WO2021185891A1 - Système et procédé de compensation de l'effet d'occlusion dans des écouteurs ou des prothèses auditives, avec perception améliorée de sa propre voix - Google Patents

Système et procédé de compensation de l'effet d'occlusion dans des écouteurs ou des prothèses auditives, avec perception améliorée de sa propre voix Download PDF

Info

Publication number
WO2021185891A1
WO2021185891A1 PCT/EP2021/056789 EP2021056789W WO2021185891A1 WO 2021185891 A1 WO2021185891 A1 WO 2021185891A1 EP 2021056789 W EP2021056789 W EP 2021056789W WO 2021185891 A1 WO2021185891 A1 WO 2021185891A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
product
input
controller
loudspeaker
Prior art date
Application number
PCT/EP2021/056789
Other languages
German (de)
English (en)
Inventor
Stefan Liebich
Johannes Fabry
Raphael Brandis
Original Assignee
Rheinisch-Westfälische Technische Hochschule (Rwth) Aachen
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Rheinisch-Westfälische Technische Hochschule (Rwth) Aachen filed Critical Rheinisch-Westfälische Technische Hochschule (Rwth) Aachen
Publication of WO2021185891A1 publication Critical patent/WO2021185891A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R1/00—Details of transducers, loudspeakers or microphones
    • H04R1/10—Earpieces; Attachments therefor ; Earphones; Monophonic headphones
    • H04R1/1016—Earpieces of the intra-aural type
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R25/00—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception
    • H04R25/50—Customised settings for obtaining desired overall acoustical characteristics
    • H04R25/505—Customised settings for obtaining desired overall acoustical characteristics using digital signal processing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R2460/00—Details of hearing devices, i.e. of ear- or headphones covered by H04R1/10 or H04R5/033 but not provided for in any of their subgroups, or of hearing aids covered by H04R25/00 but not provided for in any of its subgroups
    • H04R2460/05—Electronic compensation of the occlusion effect

Definitions

  • the present invention relates to a digital audio signal processing system and a method for digital audio signal processing for headphones or a hearing aid to compensate for the occlusion effect generated by the headphones or the hearing aid, the headphones or the hearing aid having a loudspeaker for emitting acoustic signals into the ear canal human ear and at least one internal microphone for receiving acoustic signals from the auditory canal, and the system is set up to process an error signal generated by the internal microphone and to generate a loudspeaker signal for controlling the loudspeaker, which is a compensation signal determined as a function of the error signal or contains.
  • FIG. 1 illustrates these sound components when wearing headphones 1 in the form of so-called in-ear headphones during conversations.
  • the occlusion also causes noise when swallowing, chewing or footfall noise when walking and running.
  • a measurement of the occlusion effect is usually carried out by a relation between the signal of the inner and an outer microphone.
  • These microphones can either be integrated in the headphones or the hearing aid or connected to them.
  • the inner microphone for measuring the occlusion effect can, for example, be a so-called tube microphone, as is used in hearing aids.
  • the outer microphone can be arranged in its own housing on the outside of the auricle.
  • FIG. 2 shows the basic structure of headphones 1 according to the prior art with a loudspeaker 2, an inner microphone 3, an outer microphone 4, which are arranged integrally in a housing 8.
  • the headphones 1 are arranged in the auditory canal 6 of a human ear 5 (auricle), at the end of which the eardrum 7 lies.
  • the outside of an ear pad 9 of the headphones 1 rests partially or completely on the inside of the auditory canal 6, so that the ear pad 9 closes the auditory canal 6.
  • the inner microphone 3 supplies an electrical microphone signal e (t), hereinafter referred to as the error signal.
  • the external microphone 4 supplies an electrical microphone signal x (t), hereinafter referred to as the external signal.
  • the loudspeaker 2 is controlled with an electrical loudspeaker signal u (t).
  • the relation between the inner and the outer microphone signal is described by the occlusion function OE (f) and is calculated according to the following equation from the ratio of the amounts of the Fourier transforms of the error signal e (t) and the external signal x (t) where £ (/) is the Fourier transformed frequency domain representation of the error signal e (t) and X (f) is the Fourier transformed frequency domain representation of the error signal e (t).
  • FIG. 3 shows the course of the occlusion function OE (f) on the basis of measurements in 23 people. A gain (Ampi.) At low frequencies and a weakening (Att.) At high frequencies can be seen.
  • the loudspeaker signal u (t) is or contains a compensation signal.
  • a first signal component for the compensation signal is generated, which corrects the structure-borne noise component.
  • a second signal component is generated for the compensation signal, which corrects the airborne sound components.
  • the latter is optional.
  • the electrical or analog, time-continuous signals x (t), u (t) and e (t) are processed digitally, so that AD converter 13 for the microphone signals x (t), e (t) or a DA converter 14 can be used for the loudspeaker signal u (t) in order to convert the analog signals into digital signals and vice versa, see FIG. 4.
  • the transmission from the outer microphone 4 to the inner microphone 3 is the so-called primary path 10, which is described by a transfer function P (z).
  • the transmission from the loudspeaker 2 to the inner microphone 3 is the so-called Secondary path 11, which is described by a transfer function G (z).
  • the transmission from the loudspeaker 2 to the external microphone 4 is the feedback path 12, which is described by a transfer function F (z).
  • FIG. 5 illustrates the acoustic transmission paths 10, 11, 12 in a block diagram in the form of gray boxes.
  • the error signal e (n) is filtered by means of a first digital filter 15 with the negated transfer function -K (z) in order to obtain the first signal component ux (n) and the external signal x (n) by means of a second digital filter 16 with the negated transfer function -W (z) filtered in order to obtain the second signal component uw (n).
  • the filters 15, 16 are shown in the block diagram with white boxes.
  • the first filter 15 Since the first filter 15 is located in a signal path 19 via which the error signal e (n), which also contains the signal acoustically received from the loudspeaker 2 on the inner microphone 3, is sent to the loudspeaker, this signal path 19 forms a feedback path of a control loop , in which the first filter 15 forms a feedback regulator.
  • the two signal components ui ⁇ (n), uw (n) are added up to generate the compensation signal u (n). This is played over the loudspeaker 2 in order to carry out compensation. Measurements have shown that the influence of the feedback path 12 on the signal x (n) of the outer microphone 4 is small, so that the feedback path 12 can be neglected.
  • [Liebich 2018] also discloses a weighting of the output signal of the feedback controller 15 with an adaptive factor in order to increase the robustness or stability of the control loop for all possible scenarios. This has the advantage that when designing the controller, it is possible to focus on the usual applications, i.e. the special cases that may impair the stability of the controller can be disregarded when designing the controller. This optimizes the performance of the controller. With the scaling factor, [Liebich 2018] deals with special cases that could lead to instability of the controller.
  • the scaling factor is calculated as a function of the error signal and an intermediate signal determined from the controller output signal, the intermediate signal being the convolution product of the controller output signal and a function describing the secondary path, or in other words the result of filtering the controller output signal with an estimated secondary path.
  • the scaling factor is calculated by forming the product of the error signal and the intermediate signal, recursively smoothing this product, then normalizing the smoothed product by dividing it by a denominator, and mapping the normalized product to a positive value range around which To obtain scaling factor.
  • the denominator is the product of the roots of estimated autocorrelations of the error signal on the one hand and the intermediate signal on the other. This calculation is comparatively complicated and difficult to implement in digital signal processing. In addition, the calculation takes time, so that there are only limited dynamics when calculating the factor.
  • the object of the present invention is on the one hand to simplify the calculation rule for the adaptive factor or to replace it with a new, less complex adaptation rule.
  • a digital audio signal processing system for headphones or a hearing aid to reduce the occlusion effect generated by the headphones or hearing aid having a loudspeaker for emitting acoustic signals into an ear canal and at least one internal microphone for receiving acoustic signals from the ear canal, and the system is set up, one from the inner one Process microphone-generated error signal and generate a loudspeaker signal for controlling the loudspeaker, which is or contains a compensation signal determined as a function of the error signal.
  • the system includes:
  • a feedback controller for digitally filtering a controller input signal which is or contains the error signal, a controller output signal of the feedback controller being weighted with at least one time-variant scaling factor and the weighted controller output signal forming or at least partially forming the compensation signal;
  • an adaptation unit for dynamically determining the scaling factor as a function of at least a first and a second input signal of the adaptation unit, wherein the first input signal is or contains the error signal and the second input signal is or contains a first intermediate signal determined from the controller output signal, and wherein the adaptation unit is set up ,
  • a processing unit located between the feedback controller and the adaptation unit for determining the intermediate signal from the controller output signal or from a signal containing this controller output signal, the processing unit having a transfer function describing the acoustic transfer characteristics from the loudspeaker to the internal microphone; wherein the processing unit is set up to form the denominator either from a recursively smoothed absolute amount of the product of the input signals or from a product of the respective recursively smoothed absolute amounts of the individual input signals.
  • a method for digital audio signal processing for headphones or a hearing aid to reduce the occlusion effect generated by the headphones or the hearing aid is also proposed, in which a loudspeaker signal for controlling a loudspeaker of the headphones or hearing aid is determined as a function of an error signal generated by an internal microphone of the headphones or hearing aid, the loudspeaker signal being or containing a compensation signal,
  • a controller input signal of a feedback controller is digitally filtered by the latter, the controller input signal being or containing the error signal,
  • a controller output signal of the feedback controller is weighted with a time variant scaling factor and the weighted controller output signal forms the compensation signal or at least partially forms it,
  • the scaling factor is dynamically determined by an adaptation unit as a function of at least a first and a second input signal of the adaptation unit, wherein the first input signal is or contains the error signal and the second input signal is or contains an intermediate signal determined from the controller output signal,
  • the intermediate signal is determined by a processing unit from the controller output signal or from a signal containing it, the processing unit having a transfer function describing the acoustic transfer characteristics from the loudspeaker to the internal microphone, the denominator either from a recursively smoothed absolute value of the product of the input signals or from a Product of the recursively smoothed absolute amounts of the individual input signals is formed.
  • the determination of the denominator for normalization is significantly less complex and less expensive to implement than in the prior art, since only one or two multiplications have to take place and no root formation is required.
  • the input signals are first multiplied and the amount of the product obtained is formed, which is then recursively smoothed. Since the product of the first and the second input signal has already been calculated for the numerator of the division - the numerator corresponds to the recursively smoothed product of the input signals - and is therefore already available, the denominator formation in the first variant does not require any additional multiplication.
  • a product is not formed until after the recursive smoothing, the absolute value of each of the two input signals being formed and smoothed. These recursively smoothed absolute amounts are then multiplied with one another.
  • the second embodiment variant thus has an additional multiplication as well as two absolute value formations and two recursive smoothing in the denominator compared to the first embodiment variant.
  • the formation of the absolute value for the denominator is a core aspect of the scaling factor determination according to the invention. On the one hand, this ensures that the denominator is always positive.
  • the formation of the absolute value means that the normalized product receives values between -1 and 1. Values less than 0 are achieved when a negative correlation between the compensation signal and the instantaneous error signal has been established. In this case destructive interference takes place and the closed feedback control loop leads to an attenuation in the error signal. This occurs especially with sounds that cause a strong occlusion effect (e.g. closed vowels). Values greater than zero are achieved when a positive correlation between the compensation signal and the instantaneous error signal has been established.
  • the mapping to the positive value range can be formed by negating the normalized product and adding the value 1 to the normalized and negated product. This ensures that the scaling factor is mapped to a value range between 0 and 2.
  • the partial interval between 0 and 1 has a weaker effect of the feedback controller, ie it is less attenuated.
  • a value of the scaling factor in the sub-interval between 1 and 2 leads to an increase in the controller effect, ie it is more attenuated.
  • the respective recursive smoothing can take place by means of low-pass filtering, for example by means of a first-order low-pass filter.
  • the product of the input signals in the numerator and / or the absolute amount of the product of the input signals in the denominator in the first embodiment variant or the absolute amounts of the individual input signals in the denominator in the second embodiment variant can be low-pass filtered.
  • q (n) ßq (n - 1) + (1 - ß) p (n), where q (n) is the output value of the respective smoothing / low-pass filtering at the current time step n, q (n- 1) the output value of the smoothing in the last time step n-1, p (n) is the input value of the respective smoothing / low-pass filtering in the current time step n and ⁇ is a smoothing constant.
  • ß e.g. 0.5
  • the smoothing constant is preferably between 0.9 and 0.9999, depending on the sampling rate used to calculate the adaptive factor, so that the change in the scaling factor very slowly follows the changes in the input signals.
  • the denominator can additionally have a constant that corresponds to the recursively smoothed absolute value of the product of the input signals in the first Embodiment or is added to the product of the respective recursively smoothed absolute amounts of the individual input signals in the second embodiment.
  • This constant prevents large numbers from occurring in the case of numerically small denominators of division, in particular a division by zero if the denominator is so small that it cannot be represented by the digital signal processing system or is considered NaN (Not a Number).
  • the constant thus forms a regularization constant. The signal processing system is thereby protected.
  • the constant should also be small.
  • the constant can be less than IO -3 ... preferably none than 10 -6 .
  • the headphones or a hearing aid additionally has an external microphone for picking up acoustic signals from the environment outside the auditory canal
  • the system has a forward filter in order to receive an external signal generated by the external microphone or an external signal to filter the signal containing and through this filtering to generate a forward component to the compensation signal, which is added to the weighted controller output signal in order to obtain the compensation signal or at least to form it proportionally.
  • a further processing unit can preferably be present, to which the forward component or a sum signal containing this component is fed and which has a negated transfer function describing the acoustic transmission characteristics from the loudspeaker to the internal microphone in order to generate a second intermediate signal, the controller input signal being the sum from the error signal and the second intermediate signal is or contains.
  • the first and / or the second input signal of the adaptation unit Bandpass filter or a bandpass filter can be connected upstream in order to filter the error signal or the signal containing the error signal and / or to filter the first intermediate signal or the signal containing the first intermediate signal.
  • the scaling factor can address a specific frequency range so that the feedback control loop can only be scaled in this range using the adaptive factor for the currently present noises.
  • a time variant scaling can be carried out when impact sound occurs, which is more clearly perceptible in the case of an occlusion in the ear.
  • the bandpass filter can have a frequency range of 20 Hz to 200 Hz, for example.
  • the scaling factor can be aligned specifically to the voice in that the bandpass filter has a frequency range of 100 Hz to 2 kHz.
  • the system can have an input for an external useful signal and an input filter to which the external useful signal is fed for filtering in order to generate an audio signal that is added to the forward component or the controller output signal.
  • an external audio source such as music or a voice call.
  • the system can have an output for providing a microphone signal there and an output filter connected to this output, to which the error signal or a signal containing it is fed in order to generate the microphone signal.
  • the invention further relates to headphones or a hearing aid with a loudspeaker for emitting acoustic signals into an auditory canal and at least one internal microphone for receiving acoustic signals from the auditory canal, the headphones or hearing aid comprising a digital audio signal processing system according to the invention.
  • FIG. 6 shows a digital audio signal processing system for the headphones 1 or the hearing aid 1 in FIGS. 2 or 4 to compensate for the occlusion effect generated by the headphones 1 or the hearing aid 1, the headphones 1 or the hearing aid 1 having a loudspeaker 2 for emitting acoustic signals in an auditory canal 6 and at least one internal microphone 3 for recording acoustic signals from the auditory canal 6.
  • the system processes an error signal e (n) recorded by the internal microphone 3 and generates a loudspeaker signal for controlling the loudspeaker 2, which in this embodiment is a compensation signal (u (n)) determined as a function of the error signal e (n).
  • the system includes a feedback controller 15 with time-variant scaling for adaptation to the speech activity of the wearer of the headphones 1 or hearing aid 1. More precisely, the system has a feedback controller 15 for digitally filtering a controller input signal, which in this embodiment variant contains the error signal e (n), wherein a controller output signal Uk (n) of the feedback controller 15 is weighted, ie multiplied, with at least one time variant scaling factor a (n), and the weighted controller output signal Uk (n) in this embodiment variant is the compensation signal u ( n) forms which is reproduced via the loudspeaker 2.
  • the feedback controller 15 can be designed as described in [Liebich 2018], for example. An exemplary design for the feedback controller 15 with the transfer function K (z) is shown in FIG.
  • An adaptation unit 21 of the system dynamically defines the scaling factor a (n) as a function of a first and a second input signal of the adaptation unit 21, the first input signal in this embodiment variant being the error signal e (n) and the second input signal in this embodiment variant being an intermediate signal Uk '(n), which is determined from the controller output signal Uk (n).
  • a signal processing unit 22 located between the feedback controller 15 and the adaptation unit 21, which generates a transfer function G (z) describing the acoustic transfer characteristics from the loudspeaker 2 to the internal microphone 3. possesses, ie has the transmission characteristics of the secondary path 10 described above.
  • the adaptation unit 21 is also a filter.
  • the adaptation unit 21 receives at the input on the one hand the actual microphone signal e (n), which here corresponds to the compensation signal u (n) received on the inner microphone 3, which contains the scaling of the controller output signal Uk (n), and on the other hand a theoretical microphone signal that this Does not include scaling. This makes it possible to change the scaling factor in a targeted manner in such a way that the intensity of the action of the controller 15, ie its damping or amplification, can take place as a function of the variation in one's own voice.
  • the transfer function G (z) is estimated, the estimate G (z) of the secondary path 10 reflecting a representative estimate of the current situation. It can, for example, have been determined from a measurement made before the system was used or from an averaging of several measurements of the secondary path 10. Alternatively, the estimate G (z) can also be determined or changed during the intended use of the system by measuring the secondary path.
  • the mode of action of the adaptation of the scaling factor according to the invention to one's own voice is based on knowledge about the articulated sounds.
  • Different sounds cause occlusion effects of different strengths, which can be represented, for example, by frequency-independent occlusion functions 1) E as a gain or attenuation value.
  • FIG. 9 in which the gain and attenuation values of various frequency-independent sounds are shown in comparison in two diagrams.
  • different frequency-dependent courses of ⁇ £ (f) can be determined for different sounds.
  • the diagram on the left in FIG. 9 compares the openness of 16 vowels, whereas the diagram on the right compares the voicing of 13 consonants. This is done opposite the measurable frequency-independent occlusion effect ⁇ ⁇ E.
  • E When measuring the occlusion effect 1) E, a variation of the voice depending on the articulated sound can be determined.
  • the vowels i, u and y have a very high gain of approx. 28 dB, while the vowel a has only a low gain of approx. 4 dB.
  • the feedback controller 15 can dampen more, whereas in the case of a low gain it can only dampen a little.
  • the time variant adaptation of the scaling factor a (n) and thus the feedback loop is carried out by the adaptation unit 21 to the current speech situation.
  • the scaling factor a (n) can go in the direction of its maximum value for the purpose of greater attenuation, whereas it goes in the direction of 1 in the case of a low gain.
  • a first variant of the adaptation unit 21 is shown in FIG. It comprises a first signal path for determining a numerator for division 25 and a second signal path for determining the denominator for division 25.
  • the product of the first and second input signals UK '(n), e (n) is fed to both signal paths, which is formed by a multiplication 23 at the beginning in the adaptation unit 21.
  • a smoothing unit 24 in the form of a low-pass filter, the product is recursively smoothed in the first signal path in order to obtain the counter.
  • the output signal of the smoothing unit 24 represents an estimate of the cross-correlation of the input signals UK ‘(n), e (n), which is then normalized by dividing 25 by the denominator.
  • a recursive smoothing of the product of the first and the second input signal UK '(n), e (n) is also provided by means of a smoothing unit 24 in the form of a low-pass filter; Amount of product forms.
  • the denominator is thus formed from the recursively smoothed absolute value of the product of the input signals.
  • a regularization constant e is added to the output signal of the smoothing unit 24 in the second signal path to be very small Avoid values in the denominator ( ⁇ 10 3 ) or division by zero.
  • the smoothing units in the first and second signal path are of identical design. However, they can be different in another embodiment variant.
  • the input value of the smoothing unit 24 is the product of the input signals ui ⁇ '(n), e (n) of the adaptation unit 21, and the output value is the counter of the division 25.
  • the input value of the smoothing unit 24 is the absolute value of the product of the input signals ui ⁇ '(n), e (n), and the output value a part of the denominator of division 25.
  • the smoothing constant ⁇ is 0.99, for example, so that the output value follows the input value very slowly. 99% of the current output value is formed from the previous output value and only 1% from the current input value. Signal peaks are thereby filtered out.
  • the division 25 represents a normalization of the cross-correlation or the smoothed product of the input signals ui ⁇ '(n), e (n).
  • the values of the cross-correlation or the output values of the smoothing unit 24 are normalized in the first signal path, more precisely mapped to the value range -1 and +1.
  • the sign of the output value of division 25 depends solely on the numerator.
  • the values of the normalized smoothed product of the input signals are negated after division 25 and added to 1, or in other words subtracted from 1, in order to obtain the scaling factor a (n).
  • In opposite-phase input signals (a signal ⁇ 0) now generate values for the scaling factor a (n) between 1 and 2. This means that the effect of the feedback controller 15 is increased.
  • in-phase input signals both signals ⁇
  • the fed back feedback controller 15 leads to an amplification. If a sound predominantly generates energy in this frequency range (for example in the case of unvoiced fricatives), then the feedback controller 15 generates an in-phase compensation signal compared to the noise present in the auditory canal.
  • the processing unit 21 leads to a scaling factor a (n) between 0 and 1 and weakens the effect of the feedback controller 15.
  • instability of the control loop generally leads to tonal interference, which is typically present in the amplification frequency range.
  • the processing unit 21 also feeds in this case a scaling factor a (n) between 0 and 1 and weakens the effect of the feedback controller 15.
  • q2 (ri) ßq2 (n - 1) + (1 - ß) ⁇ u K ' (n) e (n)
  • a second variant of the adaptation unit 21 is shown in FIG. It differs from the first embodiment variant in FIG. 7 only in that in the second signal path for forming the denominator the multiplication 23 only takes place after the recursive smoothing 24.
  • the second signal path thus consists of two subpaths, the magnitude of the first input signal UK '(n) being formed and recursively smoothed in a first subpath and the magnitude of the second input signal e (n) being formed and recursively smoothed in a second subpath parallel to it .
  • the recursive smoothing carried out with the corresponding smoothing unit 24 takes place, as in the first embodiment variant, in the form of low-pass filtering.
  • the output value of the respective smoothing unit in the subpath forms an estimate of the autocorrelation of the amount of the respective input value.
  • the product of the output values of the first and second sub-path is then formed by a multiplication 25 and added to the regularization constant, the sum obtained in this way forming the denominator for the division 25 in order to obtain the estimated cross-correlation of the input signals ui ⁇ '(n), e (n) normalize.
  • the processing unit 22 is thus set up to form the denominator for the division 25 from a product of the respective recursively smoothed absolute values of the individual input signals ui ⁇ ‘(n), e (n). Otherwise, the second variant is identical to the first variant.
  • This second embodiment variant calculates a quantitatively comparable, but not identical, scaling factor a (n).
  • the recursive smoothing of the two individual signals ui ⁇ '(n) and e (n) results in a single signal in the second embodiment variant
  • Estimation of the autocorrelations In the first embodiment, however, no autocorrelations are estimated, but only an estimation of the cross-correlation of the absolute value of the input signals takes place.
  • FIGS. 10 and 11 show possible combinations between a feedback control loop with controller 15 and forward filtering by means of filter 16.
  • FIG. 10 shows a system according to FIG. 6, which additionally has forward filtering of an external signal x (n) picked up by an external microphone 4, which is filtered by a forward filter 16 with the transfer function -W (z) by a forward component u (n) for the compensation signal u (n), which is added to the weighted controller output signal UKa (n), in order to obtain the compensation signal u (n).
  • the weighted controller output signal UKa (n) then forms the feedback component for the compensation signal u (n). Both components uw (n) and UKa (n) are thus summed up in order to generate the compensation signal u (n), which is emitted via the loudspeaker 2.
  • the time variant factor a (n) is set again via an adaptation unit 21 which, on the basis of the error signal e (n) and the intermediate signal UK ‘(n), calculates the scaling factor a (n) as in the above-mentioned first or second embodiment variant.
  • the forward filter 16 can be designed as described in the German patent application DE 102016011719 A1, which is hereby fully incorporated by reference.
  • An exemplary design for the forward filter 16 with the transfer function W (z) is shown in FIG.
  • the further embodiment variant shown in FIG. 11 develops the variant in FIG. 10 in such a way that the input of the feedback controller 15 is corrected as a function of the forward component u (n).
  • the system here includes one further processing unit 22 'to which the forward component u (n) is fed and which has a negated transfer function -G (z) describing the acoustic transfer characteristics from loudspeaker 2 to internal microphone 3, ie simulating secondary path 11 in the form of an estimate, to generate a second intermediate signal e (n).
  • the further processing unit 22 ' is identical to the processing unit 22 in FIG. 6, with only the sign being reversed.
  • the second intermediate signal e (n) is used to correct the error signal e (n) in the feedback control loop by addition.
  • the controller input signal of the feedback controller 15 is then formed by the sum e (n) of the error signal e (n) and the second intermediate signal e (n). Otherwise, the system is designed as described above.
  • FIGS. 12 and 13 show two further embodiment variants of the invention which differ from the previous variants in that the input signals UK ‘(n), e (n) of the adaptation unit 21 are first filtered by a bandpass filter 27 in each case.
  • the variant in FIG. 12 corresponds to the variant in FIG. 6 and the variant in FIG. 13 to the variant in FIG. 11.
  • the bandpass filters 27 for the two input signals UK ‘(n), e (n) are identical in these variants. For example, they have a frequency range of 100 Hz to 2 kHz corresponding to the frequency range of the human voice in order to adapt the control to the human voice.
  • the bandpass filters 27 can be between 20 Hz and 200 Hz for the attenuation of footfall sound.
  • the system and method according to the invention can be used in combination with an external audio source and / or external signal processing, as the exemplary embodiment in FIG. 14 shows.
  • This exemplary embodiment develops the variant in FIG. 11 according to a first aspect in that the system has an input 28 for an external useful signal a (n) and an input filter 29 with a transfer function C (z) to which the external useful signal a (n) for filtering is fed to a To generate audio signal a '(n), which is added to the forward component u (n) or is.
  • the signal fed to the further processing unit 22 ' corresponds to the sum signal from the forward component u (n) and the audio signal a' (n).
  • the sum signal from u w (n) and a '(n) is filtered with the estimate G (z) of the secondary path 11 and generates the correction signal e (n).
  • This correction signal e (n) is used as before in order to correct the error signal e (n) of the inner microphone in the feedback control loop by addition.
  • the preprocessing of the external useful signal a (n) via the input filter 29 takes into account, for example, the properties of the loudspeaker 2 in order to generate the signal a '(n).
  • the useful signal can be music or incoming voice signals from a call, for example.
  • the exemplary embodiment in FIG. 14 further develops the variant in FIG. 11 according to a second aspect in that the system has an output 30 for providing a microphone signal v (n) there and an output filter 31 connected to this output 30 with a transfer function L (z) to which a signal e (n) containing the error signal e (n) is fed in order to generate the microphone signal v (n).
  • the signal fed to the output filter 31 is the input signal e (n) of the feedback controller 15, which is derived from the forward component u (n) and the audio signal a '(n) by filtering with the estimate G (z) of the secondary path 11 comprises correction signal e (n).
  • the microphone signal v (n) provided at the output 30 is thus the corrected error signal e (n) of the internal microphone 2 filtered by the output filter 31.
  • FIG. 17 shows the influence of the adaptive scaling factor a (ri) on the amount of the closed transfer function of the feedback control loop.
  • This transfer function often known as sensitivity, is calculated as: l
  • FIG. 18 The behavior of the scaling factor a (n) in the event of speech stimulation is such that it reacts to the current situation at the inner microphone 3 and thus in the auditory canal 6.
  • Figure 18 and Figure 19 show a voice recording with the inner microphone 3 in different modes (upper plot), as well as the associated behavior of the time-variant scaling factor a (n) (lower plot).
  • a (n) 0
  • the dark curve d (n) of the signal of the inner microphone 3 results in the illustration above.
  • the inventive use of the adaptive scaling factor a (n) allows the feedback control loop to be adapted to the variation of one's own voice, to the variation of noises caused by structure-borne noise (e.g. impact sound, chewing, swallowing, etc.) or to the variation of Reach residual interference signals from the environment in the ear canal 6.
  • the new Adaptation rule for the adaptive scaling factor a (n) that is to say the signal processing carried out by the adaptation unit 21, is less complex compared to the prior art.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Acoustics & Sound (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Neurosurgery (AREA)
  • Otolaryngology (AREA)
  • Soundproofing, Sound Blocking, And Sound Damping (AREA)
  • Circuit For Audible Band Transducer (AREA)

Abstract

L'invention se rapporte à un procédé et à un système de traitement de signal audio numérique pour un écouteur (1) ou un dispositif auditif (1), lesquels procédé et système sont destinés à réduire l'effet d'occlusion généré par l'écouteur (1) ou le dispositif auditif (1). Le système comprend un dispositif de commande à rétroaction (15) pour filtrer un signal provenant d'un microphone interne (3), et une unité d'adaptation (22), un signal de sortie de dispositif de commande (uk(n)) du dispositif de commande à rétroaction (15) étant pondéré par un facteur de mise à l'échelle variant dans le temps (α(n)), et le signal de sortie de dispositif de commande pondéré (uk(n)) formant ou formant partiellement un signal de compensation (u(n)), et l'unité d'adaptation (21) définissant les facteurs de mise à l'échelle (α(n)) en fonction d'un premier et d'un second signal d'entrée, ladite unité formant un produit du premier et du second signal d'entrée (uk'(n), e(n)) et lissant de manière récursive ledit produit, puis divisant le produit lissé par un dénominateur pour une normalisation, et mappant le produit normalisé sur une plage de valeurs positives. Le dénominateur est formé soit à partir d'une valeur absolue lissée de manière récursive du produit des signaux d'entrée (uk'(n), e(n)) soit à partir d'un produit des valeurs absolues respectives lissées de manière récursive des signaux d'entrée individuels (uk'(n), e(n)).
PCT/EP2021/056789 2020-03-19 2021-03-17 Système et procédé de compensation de l'effet d'occlusion dans des écouteurs ou des prothèses auditives, avec perception améliorée de sa propre voix WO2021185891A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102020107620.7A DE102020107620B3 (de) 2020-03-19 2020-03-19 System und Verfahren zur Kompensation des Okklusionseffektes bei Kopfhörern oder Hörhilfen mit verbesserter Wahrnehmung der eigenen Stimme
DE102020107620.7 2020-03-19

Publications (1)

Publication Number Publication Date
WO2021185891A1 true WO2021185891A1 (fr) 2021-09-23

Family

ID=75143613

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2021/056789 WO2021185891A1 (fr) 2020-03-19 2021-03-17 Système et procédé de compensation de l'effet d'occlusion dans des écouteurs ou des prothèses auditives, avec perception améliorée de sa propre voix

Country Status (2)

Country Link
DE (1) DE102020107620B3 (fr)
WO (1) WO2021185891A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4304201A1 (fr) * 2022-07-08 2024-01-10 GN Audio A/S Dispositif audio doté d'annulation de bruit adaptatif et écoute active

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3005731A1 (fr) * 2013-06-03 2016-04-13 Sonova AG Procédé de fonctionnement d'un dispositif auditif et dispositif auditif
DE102016011719B3 (de) 2016-09-30 2017-09-07 Rheinisch-Westfälische Technische Hochschule Aachen Aktive Unterdrückung des Okklusionseffektes in Hörhilfen

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3005731A1 (fr) * 2013-06-03 2016-04-13 Sonova AG Procédé de fonctionnement d'un dispositif auditif et dispositif auditif
DE102016011719B3 (de) 2016-09-30 2017-09-07 Rheinisch-Westfälische Technische Hochschule Aachen Aktive Unterdrückung des Okklusionseffektes in Hörhilfen
EP3520441A1 (fr) * 2016-09-30 2019-08-07 Rheinisch-Westfälische Technische Hochschule Aachen Suppression active de l'effet ocklusion d'une prothèse auditive

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
LIEBICH ET AL.: "Active Occlusion Cancellation with Hear-Through Equalization for Headphones", INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP, 2018, pages 241 - 245, XP033401158, ISBN: 978-1-5386-4658-8, DOI: 10.1109/ICASSP.2018.8461834
LIEBICH STEFAN ET AL: "Active Occlusion Cancellation with Hear-Through Equalization for Headphones", 2018 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), IEEE, 15 April 2018 (2018-04-15), pages 241 - 245, XP033401158, DOI: 10.1109/ICASSP.2018.8461834 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4304201A1 (fr) * 2022-07-08 2024-01-10 GN Audio A/S Dispositif audio doté d'annulation de bruit adaptatif et écoute active

Also Published As

Publication number Publication date
DE102020107620B3 (de) 2021-07-29

Similar Documents

Publication Publication Date Title
EP1143416B1 (fr) Suppression de bruit dans le domaine temporel
DE69931580T2 (de) Identifikation einer akustischer Anordnung mittels akustischer Maskierung
EP1919257B1 (fr) Réduction du bruit dépendant du niveau
DE4430189A1 (de) Verfahren zur adaptiven Echokompensation
DE102006047965A1 (de) Hörhilfsgerät mit einer Okklusionsreduktionseinrichtung und Verfahren zur Okklusionsreduktion
DE102016011719B3 (de) Aktive Unterdrückung des Okklusionseffektes in Hörhilfen
EP3393143B1 (fr) Procédé de fonctionnement d'une aide auditive
EP3139633A1 (fr) Procede destine a supprimer un effet larsen dans un appareil auditif
EP2981099B1 (fr) Procede et dispositif de suppression de l'effet larsen
EP3926982A2 (fr) Procédé de réduction directionnelle du bruit pour un système auditif comprenant un dispositif auditif
DE102011086728B4 (de) Hörvorrichtung mit einer Einrichtung zum Verringern eines Mikrofonrauschens und Verfahren zum Verringern eines Mikrofonrauschens
EP3951780B1 (fr) Procédé de fonctionnement d'un appareil auditif et appareil auditif
WO2021185891A1 (fr) Système et procédé de compensation de l'effet d'occlusion dans des écouteurs ou des prothèses auditives, avec perception améliorée de sa propre voix
EP3793218B1 (fr) Procédé de fonctionnement d'un appareil auditif ainsi qu'appareil auditif
DE602004006912T2 (de) Verfahren zur Verarbeitung eines akustischen Signals und ein Hörgerät
EP1850634A2 (fr) Procédé de réglage d'une prothèse auditive avec amplification des hautes fréquences
EP1351550B1 (fr) Procédé d'adaptation d'une amplification de signal dans une prothèse auditive et prothèse auditive
DE102019211943B4 (de) Verfahren zur direktionalen Signalverarbeitung für ein Hörgerät
EP4158901A1 (fr) Procédé, dispositif, casque d'écoute et programme informatique pour supprimer activement l'effet d'occlusion pendant la lecture de signaux audio
EP4134954B1 (fr) Procédé et dispositif d'amélioration du signal audio
WO2023214020A1 (fr) Dispositif de réduction du bruit lors de la reproduction d'un signal audio à l'aide d'un casque d'écoute ou d'une prothèse auditive, et procédé correspondant
WO2000041436A1 (fr) Procede pour produire un signal electrique ou procede pour faire ressortir des signaux acoustiques d'une provenance preferentielle, ensemble de transmission et dispositif associes
WO2023104865A1 (fr) Dispositif de suppression de bruit et/ou de suppression d'occlusion actives, procédé correspondant, et programme informatique
DE102020202206A1 (de) Verfahren zur Unterdrückung eines Eigenrauschens einer Mikrofonanordnung
DE102022201943A1 (de) Verfahren zur Unterdrückung eines akustischen Nachhalls in einem Audiosignal

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21713612

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21713612

Country of ref document: EP

Kind code of ref document: A1