perceived as louder. Hence, in a preferred embodiment provided by the present invention, the detection of large inter-ear phase differences in certain frequency bands may be indicative of intentional focus with respect to said frequency bands.

In two embodiments of the device according to the present invention, relating to the measurement of said transfer function, one may either use environment sound, i.e. incident pressure, or one may use said base signal. In the former case, the device preferably comprises a first microphone for recording incident pressure and a second microphone for recording the reflected pressure. In the latter case, the device preferably comprises a sound source such as a loudspeaker for playing the base signal, as well as said second microphone for recording the reflected pressure. In both cases, the device should comprise measurement means for the measurement of incident pressure, which may preferably be implemented by said first microphone, or alternatively by another dedicated microphone not relating to the measurement of said transfer function. Furthermore, in an embodiment where the device comprises a sound source such as a loudspeaker for playback of incident waves, e.g. in case of a hearing aid, this sound source may be implemented by said loudspeaker configured for playback of said base signal, or alternatively by another dedicated sound source.

In another embodiment of the present invention, which is not intended to limit the scope in any way, the device according to the present invention discriminates between two attention-related states of the listener, whereby the listener is actively listening, in a first state, and the listener is not actively listening, in a second state. Adjustment of the device is done according to at least the transfer functions in the two states. Since the TRRS behaves differently in the first state than in the second state, also the related transfer function measurements differ significantly. Hence, detection of this difference may allow for a simplified design of the device, whereby the device preferably only measures either one of the group delay, the phase or the amplitude. Furthermore, in a preferred embodiment, the measurement is not done for the entire frequency range but only for a limited frequency band or frequency, such as a band of 700 to 1000 Hz, or a narrow frequency range around e.g. 1500 Hz. In another embodiment, the frequency range of the device is 100 to 3400 Hz, more preferably 500 to 2000 Hz.

In another embodiment of the present invention, which is not intended to limit the scope of the invention, the adjusting of the device according to the present invention comprises the activation, de-activation or adjustment of a basic function of the hearing aid. Examples of such basic functions are turning on or off of background noise filters, increasing/decreasing of background noise filters, toggling between omnidirectional processing and directional processing, the tuning of the device between omnidirectional processing and directional processing, the adjusting of volume, etc.

It is supposed that the present invention is not restricted to any form of realization described previously and that some modifications can be added to the presented example of fabrication without reappraisal of the appended claims. For example, the present invention has been described referring to a device such as a hearing aid or a cochlear implant, but it is clear that the invention can be applied to other devices that facilitate hearing, such as bone-anchored hearing aids or middle ear implants. Likewise, the results of the measurement associated with one listener can be used to steer devices for other listeners, so as to ameliorate the hearing capacities of these other listeners in a similar way as the listener wearing the measurement device. The results can be used to steer auditory attention of a robot, in a similar way as described above, and based on a top-down approach based on the working of the TRRS of the individual wearing the measuring device. Similarly the device can be used to steer non-auditory attention of characters in interactive plays such as games, interactive films and other interactive audiovisual media. Furthermore, the device can be used in 3D systems, speech recognition systems and speech-to-text systems. In a similar way as the cross-modal interactions allow for directing and modulating several forms of attention (e.g. visual attention) based on inputs from other senses, these results can be used to direct and modulate other forms of attention of these robots and characters. Furthermore, the device and method according to the present invention may be used to train people by giving them feed-back regarding the operation of their TRRS.

According to another aspect, which is not intended to limit the scope of the invention in any way, the present invention offers a method for adjusting a device according to a schedule comprising one or more test periods and operational periods, said method comprising the steps of: (a) determining a transfer function measurement associated with a tympanic membrane of a listener during a test period belonging to said one or more test periods;

(b) adjusting said device in function of said transfer function measurement; (c) letting said device process incident pressure according to said adjusting during an operational period belonging to said one or more operational periods;

(d) optionally, repeating steps (a) to (c) according to said schedule associated with said one or more test periods and said one or more operational periods; wherein said transfer function measurement comprises an amplitude measurement, a group delay measurement and preferably a phase measurement, wherein said determining in step (a) comprises detecting whether a pre-defined pattern is present in said transfer function measurement, the presence of said pre-defined pattern being manipulable by the listener, and wherein said adjusting in step (b) takes into account whether said pre-defined pattern was detected. In a preferred embodiment, said schedule comprises a plurality of test periods and operational periods, wherein step (d) comprises repeating step (a) to (c) according to said schedule associated with said plurality of test periods and operational periods. Furthermore, in a preferred embodiment, said schedule comprises at least one operational period which is at least partially non-overlapping with any test period in the time domain and/or the frequency domain. Hereby, an operational period is at least partially non-overlapping with any test period in the time domain if a time interval exists which belongs to an operational period but does not belong to any test period. Complementary to this, an operational period is at least partially non- overlapping with any test period in the frequency domain if a frequency band can be identified for which no test activity associated with a test period takes place during a time interval which belongs to an operational period.

According to another aspect, which is not intended to limit the scope of the invention in any way, the invention relates to following points 1 to 15.

1. A method for adjusting a device according to a schedule comprising one or more test periods and operational periods, said method comprising the steps of:

(a) determining a transfer function measurement associated with a tympanic membrane of a listener during a test period belonging to said one or more test periods; (b) adjusting said device in function of said transfer function measurement;

(c) letting said device process incident pressure according to said adjusting during an operational period belonging to said one or more operational periods; (d) optionally, repeating steps (a) to (c) according to said schedule associated with said one or more test periods and said one or more operational periods; wherein said transfer function measurement comprises an amplitude measurement, a group delay measurement and preferably a phase measurement, wherein said determining in step (a) comprises detecting whether a pre-defined pattern is present in said transfer function measurement, the presence of said pre-defined pattern being manipulable by the listener, and wherein said adjusting in step (b) takes into account whether said pre-defined pattern was detected. The method according to the preceding point 1, wherein said schedule comprises a plurality of test periods and operational periods, and wherein step (d) comprises repeating step (a) to (c) according to said schedule associated with said plurality of test periods and operational periods. The method according to the preceding points 1 and 2, wherein said schedule comprises at least one operational period which is at least partially non- overlapping with any test period in the time domain and/or the frequency domain. The method according to the preceding points 1 to 3, said method further comprising following steps taking place during said test period :

- playing back a base signal in direct proximity of said tympanic membrane;

- determining an unprocessed transfer function measurement associated with said tympanic membrane;

- processing said unprocessed transfer function measurement taking into account at least said base signal to obtain said transfer function measurement; wherein said transfer function measurement varies with the incident pressure; The method according to the preceding points 1 to 4, said method further comprising following steps taking place during said test period : (A) measuring an incident pressure in direct proximity of said tympanic membrane during said test period;

(B) measuring a reflected pressure in direct proximity of said tympanic membrane during said test period; whereby said determining of said transfer function measurement associated with said tympanic membrane is based on said incident pressure, said reflected pressure and an acoustic frequency.

6. The method according to any of the preceding points 1 to 5, wherein said determining of said transfer function measurement is performed for more than one frequency band.

7. The method according to any of the preceding points 1 to 6, wherein the determining of the transfer function measurement and the adjusting of said device are performed automatically by a processor, and wherein optionally any or any combination of the following is performed automatically by said processor: the measuring of the incident pressure; the measuring of the reflected pressure; the determining of said unprocessed transfer function measurement; the processing of said unprocessed transfer function measurement.

8. The method according to any of the preceding points 1 to 7, whereby said transfer function measurement further comprises any or any combination of the following : a phase delay, an attenuation and an acoustic frequency.

9. The method according to any of the preceding points 1 to 8, wherein the adjusting of said device comprises applying a gain and/or a delay and/or a filter in one or more frequency bands and/or the increasing/decreasing of background noise filters and/or the adjusting between omnidirectional processing and directional processing.

10. The method according to any of the preceding points 1 to 9, whereby said adjusting of said device is further based on information regarding listening tests specific to said tympanic membrane, preferably spectral and/or spatial listening tests. 11. The method according to any of the preceding points 1 to 10, whereby said method is applied to a first tympanic membrane belonging to a first person and a second tympanic membrane belonging to said first person, whereby said transfer function measurement comprises a first measurement associated with a said first tympanic membrane and a second measurement associated with said second tympanic membrane, whereby said adjusting comprises a first adjustment associated with said first tympanic membrane and a second adjustment associated with said second tympanic membrane, and whereby said adjusting is further based on a difference of at least one component belonging to said first measurement and at least one component belonging to said second measurement. The method according to point 11, said method further comprising the step of determining an attention-relating frequency band during said test period, whereby said determining is based on at least one transfer function measurement; whereby said adjusting maximizes an objective function relating to a combination of said first adjustment and said second adjustment with respect to said attention-relating frequency band, preferably by applying a gain and/or a delay and/or a filter in one or more frequency bands, preferably comprising said attention-relating frequency band. A device for processing an incident pressure, said device comprising — an acoustic device, configured for receiving and transmitting an acoustic energy, configured for determining a transfer function measurement associated with said acoustic energy and a tympanic membrane of a listener, and configured for operating according to a schedule comprising one or more test periods and operational periods, whereby said transfer function measurement comprises an amplitude measurement, a group delay measurement and preferably a phase measurement;

— at least one processor, configured for:

(I) processing information associated with said determined transfer function measurement during a test period belonging to said plurality of test periods;

(II) determining a transfer function measurement during said test period;

(III) adjusting at least one parameter associated with the device based on at least one component of said transfer function measurement;

(IV) receiving and transmitting acoustic energy according to said adjusting during an operational period belonging to said plurality of operational periods; wherein said determining in step (II) comprises detecting whether a pre-defined pattern is present in said transfer function measurement, the presence of said pre-defined pattern being manipulate by the listener, and wherein said adjusting in step (III) takes into account whether said pre-defined pattern was detected.

14. The device according to point 13, whereby said device is any or any combination of the following : a hearing aid, a middle ear implant, a cochlear implant, a bone- anchored hearing aid, a bone-conducting microphone.

15. The device according to any of the point 13 to 14, whereby the device is configured to perform the method according to any of the point 1 to 12.

The invention is further described by the following non-limiting examples which further illustrate the invention, and are not intended to, nor should they be interpreted to, limit the scope of the invention.

Examples

Example 1 : First device example

Figure 1 shows a first example embodiment of a hearing device 1 according to the present invention, wherein the hearing device 1 preferably concerns a hearing aid. The hearing device 1 comprises a housing 2, an input audio microphone 3, an input audio preamplifier 4, an input audio digitizer 5, an input audio wireless receiver 6, a mixer 7, a reflected audio microphone 33, a reflected audio preamplifier 34, a reflected audio digitizer 35, a signal processor 68, a power amplifier 9, a vibrator 10, an input audio correlator 57, a processed audio correlator 58 and a control unit 19. The hearing device 1 is preferably powered by means of a battery or an accumulator (not shown) in the housing 2.

The housing 2 is adapted to be worn by a human listener at the ear or in the ear canal. It is arranged in an operating position on the head of a listener of the hearing device 1, e.g. by means of an elastic band or a spring. The input audio microphone 3 is arranged to receive an acoustic input signal from the listener's surroundings when the housing 2 is in the operating position and is adapted to provide an input audio microphone signal to the input audio preamplifier 4 in dependence on the acoustic input signal. The input audio preamplifier 4 is adapted to amplify the input audio microphone signal and to provide the amplified input audio microphone signal to the input audio digitizer 5, which is adapted to digitize the amplified microphone signal and provide a corresponding first digital input audio signal. The input audio wireless receiver 6 is adapted to receive a wireless communication signal, such as e.g. an inductive signal, a radio-frequency (RF) signal or an optical signal, from a further device (not shown) and to provide a second digital input audio signal in dependence on the received wireless communication signal. The input audio microphone 3, the input audio preamplifier 4, the input audio digitizer 5 and the input audio wireless receiver 6 thus function as input means that provide input audio signals. The input means may comprise further input audio microphones 3, input audio preamplifiers 4 and/or input audio digitizers 5 providing respective further input audio signals in order to allow e.g. direction-dependent processing of the received acoustic signals.

The mixer 7 is connected to receive the first and second digital input audio signals and is adapted to provide a combined audio signal comprising a linear combination of the first and second digital input audio signals in dependence on a control signal received from the control unit 19. The signal processor 68 is adapted to processed the combined audio signal in accordance with the purpose of the hearing device 1, e.g . to improve or augment the hearing capability of the listener with respect to the listener's auditory attention and/or to amplify or convey a received audio signal to the listener. The power amplifier 9 is connected to receive the processed audio signal and is adapted to provide a corresponding amplified output signal to the vibrator 10, preferably a loudspeaker, which is arranged to transmit a structure-borne acoustic output signal to the listener in dependence on the amplified output signal when the housing 2 is in the operating position. The power amplifier 9 preferably comprises a pulse-width modulator or another type of digital-to-analog converter in order to provide the amplified output signal as an analog signal to the vibrator 10. The reflected audio microphone 33 is arranged to receive an acoustic reflected signal from the listener's tympanic membrane, wherein the tympanic membrane belongs to said listener's ear. The acoustic reflected signal is received from the tympanic membrane when the housing 2 is in the operating position at the listener's ear or in the listener's ear canal. The reflected audio microphone 33 is adapted to provide a reflected audio microphone signal to the reflected audio preamplifier 34 in dependence on the acoustic reflected signal. The reflected audio preamplifier 34 is adapted to amplify the reflected audio microphone signal and to provide the amplified reflected audio microphone signal to the reflected audio digitizer 35, which is adapted to digitize the amplified microphone signal and provide a corresponding digital reflected audio signal. In a preferred embodiment, given their related aim, the reflected audio microphone 33, the reflected audio preamplifier 34 and the reflected audio digitizer 35 are essentially identical in function and operation to the input audio microphone 3, the input audio preamplifier 4 and the input audio digitizer 5, respectively. Moreover, for reducing footprint and cost, preferably the input audio preamplifier 4 and the reflected audio preamplifier 34 are essentially provided in a single integrated component capable of handling two independent input and two independent output channels. Similarly, preferably, the input audio digitizer 5 and the reflected audio digitizer 35 are essentially provided in a single integrated component capable of handling two input signals and two output signals.

In a preferred embodiment, the input audio microphone 3, on the one hand, and the reflected audio microphone 33 and the vibrator 10, are provided at opposite sides of the hearing device 1. This advantageously allows to position the housing 2 with input audio microphone 3 facing the acoustic input signal from the listener's surroundings, while concurrently allowing to position the housing 2 with the reflected audio microphone 33 and the vibrator 10 facing the tympanic membrane.

The mixer 7 and the signal processor 68 belong to a signal processing circuit (7, 68, 57, 58, 19), which is adapted to process at least one of the first and the second input audio signals to provide a processed audio signal, i.e. the modified audio signal. To this end, the signal processing circuit (7, 68, 57, 58, 19) further comprises the input audio correlator 57 and the processed audio correlator 58, and the control unit 19 for altering the processing based on at least one output of at least one of said input audio correlator 57 and said processed audio correlator 58.

The input audio correlator 57 is connected to receive the first and second digital input audio signals from the input audio microphone 3 and the input audio wireless receiver 6, respectively, as well as to receive the digital reflected audio signal from the reflected audio microphone 33. The input audio correlator 57 is configured for repeatedly determining, by comparing said digital reflected audio signal to at least one of first and second digital input audio signals, a transfer function measurement associated with the tympanic membrane. Hereby, for each measurement associated with a given time instant or a given time period, said transfer function measurement comprises at least two dimensions, with a phase-related measurement and preferably an amplitude measurement. For each of these dimensions, preferably a plurality of measurement points associated with a plurality of frequency bands and/or phase bands is determined. To this end, the input audio correlator 57 preferably comprises means for repeatedly executing a discrete Fourier transform, e.g. by means of a dedicated FFT circuit, on each of the input audio signals and on the reflected audio signal to determine input audio spectra for each of the input audio signals and a reflected audio spectrum for the reflected audio signal, respectively, and for comparing the determined reflected audio spectrum to one or more of the determined input audio spectra, to determine amplitude difference audio spectra with respect to one or more of the audio input signals. Related, the input audio correlator 57 preferably comprises means for repeatedly executing a phase difference detection algorithm on each of the input audio signals and on the reflected audio signal to determine phase difference audio spectra with respect to each of the input audio signals, by comparing the phase of the reflected audio signal to the phase of each of the input audio signals with respect to one or more frequency bands.

The processed audio correlator 58 is connected to receive the processed audio signal from the signal processor 68 and to receive the digital reflected audio signal from the reflected audio microphone 33. The processed audio correlator 58 is configured for repeatedly determining, by comparing said digital reflected audio signal to the processed audio signal, a transfer function measurement associated with the tympanic membrane. Hereby, for each measurement associated with a given time instant or a given time period, said transfer function measurement comprises at least two dimensions, with a phase-related measurement and preferably an amplitude measurement. For each of these dimensions, preferably a plurality of measurement points associated with a plurality of frequency bands and/or phase bands determined. To this end, the processed audio correlator 58 preferably comprises means for repeatedly executing a discrete Fourier transform, e.g. by means of a dedicated FFT circuit, on the processed audio signal and on the reflected audio signal to determine a processed audio spectrum for the processed audio signal and a reflected audio spectrum for the reflected audio signal, and for comparing the determined reflected audio spectrum to the determined processed audio spectrum. Related, the processed audio correlator 58 preferably comprises means for repeatedly executing a phase difference detection algorithm on the processed audio signal and on the reflected audio signal to determine a phase difference audio spectrum, by comparing the phase of the reflected audio signal to the phase of the processed audio signal with respect to one or more frequency bands. The processed audio correlator may alternatively or additionally be connected to receive audio signals from one or more other components of the input means 3, 4, 5, 6 or of further components of the signal processing circuit 7, 68, 57 and/or the power amplifier 9, in which case the processed audio correlator 58 may be required to determine further or other transfer function measurements and/or processing algorithms in order to accurately correlate the reflected signal to the audible signal as outputted by the vibrator 10.

Furthermore, both the input audio correlator 57 and the processed audio correlator 58 are configured for repeatedly determining a correlation value of the determined transfer function measurement with respect to each of at least two pre-determined profiles. In one embodiment, the at least two pre-determined profiles may be stored and available for read-out in a dedicated memory comprised in each of the input audio correlator 57 and the processed audio correlator 58. In another embodiment, the at least two pre-determined profiles may be stored and available for read-out in a shared memory comprised in the hearing device 1 present outside of and connected to the input audio correlator 57 and the processed audio correlator 58. Said at least two predetermined profiles comprise at least one profile associated with a first tensioning of the membrane and low intentional focus and at least one profile associated with a second tensioning of the membrane and high intentional focus.

The at least two pre-determined profiles may be determined based on measurements made on one or more individuals using an arrangement equal to or similar to the arrangement of the hearing device 1. A better performance may be achieved by determining the pre-determined profiles based on measurements made on the listener intended to use the particular hearing device 1. In this case, the hearing device 1 itself is preferably used as a measurement instrument in a calibration session with the housing 2 in the operating position. In the calibration session, a range of different audible signals is presented to the listener, preferably by means of the hearing device 1, and simultaneously, transfer function measurements are determined based on the reflected audio signals received from the reflected audio microphone 33. In the calibration session, determining the two or more profiles may be performed in and by the hearing device 1 itself or, partially or entirely, in a connected computer, e.g. in a fitting system for adapting other settings of the hearing device 1 to the needs of the listener. In one example, the at least two pre-determined profiles consist of a first example profile and a second example profile. The first example profile is associated with a first tensioning of the membrane and low intentional focus and the second example profile is associated with a second tensioning of the membrane and high intentional focus are determined by means of said hearing device 1 which is an a calibration mode. Hereby, the calibration mode may comprise a calibration method, while the housing is its operating position, applied to said listener. During the calibration method, the hearing device 1 may be fully operative, except that preferably (but not necessarily) the processing by the signal processor 68 is temporarily "frozen" at some fixed state without altering of the signal processing, such as some pre-defined calibration state, or the current state of the signal processor 68 upon initiation of the calibration. The calibration method is preferably supervised by a human operator/expert but may also be performed by the listener, e.g. based on written instructions or instructions spoken out loud by the hearing device 1 upon initiation of the calibration by the listener, e.g. by means of some calibration button present on the hearing device 1. The calibration method comprises two steps.

- In a first calibration step, a series of words spoken out loud by a human is provided while the listener is not intentionally listening and/or is not prepared to intentionally listen, e.g. because a sufficiently long silence of e.g. 10 or 20 seconds or a sufficiently long period of noise of 10 or 20 seconds or a sample with nature sounds of 10 or 20 seconds precedes the start of the series of spoken words. This preferably relates to some recording of words, but may also relate to words spoken out loud by someone present with the listener such as said human operator/expert. For this calibration step, the input audio correlator 57 and the processed audio correlator 58 determine the one or more amplitude difference audio spectra and the one or more phase difference audio spectra, either as they evolve over time, or by averaging the measurements over time to one or more averaged audio spectra. The resulting one or more measurements may be further combined, interpolated and or mixed or may be used directly, to then be stored as said first example profile in said memory. Preferably, the measurements, and hence the first example profile, relate to a measurement frequency band 91 with an upper measurement frequency B, said upper measurement frequency B smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz. Such a measurement frequency band 91 advantageously utilizes the detectability of the tensioning of the tympanic membrane at low frequencies with the providing of a series of words spoken out loud at higher frequency, preventing any undesirable acoustic feedback loop between the "steering" of the tympanic membrane and the measurement of the tensioning of the membrane.

- In a second calibration step, the same series of words spoken out loud by said human is provided while the listener is intentionally listening, e.g. whereby the listener is provided with a long sequence of e.g. 10 or 20 seconds of other words spoken out loud by said human, providing the listener with the time needed to intentionally focus on the voice of said human. This preferably relates to some recording of words, but may also relate to words spoken out loud by someone present with the listener such as said human operator/expert. Also for this second calibration step, the input audio correlator 57 and the processed audio correlator 58 determine the one or more amplitude difference audio spectra and the one or more phase difference audio spectra, either as they evolve over time, or by averaging the measurements over time to one or more averaged audio spectra. The resulting one or more measurements may be further combined, interpolated and or mixed or may be used directly, to then be stored as said second profile in said memory. Preferably, the measurements, and hence the second example profile, relate to a measurement frequency band 91 with an upper measurement frequency B, said upper measurement frequency B smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz. Again, such a measurement frequency band 91 advantageously utilizes the detectability of the tensioning of the tympanic membrane at low frequencies with the providing of a series of words spoken out loud at higher frequency, preventing any undesirable acoustic feedback loop between the "steering" of the tympanic membrane and the measurement of the tensioning of the membrane.

Both the input audio correlator 57 and the processed audio correlator 58 are configured for repeatedly determining a correlation value of the determined transfer function measurement with respect to each of at least two pre-determined profiles, such as said first example profile and said second example profile. The correlation value is indicative of an intentional focus level of said tympanic membrane, and may preferably be determined as a quantitative measure of similarity between new measurements and the profiles. For instance, the correlation value may relate to a similarity between a newly measured amplitude difference audio spectrum and the amplitude difference audio spectra comprised in the at least two profiles. Alternatively or additionally, the correlation value may relate to a similarity between a newly measured phase difference audio spectrum and the phase difference audio spectra comprised in the at least two profiles. This preferably amounts to assigning a plurality of weights with respect to said at least two profiles. Examples of quantitative measures to assess similarity may relate to e.g. least squares deviations or least absolute deviations, whereby the profile for which the smallest deviation is found is attributed a higher correlation value or "weight".

The correlation value output by the input audio correlator 57 and/or the processed audio correlator 58 may be used as an indication of whether the listener pays attention to the presented audible signal or not. Paying attention hereby corresponds to high correlation values with respect to the at least one profile associated with a second tensioning of the membrane and high intentional focus and/or low correlation values with respect to the at least one profile associated with a first tensioning of the membrane and low intentional focus. Not paying attention may correspond to low correlation values with respect to the at least one profile associated with the second tensioning of the membrane and high intentional focus and/or high correlation values with respect to the at least one profile associated with the first tensioning of the membrane and low intentional focus.

In addition or alternatively, the correlation value output by the input audio correlator 57 and/or the processed audio correlator 58 may be used as an indication of whether the listener pays attention to the entire presented audible signal or only to a part of the audible signal. Such difference may relate to frequency ranges in which signals are present in the processed audio spectra but not - or to lesser degree - in the reflected audio spectra and leads to differences when comparison is performed with respect to said at least two profiles, preferably including not only the two profiles with low/high intentional focus but also one or more profiles that are exemplary for intentional focus directed at some specific attention-related frequency band. Paying attention to the entire presented audible signal hereby corresponds to high correlation values with respect to the at least one profile associated with a second tensioning of the membrane and high intentional focus but low correlation values with respect to any of the profiles associated with intentional focus directed at some specific attention- related frequency band. Paying attention only to part of the audible signal may correspond to high or moderate correlation values with respect to the at least one profile associated with the second tensioning of the membrane combined with high correlation values with respect to at least one of the profiles associated with intentional focus directed at some specific attention-related frequency band.

In view of the above, the correlation value output by the input audio correlator 57 and/or the processed audio correlator 58 may also be used as an indication to discriminate between three situations, i.e. whether

- the listener pays attention to the entire presented audible signal, - the listener pays attention to only a part of the audible signal, or

- the listener does not pay attention to the presented audible signal.

Hereby, the correlation value output by the input audio correlator 57 and/or the processed audio correlator 58 may further allow to discriminate between a plurality of pre-determined parts of the audible signal, e.g. between pre-determined different frequency bands of the audible signal, whereby for each pre-determined frequency band a corresponding profile is present in the memory, allowing to detect an attention-related frequency band in the audible signal. As discussed in details elsewhere in this document, a prime indicator for identifying attention-related frequency bands may be a large deviation in phase difference at the given frequency band when compared to the overall level of the phase difference audio spectrum. The control unit 19 is connected to receive the correlation values and preferably also the amplitude difference audio spectra and the phase difference audio spectra of the input audio correlator 57 and the processed audio correlator 58. In one embodiment, the control unit 19 operates purely on the correlation values received, and alters the processing based thereupon, according to some processing-related transfer function, associated with the combination of the mixer 7 and the signal processor 68, preferably a processing-related transfer function associated with the signal processor 68. This may for instance comprise applying one or more fir (finite impulse response) or iir (infinite impulse response) filters on the mixed audio signal provided by the mixer 7. Hereby, each of the correlation values may e.g. serve as weight to the application of a corresponding filter, whereby said corresponding filter is known to enhance listening if the corresponding correlation value is high. In another embodiment, the control unit 19 may also take into account the amplitude difference audio spectra and the phase difference audio spectra of the input audio correlator 57 and the processed audio correlator 58, allowing to process the signal according to deviations between the output of the input audio correlator 57 and the output of the processed audio correlator 58. In one related embodiment, the control unit 19 may be configured to reduce the intensity of processing if the received correlation values indicate a decrease in intentional focus, while increasing the intensity of processing if the received correlation values indicate an increase in intentional focus. In yet another related embodiment, the control unit 19 may switch the signal processing circuit (7, 68, 57, 58, 19) or the entire device 1 to a low/passive power mode if some long period of low intentional focus of e.g. 5 or 10 or 20 or 30 minutes is detected, or the control unit 19 may switch the signal processing circuit (7, 68, 57, 58, 19) or the entire device 1 to a high/active power mode if a change in the correlation values indicates that the intentional focus changes from low to high.

Furthermore, the control unit 19 may monitor the correlation values and/or the amplitude difference audio spectra and the phase difference audio spectra of the input audio correlator 57 and the processed audio correlator 58 and react to a change thereof by altering the processing such that the change is counteracted. The control unit 19 may alter the processing by altering a processing-related transfer function associated with the combination of the mixer 7 and the signal processor 68, preferably a processing-related transfer function associated with the signal processor 68. The control unit 19 may e.g. alter the transfer function such that it suppresses signals falling outside an attention-related frequency band detected based on the correlation values. Since the suppression of such frequency ranges makes the attention-related frequency bands sound relatively louder, provided that the listener maintains attention to the same parts of the sound environment, it further causes a flattening of the amplitude/phase difference audio spectra, and hence a reduction of the correlation values with respect to profiles that are indicative of an attention-related frequency band. Alternatively, the control unit 19 may set the processing-related transfer function to a predefined frequency characteristic, e.g. a frequency characteristic that is known to enhance speech signals in noisy environments, preferably when detecting that intentional focus is high. As a further alternative, the control unit 19 may set the processing-related transfer function to a predefined frequency characteristic known to enhance female speech or male speech if the correlation values indicate that the listener focuses attention respectively on a female speaker, with high correlation value with respect to a female-speaker-related profile, or a male speaker, with high correlation value with respect to a male-speaker-related profile.

Alternatively, or additionally, the control unit 19 may alter the processing by altering the relative levels of the first and second digital input audio signals in the mixer 7. The control unit 19 may e.g. compare the correlation values for the first and the second digital input audio signals and control the mixer 7 to emphasize the particular input audio signal that has the most desirable correlation values, e.g. the highest correlation values with respect to a preferred profile. Similarly as described above, this makes that the output audio signal is more similar to the one that is desired by the listener based on the reflected signal, typically resulting in "flatter" and/or lower amplitude/phase difference audio spectra. Furthermore, sound sources only present in the suppressed one of the first and second digital input audio signals are removed from the audible signal, and thus, one or more portions and/or the overall level of the amplitude/phase difference audio spectra are expected to decrease. The degree of emphasis applied by the control unit 19 may depend on the absolute differences present in the amplitude/phase difference spectra in order to avoid altering in cases where it is unlikely that a change will further reduce a measured difference.

Alternatively, or additionally, the control unit 19 may alter the processing by altering an acoustic directivity pattern of the input means 3, 4, 5, 6. A directional input audio microphone 3, e.g. a cardioid, a figure-eight or a hypercardioid microphone, may e.g. comprise two or more omnidirectional microphones, the outputs of which are combined as is already known in the art. The directivity pattern of such a directional input audio microphone 3 may be altered by altering the levels, phases and/or time delays of the microphone signals to be combined. The input means may comprise several such directional microphones 3 with different directivity patterns using the output signals of the same two or more omnidirectional microphones. The outputs of these directional microphones 3 may each be amplified in a respective input audio preamplifier 4 and digitized in a respective input audio digitizer 5 to provide respective input audio signals. In this case, the acoustic directivity pattern may be altered by altering the relative levels of the input audio signals in the mixer 7 as described above. Alternatively, the directivity patterns may be altered by controlling how the output signals of the omnidirectional microphones are combined. In this case, the control unit 19 may follow e.g. a trial-and-error approach to determine which combination of the output signals of the omnidirectional microphones to use. As a further alternative, the directivity pattern may be set to a predefined pattern which emphasizes acoustic signals arriving from the front, assuming that the listener (as is usually the case) is focusing on sound sources in front of him or her.

Furthermore, the input audio wireless receiver 6 may receive a microphone signal from an external microphone (not shown) located in the vicinity of the listener, and in this case, in a wider sense, altering the relative levels in the mixer 7 also alters the acoustic directivity pattern of the input means 3, 4, 5, 6. Instead of, or in addition to, monitoring the correlation values and react to a change thereof, preferably reacting to a decrease with respect to a target profile, the control unit 19 may access the memory comprising the at least two profiles and may determine the correlation values for different input audio signals and, in dependence on a change of one or more of the correlation values, alter the processing as described above to emphasize the input audio signal with the desired property, e.g. the highest correlation value for a certain target profile. This may e.g. be of convenience when the hearing device 1 is a hearing aid and the listener receives sound from an acoustically muted TV set through the wireless receiver and speech from a nearby person through the microphone. In this case, the hearing device 1 may automatically emphasize the sound source the listener is focusing attention on.

Ideally, the altering of the processing causes the audible signal to contain precisely those audio components that the listener is interested in listening to, e.g. in accordance to the maximization of the correlation values with respect to some predetermined target profile. The listener may therefore continue listening without the stress of having to mentally remove or disregard disturbing signals. However, the altering may remove or suppress desired audio components, and/or the desired audio components may change in frequency content, direction or input source (input audio microphone 3 or input audio wireless receiver 6) after a processing alteration has been applied. In these cases, the listener may attempt to focus on sound components only faintly present in the audible signal, which may cause the correlation values for the pre-determined target profile to decrease. The control unit 19 may in some cases try to counteract this decrease by removing or suppressing further sound components, which may eventually lead the hearing device 1 into a deadlock in which no signal at all is presented to the listener. The reason for this is that the listener cannot focus on sound components that are not present in the audible signal, and there is therefore no built-in mechanism to take back an applied suppression.

In order to avoid this deadlock, the control unit 19 may confine amplification changes caused by the processing alterations to remain within a pre-determined range, e.g. ± 10 dB, ±6 dB, ±3 dB or ±2 dB. Additionally, or alternatively, the control unit 19 may reverse a previously made processing alteration in dependence on a pre-determined event. For instance, the control unit 19 may cancel suppression of frequency ranges and/or switch from a directional input audio microphone 3 to an omnidirectional input audio microphone 3. A suitable event may e.g. be that one or more of the input audio signals changes, e.g. as a result of the listener moving to a different location. Examples of further suitable events are a change in one or more of the input audio spectra or a decrease of one or more of the correlation values. Furthermore, the control unit 19 may reverse the alteration in dependence on activation of a listener control, i.e. a reset button, on the hearing device 1 or on a remote control and/or in dependence on expiry of a pre-determined time period. The above described events may be combined arbitrarily. The detection of decreases in the correlation values provided by the input audio correlator 57 and/or the processed audio correlator 58 may comprise comparing the respective correlation values with a corresponding threshold and/or determining a change rate of the respective correlation values and comparing the change rate with a corresponding threshold. The control unit 19 may use the result of such comparisons to determine whether a decrease is large and/or fast enough to trigger an alteration of the processing.

In one embodiment (not shown), the hearing aid 1 comprises an input audio correlator 57 but no processed audio correlator 58. In such a case the input audio correlator 57 may be referred to as the first correlator 57, with no second correlator present. In another embodiment (not shown), the hearing aid 1 comprises a processed audio correlator 58 but no input audio correlator 57. In such a case the processed audio correlator 58 may be referred to as the first correlator 58, with no second correlator present.

In a preferred embodiment as shown and according to the present example, the hearing aid 1 comprises both an input audio correlator 57 and a processed audio correlator 58. Hereby, the input audio correlator 57 is referred to as the first correlator 57, and the processed audio correlator 58 is the second correlator. The control unit 19 may use the correlation values of both the input audio correlator 57 and the processed audio correlator 58 according to some fixed relative amount. Alternatively or additionally, the control unit 19 may e.g. attribute a larger weight to the correlation values of the input audio correlator 57 when the sound environment changes or some other of said suitable events is detected, demanding larger changes in the processing. Furthermore, the control unit 19 may e.g. attribute a larger weight to the correlation values of the processed audio correlator 58 when the sound environment appears to be unchanged and none of said suitable events is detected, demanding smaller changes in the processing.

Example 2: Second device example

Figure 2 shows a second example embodiment of a hearing device 1 according to the present invention. The hearing device 1 is as described in the first example, with the addition of a test audio signal generator 80 comprised in said signal processing circuit (7, 68, 57, 58, 19). The test audio signal generator 80 is connected to the signal processor 68 and is adapted for providing a test audio signal to said signal processor 68. This allows to mix the test audio signal into said processed audio signal for enabling determining of said intentional focus level of said tympanic membrane, said mixing of said test audio signal preferably triggered by a schedule and/or by a detection of an event present in said one or more input signals. Particularly, mixing the test audio signal into said processed audio signal allows to determine a transfer function measurement with respect to a known test audio signal or base signal. This allows transfer function measurements which are sufficiently accurate by controlling the power of the test audio signal, and which comprise useful information regarding the current environment of the listener and the influence of the current environment on the listener, since they comprise a registration of the conditioning of the TRRS as it was attained just before the test audio signal was provided. The trigger for mixing the test audio signal in the processed audio signal may be some suitable event as the ones mentioned in Example 1, or some schedule requiring a periodic 'snapshot' of the conditioning of the TRRS, or, related, the tensioning of the tympanic membrane. In a preferred embodiment, the test audio signal comprises human speech, i.e. the test audio signal relates to a recording or a synthetically generated sample of one or more words spoken out loud by a human or human-like voice. In a preferred embodiment, the test audio signal relates to the same words and/or the same recording as the one used in calibration of the hearing device 1, i.e. the words used for recording at least one of the at least two pre-determined profiles. This leads to a more standardized and reliable way of determining a 'snapshot' of the current tensioning of the tympanic membrane, as the input provided to the listener is essentially the same as used for the determining of the profiles. In a related preferred embodiment, the test audio signal generator 80 is also the sound source used during the calibration, which leads to an advantageous combination of calibration means and means used during actual operation of the device.

In one embodiment (as shown), the signal generator 80 is controlled through the signal processor 68. In another embodiment (not shown), the signal generator 80 is controlled directly by the control unit 19.

Example 3 : Third device example

Figure 3 shows a third example embodiment of a hearing device 1 according to the present invention. The hearing device 1 is as described in the first example, but now includes means for binaural signal processing, whereby the signal processing circuit (7, 68, 57, 58, 19, 71, 72) comprises binaural output connection means 71 and binaural input connection means 72. Hence, the hearing device 1 is associated with a first of two ears of the listener, and comprises binaural output connection means 71 for receiving a second-ear-related signal provided by a second hearing device (not shown) that is associated with the second of said two ears of the listener. Furthermore, the hearing device 1 comprises binaural input connection means 72 for receiving a first-ear-related signal provided by the second hearing device.

The binaural output connection means 71 is adapted for receiving one or more signals from other components of the signal processing circuit (7, 68, 57, 58, 19, 71, 72), preferably through a wired connection, and combining one or more of said one or more signals into a first-ear-related signal that is provided to the second hearing device, preferably to a second binaural input connection means comprised in the second hearing device. Hereby, said providing to the second hearing device may be done by a wired connection as well as via a wireless data connection, preferably a wireless connection based on Bluetooth, ZigBee, NFC or IEEE 802.11. Said one or more signals may relate to the first input audio signal (as shown) and/or the second input audio signal (not shown) and/or the output provided by the input audio correlator 57 (as shown) and/or the output provided by the processed audio correlator 58 (as shown) and/or the reflected audio signal as provided by the reflected audio digitizer 35 (as shown with dotted line). The binaural input connection means 72 is adapted for receiving a second-ear-related signal from the second hearing device. Hereby, said receiving from the second hearing device may be done by a wired connection as well as via a wireless data connection, preferably a wireless connection based on Bluetooth, ZigBee, NFC or IEEE 802.11. In one embodiment, the second-ear-related comprises one or more signals from other components of the signal processing circuit (7, 68, 57, 58, 19, 71, 72) of the second hearing device, which may be essentially identical to the first hearing device. Said one or more signals (not shown) may relate to the first input audio signal and/or the second input audio signal and/or the output provided by the input audio correlator and/or the output provided by the processed audio correlator and/or the reflected audio signal as provided by the reflected audio digitizer.

In a preferred embodiment, the signals exchanged between the hearing device of the first ear and the hearing device of the second ear, i.e. the first-ear-related signal and/or the second-ear related signal, comprise at least the repeatedly determined phase difference audio spectra as determined by one of the input audio correlator 57 or the processed audio correlator 58 and preferably also the repeatedly determined amplitude difference audio spectra as determined by one of the input audio correlator 57 or the processed audio correlator 58, since these are particularly known to be indicative of the intentional focus of the listener. As explained in this document and e.g. in (B.C. J. Moore, An introduction to the psychology of hearing, Academic Press, 1997), particularly the phase difference may be indicative of binaural masking and binaural unmasking. Hereby, the listener may consciously or unconsciously position his ears so as to maximize a phase difference with respect to an attention-related frequency band, since maximizing the phase difference for this band makes that it is perceived louder. Hence, in a preferred embodiment, the detection of large inter-ear phase differences in some frequency band may be indicative of intentional focus with respect to said frequency band. This leads to the identification of said frequency band as being attention-related and may, in its turn, trigger a first action of increasing an amplitude of said attention-related frequency band by the signal processor 68 and/or a second action of increasing an inter-ear phase difference with respect to said frequency band by said signal processor 68, since both actions may contribute to the attention-related frequency band being perceived as louder.

In one embodiment (as shown), the binaural input connection means 72 is controlled through the signal processor 68. In another embodiment (not shown), the binaural input connection means 72 is controlled directly by the control unit 19. In a preferred embodiment (not shown), the binaural output connection means is controlled by the control unit 19. In one embodiment, as shown, the first hearing device comprises the entire signal processing circuit (7, 68, 57, 58, 19, 71, 72). Preferably, therefore, the signal processing circuit (7, 68, 57, 58, 19, 71, 72) is not duplicated in its entirety in the second hearing device. Hence, in a preferred embodiment, only one of the hearing devices comprises the "full" signal processing circuit (7, 68, 57, 58, 19, 71, 72) in the role of main device, and the other of the two devices takes up the role of peripheral device, possibly lacking one or more of the means or even the components of the signal processing circuit (7, 68, 57, 58, 19, 71, 72). In yet another embodiment (not shown), the signal processing circuit (7, 68, 57, 58, 19, 71, 72) is provided in the form of some external module integrated in an external device, such as an external controller, a mobile device, a smartphone, a smartwatch, a tablet, a laptop, etc. In such cases, both the first and the second hearing device, which are preferably essentially identical, are connected to the external module by means of conducting wire and/or by means of a digital wireless communication protocol such as Bluetooth, Zigbee, NFC or IEEE 802.11 (WiFi). Such an embodiment may be particularly advantageous where the device is e.g. a pair of two hearing aids, one per ear, which are wirelessly connected to a separate hearing aid controller. In another embodiment, the two hearing devices, one per ear, belong to a binaural headphone, whereby both hearing devices connected with a wire to a single controller outside of said two hearing devices but comprised in said binaural headphone.

Example 4: Fourth device example

Figure 4 shows a fourth example embodiment of a hearing device 1 according to the present invention. This example hearing device 1 is as described in Example 3, with the addition of a test audio signal generator 80 comprised in said signal processing circuit (7, 68, 57, 58, 19, 71, 72, 80). Like in Example 2, the test audio signal generator 80 is connected to the signal processor 68 and is adapted for providing a test audio signal to said signal processor 68. This allows to mix the test audio signal into said processed audio signal for enabling determining of said intentional focus level of said tympanic membrane, said mixing of said test audio signal preferably triggered by a schedule and/or by a detection of an event present in said one or more input signals. The advantages and preferred related embodiments are at least as described for Example 2. Furthermore, the test audio signal may be provided for only one of both ears or it may be provided alternatingly to one of the ears at a time, or it may be provided concurrently to both ears. Hereby, different from Example 2, it may be useful to add an inter-aural phase or amplitude difference, preferably an inter-aural phase difference, to the test audio signal, wherein said difference is preferably chosen in function of the transfer function measurements performed in both hearing devices, more preferably in function of at least the correlation values with respect to at least two pre-determined profiles for both ears.

Example 5: Acoustic processing example

Figure 5 illustrates example aspects of acoustic processing according to the present invention. This acoustic processing may relate to calibration such as the calibration described in Example 1, which relates to the determining of said at least two profiles. On the other hand, the acoustic processing may relate to the transfer function measurements according to the present invention, either in a "normal" mode (without test audio signal) or whereby the test audio signal is mixed into the processed audio signal.

Figure 5 shows a diagram of a difference level measurement in function of the frequency (in Hz). The difference level measurement relates to the measurement of a difference between two quantities associated with a certain time or time period. The two quantities may preferably be amplitudes or phases at a certain frequency or in a certain frequency band. Figure 5 illustrates a preferred embodiment of the invention, whereby the comparing of transfer function measurements, preferably the determining of amplitude difference audio spectra and/or phase difference audio spectra, relates to a measurement frequency band 91 comprising a lower measurement frequency A and an upper measurement frequency B, and a steering frequency band 92 comprising a lower steering frequency C and a higher steering frequency D. In a preferred embodiment, the lower measurement frequency A and the upper measurement frequency B represent end points of the measurement frequency band 91, e.g. by corresponding to -3 dB, -6 dB or -10 dB signal level drop with respect to a RMS signal level associated with the measurement frequency band 91. Likewise, in a preferred embodiment, the lower steering frequency C and the upper steering frequency D represent end points of the steering frequency band 92, e.g. by corresponding to -3 dB, -6 dB or -10 dB signal level drop with respect to a RMS signal level associated with the steering frequency band 92. Hereby, an advantageous choice of the upper measurement frequency B is smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz. Such a measurement frequency band 91 advantageously utilizes the detectability of the tensioning of the tympanic membrane at low frequencies with the providing of a series of words spoken out loud at higher frequency comprised in the steering frequency band, preventing any undesirable acoustic feedback loop between the "steering" of the tympanic membrane and the measurement of the tensioning of the membrane. Preferably, hereby, at least 50% of said steering frequency band 92 is situated above said upper measurement frequency B. More preferably, at least 80% of said steering frequency band 92 is situated above said upper measurement frequency B.

In one embodiment, the lower measurement frequency A is 20 Hz or lower than 20 Hz, the upper measurement frequency B is 200 Hz or 100 Hz or less than 100 Hz, the lower steering frequency C is 200 Hz and the higher steering frequency is 1000 Hz or 8000 Hz or more than 10 000 Hz. In another embodiment, the lower measurement frequency A is 20 Hz, the upper measurement frequency B is 200 Hz, the lower steering frequency C is 20 Hz or 100 Hz and the higher steering frequency is 1000 Hz or 8000 Hz or more than 10 000 Hz.

Claims

A method for adjusting an acoustic device (1), said acoustic device (1) adapted to be worn by a human listener in an ear canal, at an ear or in the ear of the listener and suitable for being connected to a signal processing circuit (7, 68, 57, 58, 19) adapted to process one or more input audio signals to provide a processed audio signal suitable for transmission by said acoustic device (1) as an audible signal to the listener; said method comprising the steps of:

- repeatedly determining a transfer function measurement associated with a tympanic membrane belonging to said ear of said listener, by comparing a reflected audio signal to said processed audio signal and/or at least one of said one or more input audio signals; said reflected audio signal based on reflected acoustic energy reflected by said tympanic membrane; said transfer function measurement comprising at least a phase-related measurement and preferably an amplitude measurement;

- repeatedly altering said processing in dependence on said transfer function measurement; characterized in that said method comprises the further step of:

- repeatedly determining a correlation value of said transfer function measurement with respect to each of at least two pre-determined profiles; preferably by assigning a plurality of weights with respect to said at least two profiles; said correlation value indicative of an intentional focus level of said tympanic membrane; wherein said at least two pre-determined profiles comprise at least one profile associated with a first tensioning of the membrane and low intentional focus and at least one profile associated with a second tensioning of the membrane and high intentional focus; and in that said step of repeatedly altering said processing is performed dependent on each of said correlation values, wherein said altering of said processing comprises at least the altering of an amplitude and/or a phase of at least one of said one or more input signals being processed.

Method according to claim 1, wherein said at least two pre-determined profiles comprise said at least one profile associated with said first tensioning of the membrane and low intentional focus with respect to a first auditory stream and said at least one profile associated with said second tensioning of the membrane and high intentional focus with respect to said first auditory stream; said at least one profile associated with said first tensioning of the membrane preferably further associated with high intentional focus with respect to said second auditory stream; and said second tensioning of the membrane preferably further associated with low intentional focus with respect to said first auditory stream.

3. Method according to claim 1, wherein said at least two pre-determined profiles comprise said at least one profile associated with said first tensioning of the membrane and said low intentional focus and at least two profiles associated with a third tensioning of the membrane and high intentional focus to a first auditory stream and a fourth tensioning of the membrane and high intentional focus to a second auditory stream different from said first auditory stream; said at least one profile associated with said first tensioning of the membrane preferably further associated with low intentional focus with respect to both said first and said second auditory stream.

4. Method according to claims 1-3, wherein said signal processing circuit (7, 68, 57, 58, 19) is comprised in said device (1).

5. Method according to claims 1-4, wherein said step of repeatedly altering said processing comprises mixing a test audio signal, preferably provided by a test audio signal generator (80) comprised in said signal processing circuit (7, 68, 57, 58, 19, 80), into said processed audio signal for enabling determining of said intentional focus level of said tympanic membrane, said mixing of said test audio signal preferably triggered by a schedule and/or by a detection of an event present in said one or more input signals.

6. Method according to claims 1-5, wherein said method comprises the further steps of:

- if said criterion is met, perform an action relating to said signal processing circuit (7, 68, 57, 58, 19), preferably entering a passive processing mode if said intentional focus level changes from high to low and/or entering an active processing mode if said intentional focus level changes from low to high.

7. Method according to claims 1-6, wherein said method comprises the further steps of:

- evaluating an increment of said correlation of at least one of said correlation values with respect to a previous correlation value for detecting an evolution of said correlation value;

8. Method according to claims 1-7, wherein said determining of said correlation value with respect to each of said at least at least two pre-determined profiles essentially relates to a measurement frequency band (91) with an upper measurement frequency (B), said upper measurement frequency (B) smaller than or equal to 200 Hz, more preferably smaller than or equal to 100 Hz.

9. Method according to claims 5-8, wherein said test sample audio signal essentially relates to a steering frequency band (92); wherein said determining of said correlation value with respect to each of said at least at least two pre-determined profiles essentially relates to a measurement frequency band (91) with an upper measurement frequency (B), said upper measurement frequency (B) preferably smaller than 200 Hz, more preferably smaller than 100 Hz; wherein at least 50% of said steering frequency band (92) is situated above said upper measurement frequency (B), preferably at least 80% of said steering frequency band (92) is situated above said upper measurement frequency (B).

10. Method according to claims 1-9, wherein said altering of said processing comprises applying a gain and/or a delay and/or a filter in one or more frequency bands and/or the increasing/decreasing of background noise filters and/or an adjusting between omnidirectional processing and directional processing.

11. Method according to claims 1-10, wherein at least one of said at least two predetermined profiles, preferably each of said at least two pre-determined profiles, is based on a prior transfer function measurement associated with a prior- measurement-related tympanic membrane belonging to an ear of a prior- measurement-related listener, said prior-measurement-related tympanic membrane preferably being said tympanic membrane and said prior- measurement-related listener preferably being said listener; wherein said prior transfer function measurement relates to a listening test comprising listening to a human or human-like voice.

12. The method according to claims 1-11, wherein said method is applied to a first tympanic membrane belonging to a first listener and a second tympanic membrane belonging to said first listener; wherein said transfer function measurement comprises a first measurement associated with said first tympanic membrane and a second measurement associated with said second tympanic membrane and essentially concurrent with said first measurement, and wherein said repeatedly altering of said processing is performed dependent on correlation values based on each of said first and said second measurement.

13. The method according to claim 12, wherein said repeatedly determining said correlation value comprises calculating a difference between said first measurement and said second measurement, and wherein said correlation value relates to a binaural correlation of said difference with respect to at least two pre- determined profiles being binaural profiles, said correlation value indicative of an intentional focus level of said first and said second tympanic membrane; wherein said at least two pre-determined binaural profiles comprise at least one binaural profile associated with a first tensioning of the first and the second membrane and low intentional focus and at least one binaural profile associated with a second tensioning of the first and the second membrane and high intentional focus.

14. A signal processing circuit (7, 68, 57, 58, 19) for adjusting an acoustic device (1), said acoustic device (1) adapted to be worn by a human listener in an ear canal, at an ear or in the ear of the listener; said signal processing circuit (7, 68, 57, 58, 19) adapted to process one or more input audio signal to provide a processed audio signal suitable for transmission by said acoustic device (1) as an audible signal to the listener; said signal processing circuit (7, 68) comprising :

- a first correlator (57, 58) configured to process a reflected audio signal based on reflected acoustic energy reflected by a tympanic membrane belonging to said ear of said listener; said first correlator (57, 58) configured for repeatedly determining, by comparing said reflected audio signal to said processed audio signal and/or at least one of said one or more input audio signals, a transfer function measurement associated with said tympanic membrane; said transfer function measurement comprising an amplitude measurement, a group delay measurement and preferably a phase measurement; - a control unit (19) configured to alter said processing in dependence on said transfer function measurement; characterized in that said first correlator (57, 58) is further configured for: - correlating said transfer function measurement to at least two predetermined profiles whereof at least one profile is associated with a first tensioning of the membrane and low intentional focus and at least one profile is associated with a second tensioning of the membrane and high intentional focus; - repeatedly determining a correlation variable of said transfer function measurement with respect to each of said at least two pre-determined profiles for determining an intentional focus level of said tympanic membrane, preferably by assigning a plurality of weights with respect to said at least two profiles; and in that said control unit (19) is further configured for: performing said altering of said processing at least partly in function of each of said correlation variables, wherein said altering of said processing comprises at least the altering of a group delay of said at least one of said one or more input signals being processed.

15. An acoustic device (1) comprising the signal processing circuit (7, 68) according to claim 14, said acoustic device (1) further comprising :

- an input means (3, 4, 5, 6) for providing one or more input audio signals to said signal processing circuit; said input means preferably adapted for providing at least one input audio signal dependent on acoustic energy from an external acoustic input;

- a means for transmitting acoustic energy (10) adapted to provide an audible signal to the listener in dependence of the processed audio signal provided by said signal processing circuit (7, 68, 57, 58, 19);

- a means for receiving reflected acoustic energy (33, 34, 35) reflected by a tympanic membrane belonging to said ear of said listener, said means for receiving reflected acoustic energy providing a reflected audio signal;

16. Device (1) according to claim 15, whereby said device is any or any combination of the following : a hearing aid, a middle ear implant, a cochlear implant, a bone- anchored hearing aid, a bone-conducting microphone.

17. Device (1) according to claims 15-16, whereby the device (1) is configured to perform the method according to any of the claims 1 to 13.

18. Use of the signal processing circuit (7, 68, 57, 58, 19) according to claim 14 in a hearing aid.