Apparatus for reducing noise when playing audio signal by means of earphone or hearing device, and corresponding method

Abstract

According to the invention, at least one sensor (20, 22, 23) for detecting a sensor signal based on ambient sound and / or structural noise is provided in a device for reducing noise when playing an audio signal by means of an earphone (10) or a hearing device. The one or more sensor signals are first fed to a pre-processing unit (32) for pre-processing, the pre-processing unit performing filtering and / or summing for active suppression of noise, active suppression of earplug effects and / or ambient patterns. With a subsequent filter bank, the sensor signal or the output signal of the pre-processing unit is divided into frequency bands by a plurality of filters (30). One or more calculation units (31) are provided for calculating a weighting factor for a single frequency band, said weighting factor being calculated on the basis of a measure for the sensor signal in the respective frequency band and a measure of the noise signal of the sensor at silence in the frequency band or of the output signal of the pre-processing unit (32). A single frequency band is multiplied by an associated calculated weighting factor by a multiplier. The weighted output signals of the filter bank are added by an adder to form a total output signal. A compensation signal based on the total output signal is output by an output unit (21).

Description

[0001] The invention relates to a device for reducing noise when playing audio signals through an audio output device close to the head, in particular headphones or a hearing device. The invention also relates to a corresponding method.

[0002] The functional range of modern headphones usually goes far beyond listening to music or making phone calls. Therefore, active noise cancellation (English: Active Noise Cancellation, ANC) is now a standard feature of such headphones, in which the perceptibility of ambient sound is reduced by playing an acoustic compensation signal through a speaker inside the headphones. The same applies to the transparent mode provided in headphones and hearing devices, in which the ambient sound is filtered and played through a speaker in the headphones, so that there is almost no difference from the hearing sensation of naked ears, that is, not wearing headphones. Some headphones and especially hearing aids also have an atmosphere mode, in which the ambient sound is processed and / or amplified, for example, to improve speech clarity or compensate for hearing loss. However, these applications are not limited to headphones and hearing devices, but can also be implemented in audio output devices near the head of the entire structural design, such as so-called smart glasses, VR / AR headphones, collar speakers or bone conduction headphones. In order to support active noise suppression, transparency or atmosphere mode, in addition to the speaker, the headphones are additionally equipped with at least one external microphone, an internal microphone and an optional acceleration sensor, hereinafter collectively referred to as sensors. The processing of microphone signals and acceleration data detected by external microphones, internal microphones and, if applicable, acceleration sensors, collectively referred to below as sensor data, is usually performed on a digital signal processor (DSP) since only this allows sufficiently high-performance algorithms to be implemented.

[0003] When the acoustic sound is converted into an analog electrical signal by the sensor, noise is added to the useful signal of the corresponding sensor. In addition, the analog-to-digital conversion of the analog electrical signal to the digital signal increases the quantization noise, whose power depends on the bit accuracy of the converter. In addition, due to the processing and filtering of the noisy sensor data on the DSP, the noise is also amplified, so that it can be clearly heard after being played through the speaker. This can be uncomfortable for the user, especially in a quiet environment.

[0004] A number of methods for improving or de-interfering speech and useful signals are described in the prior art. The most common are analysis-synthesis systems, which analyze a portion of the input signal after a time / frequency domain transformation in the frequency domain, perform weighting of the individual frequencies, and then synthesize the time domain signal by inverse transformation. This system has proven to be particularly advantageous due to its computational efficiency, since frequency domain convolution and time / frequency domain transformation can be implemented particularly cost-effectively by fast Fourier transformation. However, this system has a high delay due to the frame-based processing. In order to make high-performance ANC feasible also for non-deterministic signals, the delay in the signal path of the ANC filter must be as small as possible. In the transparent mode, the interference of passive sound and delayed active sound can lead to the appearance of comb filter effects and a double perception of plosives. Therefore, the analysis-synthesis system in the signal path is not suitable for this application.

[0005] An alternative system, the so-called filter bank equalizer, is introduced in the document EP 1 538 749 A2, which uses a time domain filter with a finite impulse response (Finite Impulse Response, FIR) to remove interference from the input signal. The time domain filtering is generated by the product of the impulse response of the prototype low-pass filter and the effective window function. The effective window function is derived from the Fourier transform of the weighting factors for a single frequency band. By applying the time domain filter, the delay of the filter bank equalizer is significantly less than the delay of the analysis and synthesis system. However, the time / frequency domain transformation is still required due to the calculation of the filtering. Therefore, although the filtering itself has no delay, there is still a delay in the adaptation of the filter, so this method is particularly missing in the case of plosives. The linear phase prototype low-pass filter imposes an additional delay according to the filter length. In addition, only FIR filters can be realized by this method, and FIR filters are usually not supported by dedicated processors with low delay. In addition, by this method, a filter bank with nonlinear frequency resolution is only feasible when many all-pass filters are used, which increases the computational complexity.

[0006] The technical problem to be solved by the present invention is to provide a device and a corresponding method, in which the audibility of noise is reduced when an audio signal is played through an audio output device close to the head, in particular headphones or a hearing instrument, without significantly changing the useful signal.

[0007] This object is achieved by a device having the features of claim 1 and a corresponding method according to claim 12. Preferred embodiments of the invention are the subject matter of the dependent claims.

[0008] The device according to the invention for reducing noise when playing an audio signal through an audio output device close to the head, in particular headphones or a hearing device, comprises

[0009] - at least one sensor for detecting a sensor signal based on ambient sound and / or structure-borne noise;

[0010] - a preprocessing unit for preprocessing the one or more sensor signals, wherein the preprocessing unit performs filtering and / or summing for active noise suppression, active occlusion suppression and / or ambience mode;

[0011] - a filter bank for dividing the sensor signal or the output signal of the preprocessing unit into frequency bands by means of a plurality of filters;

[0012] one or more calculation units for calculating weighting factors for individual frequency bands, wherein the weighting factors are calculated based on a measure of the signal of the sensor or the output signal of the preprocessing unit in the corresponding frequency band and a measure of the noise signal of the sensor or the output signal of the preprocessing unit in silence in the frequency band;

[0013] - a multiplier, by which the individual frequency bands are multiplied by the associated calculated weighting factors;

[0014] - an adder, by which the weighted output signals of the filter banks are added together to form a total output signal; and

[0015] - an output unit, configured to output a compensation signal based on the total output signal.

[0016] According to a preferred embodiment of the present invention, the weighting factor is calculated based on the estimated power or standard deviation of the sensor signal or the output signal of the preprocessing unit in the corresponding frequency band and the estimated power or standard deviation of the noise signal of the sensor or the output signal of the preprocessing unit in silence in the frequency band.

[0017] According to a further preferred embodiment, the preprocessing unit, the filters of the filter bank, the multipliers and the adders are implemented at a first sampling rate and the calculation unit is implemented at a second, lower sampling rate.

[0018] Advantageously, the pre-processing unit, the filters of the filter bank, the multipliers and the adders are implemented by the first processor.

[0019] Advantageously, the calculation unit is implemented by a separate second processor.

[0020] Advantageously, output signals of the respective filters of the filter bank are transmitted from the first processor to the second processor via an interface between the processors, and calculated weighting factors for the individual frequency bands of the filter bank are transmitted from the second processor back to the first processor.

[0021] According to one embodiment of the invention, the rate of the output signals of the filters of the filter bank is adapted to the sampling rate on the second processor by means of a sampling rate converter.

[0022] According to another embodiment of the present invention, an input signal of a filter group is transmitted from a first processor to a second processor via an interface between processors, wherein the rate of the input signal to be transmitted is converted by a sampling rate converter and a second filter group is simulated on the second processor to calculate weighting factors, which implement the filter at a correspondingly lower sampling rate, and wherein the calculated weighting factors for individual frequency bands of the filter group are transmitted from the second processor back to the first processor.

[0023] Preferably, in the device according to the present invention, at least one sensor includes one or more internal microphones arranged in the earplug for detecting sound signals in the user's ear canal and / or external microphones for detecting sound signals outside the ear canal and / or an acceleration sensor for detecting structure-borne noise transmitted to the earplug via the ear canal, and the output unit includes at least one speaker.

[0024] Furthermore, in one embodiment it is advantageously provided that the filters of the filter bank can be implemented as low-pass, high-pass and / or band-pass filters by means of cascaded biquad filters.

[0025] The device according to the invention can be integrated in an audio output device that is close to the head, in particular a headset or a hearing aid.

[0026] Accordingly, in the method according to the invention for reducing noise when playing an audio signal via an audio output device close to the head, in particular headphones or a hearing instrument, the following steps are performed:

[0027] - detecting a sensor signal based on ambient sound and / or structure-borne noise by means of at least one sensor arranged in the earphone or the hearing instrument;

[0028] - preprocessing of the one or more sensor signals, wherein the preprocessing unit performs filtering and / or summing for active noise suppression, active occlusion suppression and / or ambience mode;

[0029] - applying noise reduction to the sensor signal or the output signal of the pre-processing unit, wherein the signal

[0030] - split into multiple frequency bands through filter banks,

[0031] - a single frequency band is multiplied by a weighting factor for the single frequency band, wherein the weighting factor is calculated based on a measure for the signal of the sensor or the output signal of the preprocessing unit in the corresponding frequency band and a measure for the noise signal of the sensor or the output signal of the preprocessing unit in silence in the frequency band, and

[0032] - then adding the weighted output signals of the filter bank to form a total output signal; and

[0033] - Output a compensation signal based on the total output signal.

[0034] The invention also relates to a computer program with instructions for causing a computer to execute the steps of the method according to the invention.

[0035] Other features of the present invention are apparent from the following description and claims taken in conjunction with the accompanying drawings.

[0036] Figure 1 Schematically illustrates an in-ear headphone with basic electronic components in the user's ear canal;

[0037] Figure 2 shows a block diagram for noise suppression by filter banks, in the time domain and with adaptation per frequency band;

[0038] Figure 3 A block diagram showing filtering with noise suppression according to the present invention;

[0039] Figure 4 schematically illustrates the amplitude characteristics for a filter bank having a bandpass filter;

[0040] Figure 5 A calculation unit for calculating a weighting factor for each frequency band is shown;

[0041] Figure 6 A calculation unit for numerically robustly calculating the weighting factors per frequency band is shown;

[0042] Figure 7 A calculation unit for calculating a weighting factor is shown, wherein the noise power is estimated from the input signal;

[0043] Figure 8 A block diagram showing a hybrid multi-channel ANC system with noise suppression;

[0044] Fig. 9 A block diagram showing a hybrid multi-channel ANC system with noise suppression and feedback control loop compensation;

[0045] Fig.10 A flow chart of the method according to the invention is shown.

[0046] In order to better understand the principle of the present invention, the embodiments of the present invention are described in detail below according to the accompanying drawings. The present invention is of course not limited to these embodiments and the features can also be combined and modified without leaving the protection scope of the present invention defined in the claims.

[0047] For use of the device according to the present invention, Figure 1In-ear headphones are shown as an example. However, the device can also be used for other types of audio output devices close to the head, such as headphones, hearing devices, smart glasses, VR / AR headsets, collar speakers or bone conduction headphones. The in-ear headphones 10 are held in the ear canal 12 by the earplug 11 and the ear canal is completely or partially acoustically sealed. As a result, the user can no longer clearly perceive his environment. However, in order to enable the user to have a clear auditory impression of his environment when wearing the headphones, the headphones are equipped with a microphone 22 and at least one processor 24, which records, processes and plays the ambient sound through the speaker 21 in the headphones.

[0048] The closure of the ear canal 12 by the earphone 10 also has the effect that structure-borne noise, such as footsteps or speech, which is emitted into the ear canal 12 by vibrating the ear canal wall can hardly escape the ear canal 12 anymore. This manifests itself as a low-frequency amplification of the structure-borne noise relative to an open ear canal, which, in combination with the perceived reduction of the ambient sound, is called the occlusion effect. In order to compensate for the occlusion effect (English: Active Occlusion Cancellation, AOC), additional sensors in the form of a microphone 20 and an acceleration sensor 23 pointing inwards on the side of the earphone pointing toward the ear canal can be used to record information about the structure-borne noise in the ear canal 12.

[0049] In addition to creating acoustic transparency or compensating for occlusion effects when the earphones are worn, the sensors 20, 22, 23 and the processor 24 can also be used to actively attenuate loud ambient sounds. In active noise suppression, an acoustic compensation signal is played by the earphone internal speaker 21 based on the sensor data, which destructively interferes with the ambient sounds at the eardrum 13 and thus reduces the perceived volume of the ambient sounds. In addition, the sensors can be used in an ambient mode, in which the ambient sounds are processed and / or amplified, for example, to improve speech intelligibility or compensate for hearing loss.

[0050] As already mentioned, one problem solved by the method according to the invention is that the sensor not only records the useful signal x(n), n being a discrete time index, but also causes noise v(n). In the context of the invention, the useful signal also includes ambient noise in addition to speech, some of which is considered undesirable in conventional methods for interference sound or noise suppression, but is here assigned to the useful signal instead of noise. The output signal of the sensor is here given by the sum of the useful signal and the sensor noise

[0051]

[0052] Through the processing and filtering of the sensor signal, the noise is significantly amplified, which is perceived as unpleasant when played through the speakers for AOC or ANC.

[0053] Figure 2The structure for noise suppression is shown. Here, the disturbed useful signal First, by having a corresponding impulse response b k The filter groups of K filters 30 of (n) are divided into different frequency ranges, where k=1, 2, ..., K. As a result of the filtering, each of the parallel processing chains is referred to herein as a (frequency) band. The output signals of the filters 30 are respectively

[0054]

[0055] For each frequency band k, a time-varying weighting factor g is calculated in the calculation unit 31 k (n), and then with the frequency band signal An estimate of the useful signal is then obtained at the output of the noise suppression system 63 It is based on the weighted sum of the frequency band signals:

[0056]

[0057] Figure 3 The device according to the invention is shown, in which the useful signal For example, ambient sound or structure-borne noise is recorded by the external microphone 22 of the headset 10, such as Figure 1 , and is additionally disturbed by noise v(n). The microphone signal is then processed by a preprocessing unit 32 on a processor for fast filtering 34 and then by a preprocessing unit with an impulse response b k The individual frequency bands are then filtered by a filter bank of filters 30 of (n). k (n) weighting, wherein the weighting factors are calculated on one or more further processors 35. The output signals of the respective filters of the filter bank are transmitted from the processor for fast filtering to the further processor via an interface. The weighting factors for the individual frequency bands of the filter bank are likewise transmitted from the further processor back to the processor for fast filtering via the interface.

[0058] The preprocessing unit can filter the microphone signal, for example, by a filter with an impulse response w(n), which is designed for a transparent or ambient mode, ANC or AOC. In addition, the preprocessing unit can filter a plurality of sensor signals, for example, by different filters and then add them to generate an output signal.

[0059] The output signal of the filter w(n) is generated as follows:

[0060]

[0061] In order to obtain the lowest possible input-to-output delay on the processor for fast filtering 34, a high sampling rate is preferably used. Advantageously, the input-to-output delay should be less than 1 millisecond. Since the characteristics of the useful signal usually change only slowly, the calculation of the weighting factors can be performed at a lower sampling rate. Advantageously, the calculation of the weighting factors can be performed on one or more separate processors 35. In order to avoid aliasing effects when the band signal is transmitted to a lower sampling rate for calculating the weighting factors, it is optionally provided that the sampling rate is converted by a sampling rate converter 33. In contrast, the weighting factors do not require sampling rate conversion because the weighting factors are low-frequency signals.

[0062] Instead of switching the rates of the K output signals of the filter 30, the rates of the input signals of the filter bank can alternatively be switched, and a second filter bank can be simulated on one or more processors to calculate weighting factors 35, which implement the filter 30 at a correspondingly lower sampling rate. The factors thus calculated are then transmitted to the processor for fast filtering 34 as described above and applied there. Since the filter bank on the processor for fast filtering 34 is still necessary, no complexity reduction is achieved on the processors 34, 35, but the number of sampling rate converters 33 and communication channels between the processors 34, 35 is reduced.

[0063] Figure 4 The amplitude characteristic 40 of a filter bank is shown by way of example, wherein the filter 30 is implemented as a bandpass filter. k Each bandpass filter of (n) is designed here so that, ideally, it is below the lower limit frequency f g，k and above the upper frequency f g，k+1 The input signal frequency is blocked, and the g，k and f g，k+1 The frequencies between 1 and 2 are passed without distortion. This results in the following objective function f of the amplitude characteristic of the bandpass filter according to frequency:

[0064]

[0065] In addition, the sum of the bandpass filters 30

[0066]

[0067] It should be advantageously optimized for the following objective function

[0068]

[0069] In particular, when the bandpass filters are connected in parallel, no undesired cancellation or amplification occurs in the transition region. The lower limit frequency f of the first bandpass filter g，1Here it can be equal to 0Hz, generally known as a low-pass filter, and the upper limit frequency f of the final band-pass filter g，K+1 Here, it can be set equal to the Nyquist frequency, which is generally called a high-pass filter. In order to perform processing that simulates human hearing, it is advantageous if the limit frequency f g，k is linearly distributed on a psychoacoustic frequency scale, such as the Bark scale. Furthermore, the filter 30 can advantageously be optimized so that the transfer function The group run time is minimized taking into account the target curve and the maximum deviation from the amplitude characteristic.

[0070] Advantageously, the filter 30 may be implemented as a cascade of biquad filters.

[0071] Thus, the filter bank contributes only minimally to the delay of the signal path, whereby high performance ANC as well as transparency modes are still possible without comb filtering effects and without the double perception of plosives, and can be computationally efficient. The amplitude and phase characteristics of are taken into account for the design of filters for ANC or AOC, for example, in the preprocessing unit 32 .

[0072] However, an implementation corresponding to an orthogonal filter bank is also possible. On the one hand, this results in a filter bank with linear frequency resolution and possibly additional delays due to sampling rate conversion in the signal path, but on the other hand results in a reduction in computational complexity.

[0073] The weighting factor can be based on the spectrum subtraction rule using the short-term power of the interfered useful signal and the short-term power of the noise In the following, it is shown that regardless of the band index k=1, 2, ..., K, the subtraction rule can be given by the cost function

[0074]

[0075] Derived and derived, assuming that the noise and useful signal are uncorrelated,

[0076]

[0077] According to this rule, the weighting factors can be negative, which leads to undesired phase images. Therefore, the general premise is 0≤g(n)≤1. A hard limit of g(n) in this value range can lead to undesired time artifacts due to discontinuous changes. Dependence on the estimated short-term power of the undisturbed useful signal According to the cost function Mathematically the optimal weighting factor is defined as

[0078]

[0079] This maintains the desired value range g(n). However, x(n) is unknown and therefore Also unknown.

[0080] Since the power of sensor noise is usually low relative to the power of the useful signal, if a useful signal is present, it can be assumed that This gives for the frequency band k the value according to Figure 5 The calculation rules of block 53 in the calculation unit 31. The power of the sensor and quantization noise is usually constant and can therefore be determined for each frequency band, for example, by measurements in silence, hearing tests, consideration of component specifications, estimates during operation in a quiet environment or by mathematical methods. Sensor and quantization noise are often also known as system noise or noise floor. For measurements in silence, for example, the digitized signals of sensors 20, 22, 23 or the output signal of preprocessing unit 32 can be recorded, wherein earphone 10 is activated and located in a room acoustically protected from ambient sound. Advantageously, these values It can be chosen such that a defined useful signal with a selected sound pressure level at one of the sensors 20, 22, 23 or at a separate reference sensor leads to a defined maximum attenuation. This is in contrast to the goal of typical hearing instrument processing, in which the ambient noise is reduced in order to extract the speech signal. Input signal Estimated short-term power For example, the block 50 can be used to smooth the exponential smoother with a smoothing factor of 0<<λ<1 according to

[0081]

[0082] calculate.

[0083] For reasons of cost and battery efficiency, processors that only have fixed-point arithmetic are often used in integrated systems, such as headphones. In particular, the division operation is numerically ill-conditioned, so it is desirable to reduce the dynamics of the input values ​​of the division operation when calculating the weighting factors. This can be achieved by Figure 6 The calculation unit 31 shown is implemented, wherein instead of the corresponding short-term power σ 2 (n), using the standard deviation or square root of the signal variance, σ 2 (n), and then the quotient is squared according to block 53. This calculation rule of block 53 is equivalent to the previous one, ignoring the cross terms of the binomial formula. Figure 5 The calculation rule of block 53 of . The estimation of the standard deviation can be performed by block 51 through an exponential smoother according to

[0084]

[0085] calculate.

[0086] like Figure 7 As shown, the short-term power of the noise signal Based on the output signal of the filter 30 For this purpose, an algorithm such as that established in block 52 can be used, such as a so-called baseline tracker, which is based on the assumption that the power of the noise is constant in time and can be estimated in pauses in the useful signal. When a quiet environment is detected, these methods can be applied in order to perform an estimation of the system noise in the running system. In contrast to hearing device applications, the goal here is not to identify interfering noise from the environment, but to identify system noise.

[0087] In addition, the weight coefficient g can be set as follows k (n), so that the power of the corresponding weighted bandpass signal does not exceed a predetermined threshold. This can be used to protect the hearing of the user in a loud environment.

[0088] Figure 8 A further embodiment of the device according to the invention is shown, which comprises M external microphones 22, a microphone 20 directed towards the ear canal and a loudspeaker 21. In addition or alternatively to the external microphone 22 or the internal microphone 20, one or more acceleration sensors 23 can also be used. The signal of the microphone 20 directed into the ear canal is supplied to a filter 62k(n), which is designed, for example, to attenuate ambient sounds or structure-borne noise d(n) in a certain frequency range. For the design of the feedback filter 62, in particular, a secondary path 61 with an impulse response gg(n) is taken into account, which characterizes the transfer function between the loudspeaker 21 and the internal microphone 20. The signal of the external microphone 22 is supplied to a preprocessing unit 32, which, for example, respectively transmits the signal to the preprocessing unit 32 via a filter w m (n), m=1, 2, ..., M filter the microphone signals and then sum them. The filter can be designed for a transparent or ambient mode, ANC or AOC. By filtering and subsequently summing a plurality of microphones, a filter-and-sum beamformer can be implemented implicitly by the preprocessing unit, for example, which processes the signals of the microphone array in such a way that the output signal mainly only includes sounds from a certain direction. The output signal of the preprocessing unit is then processed according to Figure 3is delivered to the noise suppression system 63. The preprocessing unit 32 and the feedback filter 62 are also implemented at a high sampling rate, for example on a processor for fast filtering, while the calculation unit for calculating the weighting factors for the noise suppression system 63 is implemented at a lower sampling rate, for example on one or more separate processors. The output signal of the feedback filter 62 can be provided via a switch 64 to the sum of the loudspeaker signals (switch position 1) or to the sum of the input signals for the noise suppression system 63 (switch position 2). In addition, the playback of a multimedia signal a(n), such as music or speech from a telephone call, is provided, which is corrected by an equalizer 60 with an impulse response q(n). The output signal of the equalizer is directly transmitted to the loudspeaker 21 together with the output signals of the feedback filter 62 and the noise suppression system 63.

[0089] Fig. 9 Another embodiment of the device according to the invention is shown, which is Figure 8 The embodiment of the present invention has multiple external microphones 22 and filters 32 for filtering, microphones 20 directed toward the ear canal and feedback filters 62, loudspeakers 21, multimedia signals a(n) and equalizers 60. Figure 8 In contrast to the embodiment of The secondary path estimator 65 is connected to the secondary path estimator 65, to which the sum of the corrected multimedia signal and the de-noised signal is fed. The output signal of the secondary path estimator 65 is added to the signal of the internal microphone 20, so that the closed-loop control loop formed by the feedback filter 62 and the secondary path 61 only slightly affects the signal fed to the secondary path estimator 65. In other cases, the transmission characteristics to the internal microphone 20 and even to the eardrum 13 may change in an audible manner according to the feedback filter 62.

[0090] The device according to the invention can be integrated in particular in headphones, wherein such headphones can be designed in different ways. For example, they can be shell headphones, hearables or so-called in-ear monitors, which are used, for example, in live performances of musicians or TV presenters to check their own voice, or they can also be a combination of earphones and a mouth microphone for detecting speech in the form of headphones. However, the device can also be part of a hearing device. In addition, the device can be integrated into smart glasses, VR / AR headphones, collar speakers or bone conduction headphones. Finally, however, parts of the device can also be components of external devices, such as smartphones.

[0091] Fig.10 The basic scheme of the method for reducing noise when playing an audio signal is schematically shown, for example, by Figure 3 Here, the following exemplary reference is made to the application of the method on headphones, but the method is not limited thereto and can also be applied to other audio output devices close to the head, especially hearing devices.

[0092] In the method, ambient sound and / or structure-borne noise, such as speech output of a user wearing the headset or the sound of the user's footsteps, is detected by at least one sensor arranged in the headset in a first step 70. In a subsequent step 71, the corresponding sensor signals are processed by a preprocessing unit, which performs filtering and / or summing for active noise suppression, active occlusion suppression and / or ambience mode.

[0093] Then, in a subsequent step 72, the sensor signal or the output signal of the preprocessing unit is subjected to noise suppression. In this case, the method according to the invention uses a filter bank integrated in the signal path between the sensor and the loudspeaker of the headset. The filtered sensor signal is divided into a plurality of frequency bands by the filter bank.

[0094] The algorithm sets weighting factors which are multiplied by the respective output signals of the filter bank so that frequency bands in which there is hardly any useful signal are significantly attenuated. In particular, the weighting factors can be calculated based on a measure of the signal of the sensor or the output signal of the preprocessing unit (32) in the respective frequency band and a measure of the noise signal of the sensor or the output signal of the preprocessing unit (32) in the frequency band when silent. The weighted output signals of the filter bank are then summed to form the overall output signal.

[0095] Then, in a subsequent step 73, a compensation signal based on the total output signal is supplied to and output by the loudspeaker of the earphone.

[0096] In the case of headphones which include sound transducers for both ears of the user, the method can be carried out separately for both ears or also jointly for the sound transducers of both ears.

[0097] Reference numerals list

[0098] 10 In-ear headphones

[0099] 11. Earplugs

[0100] 12 Ear Canal

[0101] 13 Tympanic membrane

[0102] 20 Internal microphone

[0103] 21 Speakers

[0104] 22 External Microphone

[0105] 23 Acceleration sensor

[0106] 24 Signal Processor

[0107] 30 Filter

[0108] 31 Calculation unit for weighting factors

[0109] 32 Preprocessing Unit

[0110] 33 Sample Rate Converter

[0111] 34 Processors for fast filtering

[0112] 35 Processor for calculating weighting factors

[0113] 40 Filter Amplitude Characteristics

[0114] 50 Power Estimator of Useful Signal

[0115] 51 Estimator of the standard deviation of the useful signal

[0116] 52 Power Estimator for Noise Signals

[0117] 53 Blocks with calculation rules

[0118] 60 Equalizer

[0119] 61 Secondary Path

[0120] 62 Filter

[0121] 63 Noise Suppression System

[0122] 64 Switch for filter output signal

[0123] 65 Secondary Path Estimator

[0124] 70 Method steps for detecting sensor data

[0125] 71 Method steps for filtering sensor data

[0126] 72 Method steps for applying noise suppression

[0127] 73 Method steps for outputting compensation signal

Claims

1. A device for reducing noise when playing an audio signal through an audio output device close to the head, in particular a headset (10) or a hearing device, the device comprising: - at least one sensor (20, 22, 23) for detecting a sensor signal based on ambient sound and / or structure-borne noise; - a preprocessing unit (32) for preprocessing one or more sensor signals, wherein: The pre-processing unit (32) performs filtering and / or summing for active noise suppression, active suppression of occlusion effect and / or ambience mode; - a filter bank for dividing the sensor signal or the output signal of the preprocessing unit (32) into frequency bands by means of a plurality of filters (30); - one or more calculation units (31) for calculating weighting factors for individual frequency bands, wherein the weighting factors are calculated based on a measure of the signal of the sensor or the output signal of the preprocessing unit (32) in the corresponding frequency band and a measure of the noise signal of the sensor or the output signal of the preprocessing unit (32) in silence in the frequency band; - a multiplier, by which the individual frequency bands are multiplied by the associated calculated weighting factors; - an adder, by which the weighted output signals of the filter banks are added together to form a total output signal; and - an output unit (21), configured to output a compensation signal based on the total output signal.

2. The device according to claim 1, wherein: The weighting factor is calculated based on the estimated power or standard deviation of the sensor signal or the output signal of the preprocessing unit (32) in the corresponding frequency band and the estimated power or standard deviation of the noise signal of the sensor or the output signal of the preprocessing unit (32) in silence in the frequency band.

3. The device according to claim 1, wherein: The preprocessing unit (32), the filters of the filter bank, the multipliers and the adders are implemented at a first sampling rate, and the calculation unit (31) is implemented at a lower second sampling rate.

4. The device according to claim 3, wherein: The pre-processing unit (32), the filters (30) of the filter bank, the multipliers and the adders are implemented by a first processor (34).

5. The device according to claim 4, wherein: The calculation unit (31) is implemented by a second processor which is separate therefrom.

6. The device according to claim 5, wherein: Via an interface between the processors, the output signals of the corresponding filters (30) of the filter bank are transmitted from the first processor (34) to the second processor (35) and the calculated weighting factors for the individual frequency bands of the filter bank are transmitted from the second processor (35) back to the first processor (34).

7. The device according to claim 6, wherein: The rate of the output signal of the filter (30) of the filter bank is adapted to the sampling rate on the second processor by means of a sampling rate converter (33).

8. The device according to claim 5, wherein: An input signal of the filter group is transmitted from a first processor (34) to a second processor (35) via an interface between the processors, wherein the rate of the input signal to be transmitted is converted by a sampling rate converter and a second filter group is simulated on the second processor (35) to calculate weighting factors, the second filter group implements filters at a correspondingly lower sampling rate, and wherein the calculated weighting factors for the individual frequency bands of the filter group are transmitted from the second processor (35) back to the first processor (34).

9. The device according to any one of the preceding claims, wherein: The at least one sensor comprises one or more internal microphones (20) arranged in the earplug (11) for detecting sound signals in the ear canal (12) of the user and / or an external microphone (22) for detecting sound signals outside the ear canal and / or an acceleration sensor (23) for detecting structure-borne noise transmitted to the earplug (11) via the ear canal, and the output unit comprises at least one loudspeaker (21).

10. The device according to any one of the preceding claims, wherein The filters (30) of the filter bank are implemented as low-pass, high-pass and / or band-pass filters by cascading biquad filters.

11. The device according to any one of the preceding claims, wherein The device is integrated into an audio output device close to the head, in particular headphones or a hearing aid.

12. A method for reducing noise when playing an audio signal through an audio output device close to the head, in particular headphones (10) or a hearing device, the method comprising the steps of: - detecting (70) a sensor signal based on ambient sound and / or structure-borne noise by means of at least one sensor arranged in the earphone or the hearing instrument; - preprocessing (71) one or more sensor signals, wherein: Filtering and / or summing for active noise suppression, active occlusion suppression, and / or ambience mode; - applying (72) noise reduction to the sensor signal or the output signal of the pre-processing unit, wherein the signal - split into multiple frequency bands through filter banks, - multiplying the individual frequency bands by a weighting factor for the individual frequency bands, wherein the weighting factor is calculated based on a measure for the signal of the sensor or the output signal of the preprocessing unit (32) in the respective frequency band and a measure for the noise signal of the sensor or the output signal of the preprocessing unit (32) in silence in the frequency band, and - then adding the weighted output signals of the filter bank to form a total output signal; and - output (73) a compensation signal based on the total output signal.

Images