Noise-canceling headphones

Abstract

Noise-canceling headphones. Headphones (10) include: - an electro-acoustic transducer (16); - at least one noise-canceling chain including: - an external microphone (31A, 31B) and an internal microphone (32); - a noise-canceling filter (30) for processing the signal from the external microphone (31A, 31B) to produce an external noise-canceling signal (sE), which noise-canceling filter (30) includes: - mounted between the external microphone (31A, 31B) and the output delivering the external noise-canceling signal (sE): at least one finite impulse response (FIR) filter (38); and at least one recursive (IIR) filter (40), - an update unit (252) for each finite impulse response (FIR) filter (38);and - a unit (42) for updating the recursive filters (IIR) based on at least one external anti-noise setpoint signal for the recursive filter (sRABE-f 1, sRABE-f2) from the external microphone (31A, 31B) and a raw error signal (em) from the internal microphone (32). Figure for the abstract: 2;

Description

Title of the invention: Noise-canceling headphones

[0001] The present invention relates to headphones of the type comprising:

[0002] - an electro-acoustic transducer placed in a sound reproduction cavity;

[0003] - at least one noise reduction chain comprising:

[0004] - at least one external microphone for capturing ambient sound outside the cavity;

[0005] - an internal microphone for capturing internal sound in the cavity;

[0006] - a noise reduction filter for each signal originating from each external microphone to produce an external noise-canceling signal, which noise-canceling processing filter comprises:

[0007] - mounted between the external microphone or microphones and the output delivering the signal external noise cancellation:

[0008] - at least one finite impulse response filter; and

[0009] - - at least one recursive filter,

[0010] - an update unit of the or each finite impulse response filter in function of at least one external anti-noise setpoint signal for the finite impulse response filter from the external microphone(s) and a first raw error signal from the internal microphone;

[0011] - amplification means for the excitation of the electro-acoustic transducer at start at least from the external anti-noise signal.

[0012] In noise reduction chains using the signal from the external microphone capturing ambient sound outside the cavity, the quality of the noise reduction depends a lot on the isolation of the sound reproduction cavity, and in particular on the seal ensured between the ear and the mechanical structure of the helmet delimiting the cavity.

[0013] To take into account possible leaks resulting from the fitting of the helmet, it is known that the noise-canceling filter is adaptive, so that the processing chain is modified over time.

[0014] For this purpose, it is known to use finite impulse response filters, known by the acronym FIR, which are easy to adapt. However, these filters are difficult to implement in a low-latency process such as in the case of a noise reduction system, resulting in reduced noise cancellation performance.

[0015] In this context, it is known to use, in addition, adaptive FIR filters, recursive filters or infinite impulse response filters known by the acronym IIR for Infinity Impulse Response in English.

[0016] The implementation of such filters is described, for example, in US patent 2015 / 0243271. In the embodiment of [Fig. 5], in addition to the adaptive FIR filter, An IIR filter is implemented. The IIR filter implemented is chosen from a predetermined table, visible in [Fig.3], according to the leakage level measured in the helmet.

[0017] Even if this arrangement allows the use of a short FIR filter, noise reduction is not optimal, particularly when acoustic leakage exists between the inside of the helmet and the surrounding world.

[0018] The invention aims to provide an audio headset with an anti-noise processing algorithm that compensates for acoustic leakage and is compatible with limited processor resources and higher latency.

[0019] To this end, the invention relates to an audio headset of the aforementioned type, characterized in that it comprises a unit for updating the recursive filters based on at least one external anti-noise setpoint signal for the recursive filter from the external microphone or each external microphone and a second raw error signal from the internal microphone.

[0020] According to particular embodiments, the headphones comprise one or more of the following characteristics: - the first and second raw error signals are identical; - it includes an input for an audio signal to be reproduced connected to the input of the amplification means, in that it includes means for calculating a raw error signal so that the raw error signal comprises a combination of the signal supplied by the external microphone and the audio signal to be reproduced multiplied by an estimation filter of a secondary path of the finite impulse response filter corrected by a feedback loop of the secondary path if it exists; - The recursive filter includes:

[0021] - a filter bank comprising, connected in parallel, at least one filter elementary recursive variable filters, each comprising at least one elementary recursive filter, each associated in series with an elementary variable-gain amplifier; and

[0022] - the recursive filter update unit is specific to controlling the gain of each elementary amplifier as a function of an external anti-noise setpoint signal for the recursive filter from the external microphone(s) and the second raw error signal from the internal microphone; - the recursive filter includes, for the external microphone or each external microphone, a main open-loop noise-reduction filter connected at the input to the associated external microphone and whose output is connected to the input of the variable recursive elementary filter or filters; - Elementary recursive filters are filters whose frequency response decreases for frequencies above 8000 Hz; - it also includes a feedback loop to add to the external anti-noise signal, at the input of the amplification means, an internal anti-noise signal equal to the signal provided by the internal microphone multiplied by a closed-loop anti-noise processing filter; - it includes means for calculating the external anti-noise setpoint signal for the finite impulse response filter connected at the output of the or each external microphone comprising a filter for estimating a secondary path of the recursive filter; - it includes means for calculating the external anti-noise setpoint signal for the recursive filter connected at the output of the or each external microphone comprising a filter for estimating a secondary path of the finite impulse response filter; - at least the finite impulse response filter and update units include, at the input, means of subsampling in order to reduce the computing power required; - at least one of the update units is capable of implementing a recursive least squares algorithm; - at least one of the update units is capable of implementing a least-mean-squares algorithm; and - the update units include means for simulating and optimizing the finite impulse response filter and the recursive filter and means for transferring the characteristics of the simulated finite impulse response filter and the recursive filter to the finite impulse response filter and the recursive filter; - at least one finite impulse response filter and at least one recursive filter are connected in parallel.

[0023] The invention will be better understood upon reading the following description, given solely by way of example and made with reference to the drawings in which:

[0024] - [Fig. 1] [Fig. 1] is a schematic view of a headset according to the invention placed on one ear;

[0025] - [Fig.2] [Fig.2] is a diagram of the headphones illustrating the signal processing in the helmet;

[0026] - [Fig.3] [Fig.3] is a diagram of an IIR filter implemented in the helmet of the [Fig.l];

[0027] - [Fig.4] [Fig.4] is a combination of the characteristic curves of the amplitude and of the phase as a function of the frequency of IIR filters implemented in the headset;

[0028] - [Fig.5] [Fig.5] is a combination of the characteristic curves of the amplitude and the phase as a function of the frequency of the Sftat transfer function of the secondary path and the noise reduction chain; and

[0029] - [Fig.6] [Fig.6] is an amplitude curve of the transfer function of a filter compensation.

[0030] Figure 1 schematically represents an ear 8 covered by one of the two half-shells of a headset 10. In the example considered, this headset has two half-shells, one for each ear, connected by a headband 11. Only one will be described here.

[0031] Each half-shell comprises a dome-shaped casing 12 delimiting a sound-reproducing cavity 14 that surrounds the ear when the headphones are worn. The casing 12 houses the electronic components of the headphones, as is known per se.

[0032] An electroacoustic transducer 16 carried by the casing 12 is disposed in this cavity opposite the ear canal. This transducer is capable of emitting a noise-canceling signal in the cavity 14 and optionally of reproducing a sound signal such as music or voice.

[0033] Alternatively, cavity 14 is delimited by a flexible tip inserted into the ear canal. Each half of the helmet then forms an earpiece. They are then not connected to each other. The helmet is then formed of two in-ear earphones.

[0034] The transducer 16 is connected for its excitation to an amplifier 18 assumed to have unity gain receiving a digital signal to be reproduced through a digital-to-analog converter not shown.

[0035] The headset has an input 22 for a musical signal to be reproduced. This input 22 is formed, for example, by a Bluetooth or Wi-Fi receiver suitable for receiving a digital audio signal.

[0036] The headset includes a noise-canceling filter 30 whose output is connected to the amplifier 18.

[0037] In addition, the headset includes two external microphones 31 A, 31 B suitable for capturing ambient sound outside the cavity 14. Alternatively, it includes a single external microphone or more than two external microphones.

[0038] The microphones 31 A, 31 B are fixed on the outer face of the enclosure 12 and are spaced apart to capture noises coming from different directions.

[0039] The headset further comprises an internal microphone 32 disposed in the cavity 14 for capturing internal sound within the cavity. The internal microphone 32 is disposed along the emission axis of the transducer 16.

[0040] Microphones 31 A, 31B and 32 are connected to the noise-cutting filter 30.

[0041] An accelerometer 34 is provided in the earpiece housing, this accelerometer being it is also connected to the noise reduction filter 30.

[0042] In [Fig.2], the components of the helmet illustrated in [Fig.1] bear the same reference number. Thus, we find the two external microphones 31A, 31B, the internal microphone 32 and the transducer 16 connected to the amplifier 18.

[0043] The noise-cutting filter 30 comprises two external, open-loop noise-cutting chains 38, 40 mounted in parallel between the external microphones 31A, 31B and the amplifier 18.

[0044] One of the chains 38 comprises a set of finite impulse response filters, known by the acronym FIR for Finite Impulse Response, providing a finite external anti-noise signal denoted sABEE

[0045] The other chain 40 comprises a set of infinite impulse response filters known by the acronym IIR for Infinity Impulse Response, referred to hereafter as recursive. They provide an infinite external noise-canceling signal denoted sABE1.

[0046] The outputs of the two chains are connected by a summing 41 at the output of which is obtained an external anti-noise signal noted sE applied to the amplifier 18. The external anti-noise signal sE includes other signals described later.

[0047] The secondary path between the output of the finite impulse response (FIR) filters and the internal microphone 32 is designated by Ss on [Fig.2] while the secondary path between the output of the recursive IIR filters and the internal microphone 32 is designated by Sf on the same figure.

[0048] Similarly, the primary path from the external microphones 31 A, 31B to the internal microphone 32 is designated by the letter P, it corresponds to the passive attenuation of the headset.

[0049] The primary and secondary paths are represented in dotted lines, since they do not correspond to a part of the electronic circuit but are part of the signal processing chain represented here.

[0050] The external noise-cancellation processing chain 40 comprises a recursive filter 40A, 40B at the output of each external microphone 31A, 31B respectively. The outputs of the recursive filters are connected by a summing 40C, at the output of which the infinite external noise-cancellation signal sABEE is obtained.

[0051] Microphones 31 A, 31 B are suitable for providing signals denoted xFFA, xFFB respectively representative of the measured acoustic pressure.

[0052] The IIR 40A and 40B recursive filters have identical structures.

[0053] Each filter 40A, 40B comprises a set of noise reduction filters, some of which are associated with a variable gain amplifier. The filter 30 includes a unit 42 for updating the gain of each elementary variable gain amplifier.

[0054] Each filter 40A, 40B includes first, after the microphone, a main static open-loop filter 44A, 44B whose transfer function is denoted FFA, FFB respectively and whose output signal is denoted xAf, xBf respectively.

[0055] At the output of the main filter 44A, 44B is provided a bank of variable elementary filters IIR 45A, 45B, one of which 45A is shown in more detail in [Fig.3].

[0056] The bank 45A comprises parallel IIR elementary filters 48A1, 48B1, 48C1 with transfer functions HA1, HA2, HA3 respectively, for example, three in number. The filter bank comprises one or more filters.

[0057] Each of the elementary filters is connected in input to the output of the main filter 44A and are connected in series to a variable gain amplifier 50A1, 50B1, 50C1, with gain denoted gAi, gA2, gA3 respectively.

[0058] Amplifiers 50A1, 50B1, 50C1 are connected to the update unit 42 to receive a gain value gAi, gA2, gA3 to be applied.

[0059] The outputs of amplifiers 50A1, 50B1, 50C1 are connected to a summing 52A.

[0060] The structure of bank 45B is identical to that of bank 45A and the elements the constituent parts are designated by the same reference numbers, replacing A with B.

[0061] The anti-noise signals at the output of banks 45A, 45B are denoted yAf, yBf respectively.

[0062] In [Fig. 3], the elementary IIR filters 48A1, 48B1, 48C1 are all filters whose The frequency response decreases at high frequencies, typically above 8000 Hz for headphones with an earcup covering most of the ear and above 2000 Hz for an earpiece. They exhibit a characteristic denoted HA1, HA2, HA3 as illustrated in [Fig. 4] in the case of an earpiece.

[0063] In this figure, the HA1 feature is represented in short dashed lines, the HA2 feature is represented in long dashed lines and the HA3 feature is represented in solid lines.

[0064] Thus in the example considered, the cutoff frequencies are respectively 100 Hz, 500 Hz and 1000 Hz for the elementary filters 48B1, 48C1 and 48A1.

[0065] The three elementary filters are designed to act in different frequency ranges, so that the total gain of the external noise-control chain 40A or 40B grouping the combined filters ensuring noise control can modulate the main static filter 44A, 44B over the entire frequency range, potentially in different ways depending on the frequencies.

[0066] The elementary filters are designed so that a linear combination of the static main filter 44A and the elementary filters 48A1, 48B1, 48C1 associated with their gain models all the transfer functions of the optimal external noise reduction filter which depends on the primary path P and the secondary path Sf, which can vary for the different usage conditions that may occur.

[0067] The filter formed by the external noise-canceling chain 40A depends on the secondary path Sf and the primary path P. Both the secondary path Sf and the primary path P can vary depending on how the headphones are positioned on the ear, the shape of the ear canal, the level of leakage, etc. During the design phase, the optimal filter is calculated for each combination of the primary path P and the secondary path Sf. The elementary filters are designed so that the difference between each optimal filter and the main filter 44A can be described using the elementary filters. An internal noise-canceling feedback loop 55 connects the internal microphone 32, which provides an sFB signal, to the output of the noise-canceling filter 30 via a summing junction 56.

[0068] This loop includes a static internal noise filter 58 whose transfer function is denoted HFB. This filter, as is known per se, is chosen to maximize the attenuation at the eardrum (by maximizing its amplitude) while ensuring that the feedback loop remains stable. At the output of the internal noise filter 58 is obtained an internal noise-canceling signal denoted sAB1

[0069] This filter 58 is a feedback-type noise-canceling filter (feedback loop) that increases the perceived attenuation at the eardrum. Alternatively, the internal noise-canceling correction feedback loop 55 does not exist.

[0070] The input 22 for the audio signal m to be reproduced is connected by a summing 59 to the main branch at the output of the noise-cutting filter 40 so that the signal to be reproduced m, the infinite external noise-cutting signal sABFi, and the internal noise-cutting signal sAB1 and the finite external noise-cutting signal sABFF are added to form the excitation signal denoted sF applied to the transducer 16 after amplification by the amplifier 18.

[0071] The update unit 42 receives, as input, for each filter 40A, 40B, the anti-noise signal xAf, xBf obtained at the output of the main filter 44A, 44B. This signal is subsampled by a subsampler 62A, 62B to go from a sampling frequency for example of 384kHz to 8kHz or any other value between 1 kHz and 348 kHz.

[0072] The signal thus undersampled is addressed in a filter 64A, 64B whose transfer function denoted Sfhat is an estimate of the transfer function Sf of the secondary path between the external anti-noise signal sF which comes out of the summing 41 and the signal measured by the internal microphone 32.

[0073] This transfer function Snw is an estimate of the transfer function of the secondary path Sf with the internal noise correction feedback loop 55.

[0074] It is defined by Sftat= Sf / (l-SfHFB) where Sf is the secondary path and HFB is the transfer function of the internal noise filter.

[0075] An estimate of Sfhat is obtained by fitting a series of biquad circuits to the above expression using transfer functions measured at the acoustic cavities while the internal static noise filter 58 is active.

[0076] Figure [Fig. 5] represents the amplitude and phase of the transfer function.

[0077] On [Fig.5] the transfer function of the secondary path with the closed-loop noise filter Sf / (l-SfHFB) denoted Starget is represented in short dashed lines and the estimate of Sfhat retained for implementation is represented in long dashed lines.

[0078] Each signal obtained at the output of the filter 64A, 64B forms an external anti-noise setpoint signal for the recursive filter s^BE-net sRAHH [2 respectively. Each signal is addressed in three filters 68A1, 68B1, 68C1, respectively 68A2, 68B2, 68C2, identical to the filters 48A1, 48B1, 48C1, respectively 48A2, 48B2, 48C2. The external anti-noise setpoint signals for the recursive filter thus filtered are addressed as input to an adaptive gain calculation unit 70.

[0079] For each filter 40A, 40B, the outputs of the filters 68A1, 68B1, 68C1, respectively 68A2, 68B2, 68C2 are connected to variable gain amplifiers 72A1, 72B1, 72C1, respectively 72A2, 72B2, 72C2 identical to the variable gain amplifiers 50A1, 50B1, 50C1, respectively 50A2, 50B2, 50C2.

[0080] The outputs of the variable gain amplifiers 72A1, 72B1, 72C1, respectively 72A2, 72B2, 72C2 corresponding to the same filter 40A, 40B are summed by a summing 74A, 74B respectively whose outputs are themselves connected to a summing 76.

[0081] The adaptive gain calculation unit 70 also receives as input an error signal ef from the internal microphone 32.

[0082] For this purpose, the microphone 32 is connected to a subsampler 80 whose output is connected to the computing unit 70.

[0083] Advantageously, but optionally, an echo suppression filter 120 with a transfer function equal to S^t receives as input the signal to be reproduced m from input 22 after subsampling in a subsampler 121. Its output is connected to the output of the microphone 32 by a summing 122 ensuring the sum of the internal subsampled signal sFB denoted e, and the signal to be reproduced filtered by the filter 120, which sum is designated by emet forms a raw error signal.

[0084] The raw error signal em is the error signal from the internal microphone 32 without the signal to be reproduced from the input 22 and making an anti-noise signal.

[0085] A control unit 123 adapts the filter 120 according to the raw error signal em obtained at the output of the summing 122 by implementing an algorithm of Least Mean Squares (LMS) is used as an alternative method. Alternatively, a recursive least squares (RLS) algorithm is implemented.

[0086] Unit 70 receives an error signal ef, which is an estimate of the sFB signal from the internal microphone 32, downsampled, from which have been subtracted the signal to be reproduced m from input 22, and the correction currently provided by the IIR filters, 40A, 40B, and to which has been added the correction currently provided by the IIR filters as simulated in unit 42.

[0087] To cancel the current correction of the IIR filters, 40A, 40B, the signals yAf and yBf at the output of the filters 40A, 40B are summed in a summing 130 after being downsampled in subsamplers 132A, 132B. The summed signals filtered in a filter with a gain equal to Sfhat are summed by a summing 136 receiving on its other input the signal from the summing 122.

[0088] Similarly, the output of the summing 76 is connected to the negative input of a subtractor 150 whose positive input is connected to the output of the summing 134 to subtract from the error signal em, after suppression of the correction by the current IIR filters, 40A, 40B, the simulated correction of the filters by the unit 42. The final error signal formed is ef and it is addressed to the unit 70.

[0089] Unit 70 is connected to the variable gain amplifiers 72A1, 72B1, 72C1, 72A2, 72B2, 72C2 to ensure the fixing of their gain by implementation of a Least Mean Squares algorithm known by the acronym LMS.

[0090] The equations for updating the gains gA1, gA2, gA3 are given below:

[0091] g = g + e A1 & Al i Æs,min

[0092] g = g > e ---

[0093] g = ff„+ef„ *

[0094] where is the step size of the least mean squares algorithm;

[0095] xi is the external anti-noise setpoint signal filtered by the filter of characteristic H; ;

[0096] Px is the power of the signal xi;

[0097] Px mj„ is a fixed constant; and

[0098] is the error signal from summing 150.

[0099] Alternatively, and for faster convergence, a recursive least squares algorithm known by the acronym RLS for Recursive Least Square (RLS) algorithm in English is implemented by unit 70.

[0100] The noise reduction processing chain 38 includes a finite impulse response (FIR) filter, denoted 240A, 240B for each external microphone 31A, 31B respectively. The transfer function filters W1, W2 receive as input a signal xsA, xsB respectively obtained by subsampling the signals xFFA, xFFB by subsamplers 242A, 242B.

[0101] The outputs of the two filters 240A, 240B are connected by a summing 244 to produce at output a finite, undersampled external noise reduction signal ys. This signal is addressed to a sampler 246 to form the finite external noise reduction signal sABEF, which is fed back into the summing 4L.

[0102] For the adjustment of the FIR filters 240A, 240B, the circuit includes an update unit 252 suitable for pre-defining the transfer function W1, W2 before ensuring their implementation by the filters 240A, 240B.

[0103] This update unit has an input for each xsA, xsB signal, connected to the output of subsamplers 242A, 242B.

[0104] At the input, these signals are addressed in a filter 254A, 254B whose transfer function denoted Sshat is an estimate of the transfer function of the secondary path Ss which is the transfer function Ss of the secondary path also taking into account the sampler 246 and the subsamplers 80 and 268.

[0105] Each signal obtained at the output of filter 254A, 254B forms an external anti-noise setpoint signal for the finite impulse response filter Srabe siet Srabe S2 respectively. The update unit 252 comprises two variable FIR filters 256A, 256B whose input is connected to the output of filters 254A, 254B respectively and whose outputs are summed by the summing 258.

[0106] A control unit 260A, 260B is associated with each variable FIR filter 256A, 256B to modify their parameters.

[0107] Each unit 260A, 260B receives the input signal from the associated filter 256A, 256B as well as an error signal to adapt the filter by implementing a least mean squares algorithm.

[0108] For the formation of the error signal es, the output of the summing 244 providing the signal ys is connected to a first input of the summing 264 through a filter 266 having Sshat as its transfer function. The output of the summing 122 providing the signal em is connected to the other input of the summing 264 through a subsampler 268.

[0109] At the output of the summing 264 is obtained the error signal e measured by the internal microphone 32, from which have been removed the signal m to be reproduced and the effect of the anti-noise signal generated by the anti-noise processing chain 38 currently in operation estimated by the signal ys filtered by the transfer function Sshat.

[0110] The error signal es is obtained by adding to the previous signal the anti-noise signal estimated in the update unit 252 obtained at the output of the summing 258.

[0111] Thus, the output of the summing 258 is connected to the negative input of a subtractor 270 whose positive input is connected to the output of the summing 264. The signal es is obtained at the output of the summing 70 whose output is connected to the input of each control unit 260A, 260B.

[0112] Furthermore, the external microphones 31 A, 31B, the internal microphone 32 and the accelerometer 34 are each connected to an input of a convergence control unit 340 which is itself connected to the adaptive gain calculation unit 70 and to the control units 260A, 260B for modifying the parameters of the filters 256A, 256B to stop the adaptive calculation of the gains and parameters according to the values ​​of the calculated energies.

[0113] The convergence unit 340 is suitable for calculating the energy of the ambient external sound captured by the microphones 31A and 31B in the frequency range of 50 to 2000 hertz.

[0114] Similarly, over the same frequency range, it is suitable for calculating an internal sound energy corresponding to the energy of the signal captured by the internal microphone 32.

[0115] Finally, it is suitable for calculating the acceleration energy corresponding to the energy of the signal measured by the accelerometer 34 in the frequency band from 70 to 1500 hertz.

[0116] From these calculated energies, the convergence control unit 140 is suitable for controlling the gain calculation unit 70 and the control units 260A, 260B.

[0117] For this purpose, the energy of the ambient external sound is compared to a threshold. If the value of the energy of the ambient external sound is below this threshold for a predetermined duration, for example one second, thus indicating that the ambient external sound is low, then the gains gAb gA2, gA3 and gBb gB2, gB3 are set to predetermined reference values, without taking into account the gain evolution algorithm implemented by the adaptive gain calculation unit 70. Similarly, under these conditions, the parameters of the FIR filters are set to reference values.

[0118] Similarly, if the internal sound energy is much greater than the external ambient sound energy, for example by a ratio greater than 10, the gains gAb gA2, gA3 and gB1, gB2, gB3 are reset to their initial reference values, as are the parameters of the FIR filters. The algorithms for defining the gains gAb gA2, gA3 and gBb gB2, gB3 in the computing unit 70 and for setting the filter parameters in the control unit 260A, 260B are only restarted when the internal sound energy is no longer much greater than the external ambient sound energy.

[0119] Finally, if the acceleration energy exceeds a threshold value, thus reflecting skin vibrations of the user representative of speech spoken by the user or parasitic vibrations, the gains of the various elementary IIR filters and the parameters of the FIR filters are locked at their current values ​​and the algorithms for defining the gains in the computing unit 70 and fixing the Filter parameters in control unit 260A, 260B are stopped until the acceleration energy is less than the predetermined threshold value.

[0120] During operation, and outside of the blocking phases by the convergence control unit 340, the update units 42 and 252 define the gains and parameters for respectively the IIR filter chain 40 and the FIR filter chain 38.

[0121] The gains and parameters are defined to minimize the errors ef and es respectively. When the minimization of the errors is satisfactory, for example by comparing the errors ef and es and by including the error es at the output of the decimator 268, the gains and / or parameters are addressed to the processing chains 38 and 40 and implemented by them.

[0122] It is understood that the simultaneous adaptation of FIR filters and IIR filters in continuous operation makes it possible to obtain good noise attenuation in both low and high frequencies.

[0123] The described embodiment comprises two microphones. Alternatively, it comprises a single microphone.

Claims

Demands

1. Headphones (10) comprising: - an electro-acoustic transducer (16) placed in a sound reproduction cavity (14); - at least one noise-cancelling chain comprising: - at least one external microphone (31A, 31B) for capturing ambient sound outside the cavity (14); - an internal microphone (32) for capturing internal sound within the cavity (14); - a noise-cancelling filter (30) for processing the signal or signals from the external microphone (31A, 31B) to produce an external noise-cancelling signal (sE), which noise-cancelling filter (30) comprises: - connected in parallel between the external microphone (31A, 31B) and the output delivering the external noise-cancelling signal (sE): • at least one finite impulse response (FIR) filter (38);and • at least one recursive filter (IIR) (40), - one unit (252) for updating the or each finite impulse response (FIR) filter (38) as a function of at least one external anti-noise setpoint signal for the finite impulse response filter (srabe si, Srabe si) from the or each external microphone (31 A, 3IB) and a first raw error signal (em) from the internal microphone (32); - amplification means (18) for exciting the electro-acoustic transducer (16) from at least the external anti-noise signal (sE), characterized in that it comprises: - a unit (42) for updating the recursive filters (IIR) as a function of at least one external anti-noise setpoint signal for the recursive filter (srabe fi, sRABE q) from the or each external microphone (31 A, 3IB) and a second raw error signal (em) from the internal microphone (32).;

2. Headphones according to claim 1, characterized in that the first and second raw error signals (em) are identical.

3. Headphones according to the preceding claim, characterized in that it comprises an input (22) for an audio signal (m) to be reproduced connected to the input of the amplification means (18), in that it comprises means (122) for calculating a raw error signal (em) such that the raw error signal (em) comprises a combination of the signal provided by the internal microphone (32) and the audio signal to be reproduced (m) multiplied by a filter (Sfhat) (120) for estimating a secondary path of the finite impulse response (FIR) filter (38) corrected for a feedback loop (55) of the secondary path if it exists.

4. Headphones according to any one of the preceding claims, characterized in that the recursive filter (IIR) (40) comprises: - a bank (45A, 45B) of filters comprising, mounted in parallel, at least one elementary variable recursive filter (48A1, 48B1, 48C1) each comprising at least one elementary recursive filter (48A1, 48B1, 48C1) each associated in series with an elementary variable gain amplifier (50A1, 50B1, 50C1); and - the recursive filter update unit (42) is suitable for controlling the gain of each elementary amplifier (50A1, 50B1, 50C1) as a function of an external anti-noise setpoint signal for the recursive filter (srabe fi, Srabe q) from the one or each external microphone (31 A, 31B) and the second raw error signal (em) from the internal microphone (32).

5. Headphones according to claim 4, characterized in that the recursive filter (40) comprises, for the external microphone or each external microphone (31 A, 31B), a main open-loop noise-cancelling filter (44A, 44B) connected at the input to the associated external microphone (31 A, 31B) and whose output is connected, at the input of the variable recursive filter or each elementary filter (45A, 45B).

6. Headphone according to claim 4 or 5, characterized in that the elementary recursive filters (48A, 48B, 48C) are filters whose frequency response decreases for frequencies above 8000 Hz.

7. Headphones according to any one of the preceding claims, characterized in that they further comprise a feedback loop (55) for adding to the external anti-noise signal (sABEf), at the input of the amplification means (18), an internal anti-noise signal (sABi) equal to the signal (e) provided by the internal microphone (32) multiplied by a closed-loop noise-processing filter (58).

8. Headphone according to any one of the preceding claims, characterized in that it comprises means for calculating the external anti-noise setpoint signal for the finite impulse response filter (srabe ^ i , sRABE_s2) connected at the output of the or each external microphone (31 A, 31B) comprising a filter (Sshat) (254A, 254B) for estimating a secondary path of the recursive filter (IIR) (40).

9. Headphones according to any one of the preceding claims, characterized in that they comprise means for calculating the external anti-noise setpoint signal for the recursive filter (s^Ef i, sRABE f2) connected at the output of the or each external microphone (31 A, 31B) comprising a filter (Sfhat) (254A, 254B) for estimating a secondary path of the finite impulse response (FIR) filter (38).

10. Headphones according to any one of the preceding claims, characterized in that at least the finite impulse response (FIR) filter 38 and update units (42, 252) comprise, at the input, downsampling means (62A, 62B, 80, 132A, 132B, 242A, 242B, 268) in order to reduce the computing power required.

11. Headphones according to any one of the preceding claims, characterized in that at least one of the update units (42, 252) is suitable for implementing a recursive least squares (RLS) algorithm.

12. Headphones according to any one of the preceding claims, characterized in that at least one of the update units (42, 252) is suitable for implementing a least-mean squares (LMS) algorithm.

13. Headphones according to any one of the preceding claims, characterized in that the update units (42, 252) comprise means (72A1, 72B1, 72C1, 72A2, 72B2, 72C2, 256A, 256B) for simulating and optimizing the finite impulse response (FIR) filter (38) and the recursive (IIR) filter (40) and means for transferring the characteristics of the simulated finite impulse response (FIR) filter (38) and the recursive (IIR) filter to the finite impulse response (FIR) filter (38) and the recursive (IIR) filter (40).

Images