Audio rendering enhancement by frequency band extraction

Abstract

Audio device (14) comprising: - a frequency component extraction module (24); - a selection module (30) arranged to select at least one selected audio channel (Cs) and to control the extraction module (24) so ​​that it extracts said frequency components only from at least one selected audio channel; - a processing module (25) arranged to perform at least one processing operation inducing an inter-channel delay on the at least one selected audio channel (Csp) deprived of said frequency components and on the at least one unselected audio channel (Cns); - an injection module (26) arranged to sum said frequency components on each processed audio channel.

Description

[0001] The invention relates to the field of multichannel audio systems. BACKGROUND OF THE INVENTION

[0002] A set-top box, or STB (for Set-Top Box ) , is an audio-video broadcasting device, whose primary function is to acquire an audio-video stream, decode this stream, and broadcast the video stream through a television and the audio stream through the television speakers and / or possibly through other audio playback equipment (soundbar, connected speakers, etc.).

[0003] Some recent set-top boxes include an audio device with one or more speakers. These speakers can be used for voice assistant functions, but also to contribute to a multi-channel audio system that reproduces the audio stream from the video stream. Multi-channel playback provides optimal and immersive sound.

[0004] We know of such a decoder box whose audio device allows the Left (or L) channels to be "physically" reproduced. Left ), Right (or R for Right ) and Center (or C for Center ) .

[0005] The audio device can also perform a virtualization process to give the illusion that audio channels not physically present are (virtually) present.

[0006] This allows, for example, the implementation of 5.1.2 audio playback, resulting in a more immersive sound experience.

[0007] The 5.1.2 format normally requires five main speakers (front left, front center, front right, rear left, rear right), a subwoofer ( subwoofer ) and two height speakers ( up-firing ) which can be placed on the ceiling or directed towards the ceiling to reflect sound. Height speakers, for example, are positioned in what is called a " top middle "

[0008] In the situation described here, the rear channels, also called channels surround ( Left surround (L s ) and Right surround (R s )), and elevated canals, also called top middle ( Left top middle (L tm ) and Right top middle (R tm )), are not physically present but are created and then virtualized. The channels surround And top middle are then sub-mixed (we are talking about downmixing ) on the physically present channels (L, R, C), which gives the user the illusion that these channels are indeed present and therefore that part of the audio content is coming from behind or above them.

[0009] Alternatively, the set-top box can be connected to other audio playback devices and work with them to achieve multichannel playback. These other audio playback devices could be, for example, two satellite speakers that physically reproduce the two channels. surround. In this case, only the channels top middle are virtualized.

[0010] With reference to the figure 1 The processing chain 1 implemented by the audio device 2 of the decoder box 3 includes a module upmixer 4, a processing module 5 performing the virtualization, a module downmixer 6 and a low-frequency creation module 7.

[0011] We can translate " upmix » by “overmixing” and downmix by "sub-mixing".

[0012] Upmixing is an audio processing technique that converts an input audio signal with a limited number of channels into a signal with a larger number of channels. The input audio signal typically contains between two and eight audio channels. For example, a stereo signal with two left and right channels. The resulting signal is, for example, in a 5.1.2 configuration.

[0013] Upmixing uses algorithms to redistribute the sound elements of the original signal to different speakers. For example, dialogue can be directed to the center speaker, sound effects to the rear speakers, and ambient sounds to the height speakers.

[0014] Submixing, on the other hand, consists of reducing the number of channels in a multichannel signal.

[0015] The module upmixer 4 therefore acquires input audio channels C e contained in the audio stream and creates the L / R / C / L s / R s / L tm / R tm channels from the input audio channels C e.

[0016] The processing module 5 performs the virtualization.

[0017] The module downmixer 6 uses a matrix to implement the matrixing of virtualized channels C v onto physically present channels (for example L, R and C for decoder box 3). The matrixing allows, by means of coefficients of a matrix M, the multiplexing of the input channels of the matrix onto the output channels.

[0018] The low-frequency creation module 7 extracts the low-frequency components (down to the so-called "frequency of crossover ") of previously generated channels.

[0019] Then, the low-frequency components are summed to produce the LFE channel (for Low Frequency Effect ) .

[0020] For example, the frequency of crossover is fixed at 500 Hz, so that all frequencies below 500 Hz are extracted as low frequencies by the so-called filtering crossover.

[0021] The C so output channels are intended to be played back by the multichannel audio system.

[0022] The virtualization processes implemented by processing module 5 have the drawback of inducing an inter-channel delay, which generates a phase shift between the channels. The low frequencies thus phase-shifted by virtualization are summed in the LFE channel by the low-frequency generation module 7, resulting in a power loss of up to 10 dB SPL (for Sound Pressure Level (This is a measurement of the sound level received at a point).

[0023] Such a loss of power must be avoided.

[0024] Generally, the ability to adjust the frequency at which virtualization processes are applied avoids this problem by acting directly in the processing module 5. However, this setting may not be available, or may not allow coverage of all the low frequencies extracted by the low-frequency management module 7, particularly in the case of a frequency of crossover high imposed by the acoustic elements.

[0025] For example, virtualization could be applied starting from a maximum of 250 Hz (e.g., the virtualization window starts at 250 Hz, which is a maximum value for this threshold; in practice, the virtualization window could start at a frequency lower than 250 Hz), and the acoustic elements of the audio device 2 of the decoder box 3 could impose a frequency of crossover of 500 Hz (e.g., speakers designed for reproducing low frequencies are characterized by a maximum acoustic frequency of 500 Hz). In this case, the adjustment does not resolve the problem just mentioned.

[0026] It should be noted that this problem does not necessarily concern virtualization processes, but more generally any process that introduces inter-channel delay. Similarly, the frequency band to which the extracted frequency components belong is not necessarily a low-frequency band.

[0027] We are therefore seeking to resolve this power loss issue within a general context. Of course, the solution chosen to address the problem must not degrade the acoustic performance for the user. SUBJECT OF THE INVENTION

[0028] The invention aims to improve the audio performance of a multi-channel audio system. SUMMARY OF THE INVENTION

[0029] To achieve this goal, an audio device is proposed comprising a processing chain arranged to acquire input audio channels and to produce output audio channels intended for playback by a multichannel audio system, the processing chain comprising: a frequency component extraction module belonging to a predefined frequency band, to the input of which are applied first audio channels from the input audio channels; a selection module arranged to select at least one selected audio channel from among the first audio channels, and to control the extraction module so that it extracts said frequency components only from at least one selected audio channel, and not from at least one unselected audio channel; a processing module arranged to perform at least one processing operation inducing an inter-channel delay on the at least one selected audio channel deprived of said frequency components and on the at least one unselected audio channel, so as to produce processed audio channels;an injection module arranged to sum said frequency components on each processed audio channel in order to produce second audio channels intended to be used to produce the output audio channels.

[0030] The audio device extracts the frequency components from the predefined frequency band (for example, low-frequency components) upstream of the processing module, which introduces the inter-channel delay. The injection module then sums these frequency components across the processed audio channels, downstream of the processing module. The processing module therefore does not generate any phase shift on these frequency components, which can then be summed without any loss of power.

[0031] However, it was observed that injecting the frequency components of a predefined frequency band from certain channels onto all processed audio channels could degrade the audio quality. In particular, when an audio channel contains a voice, reinjecting the frequency components from that channel onto all other channels impairs voice intelligibility and generates an echo effect. The selection module therefore allows the extraction of frequency components from selected channels only, thus mitigating this problem.

[0032] The audio device therefore solves the problem of power loss without degrading audio quality. This significantly improves the audio performance of the multichannel audio system.

[0033] We also propose an audio device as previously described, in which at least one audio channel selected by the selection module includes at least one predefined audio channel, statically selected by the selection module.

[0034] We further propose an audio device as previously described, in which the at least one selected audio channel comprises a Left channel and a Right channel, and the at least one unselected audio channel comprises a Center channel.

[0035] We also propose an audio device as previously described, the processing chain being arranged to analyze audio channels present upstream of the extraction module, with at least one audio channel selected by the selection module including at least one channel selected according to this analysis.

[0036] We also propose an audio device as previously described, in which the predefined frequency band includes low frequencies below a predefined threshold.

[0037] We also propose an audio device as previously described, in which the processing module implements virtualization.

[0038] We also propose an audio device as previously described, the processing chain further including a module upmixer arranged to produce the first audio channels from the input audio channels.

[0039] We also propose an audio device as previously described, the processing chain further including a module downmixer arranged to produce third audio channels from second audio channels, the output audio channels comprising the third audio channels.

[0040] We also offer equipment including an audio device as previously described.

[0041] We also offer equipment as previously described, the equipment being a decoder box.

[0042] We also propose a processing method, implemented in the audio device as previously described, and comprising the following steps: select at least one selected audio channel from the first audio channels, and command the extraction module so that it extracts the frequency components only from at least one selected audio channel, and not from at least one unselected audio channel; perform at least one processing inducing an inter-channel delay on the at least one selected audio channel deprived of said frequency components and on the at least one unselected audio channel, so as to produce processed audio channels; sum the said frequency components on each processed audio channel so as to produce second audio channels intended to be used to produce the output audio channels.

[0043] We also propose a computer program comprising instructions which lead the audio device of the equipment as previously described to execute the steps of the processing method as previously described.

[0044] In addition, a computer-readable recording medium is proposed, on which the computer program as previously described is recorded.

[0045] The invention will be better understood in light of the following description of particular, non-limiting embodiments of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Reference will be made to the attached drawings, among which: [ Fig. 1 ] there figure 1 represents a processing chain for a prior art audio device; [ Fig. 2 ] there figure 2 represents the set-top box, the television, and satellite speakers; [ Fig. 3 ] there figure 3 represents a processing chain for an audio device according to a first embodiment of the invention; [ Fig. 4 ] there figure 4 is a graph representing, in the frequency domain, the curves of the LFE channel signals with and without low-frequency extraction; Fig. 5 ] there figure 5 is a graph representing, in the frequency domain, the curves of the left channel signals at the input of the processing chain, and at the output with and without channel selection; Fig. 6 ] there figure 6 is a figure similar to the figure 5 , with the Centre channel; Fig. 7 ] there figure 7 is a figure similar to the figure 5 , with the LFE channel; [ Fig. 8 ] there figure 8 represents a processing chain for an audio device according to a second embodiment of the invention.

[0047] The curves of figures 4 à 7 represent on the ordinate the level of the audio channel considered, expressed in dBFS ( Decibel Full Scale ) as a function of the acoustic frequency, expressed in Hertz (Hz) on the x-axis. DETAILED DESCRIPTION OF THE INVENTION

[0048] With reference to the figure 2 , the set-top box 10 (also called TV set-top box, or audiovisual set-top box, or Set-Top Box ) is connected to a TV 11 and satellite speakers 12.

[0049] The decoder box 10 is an "advanced" decoder box, which incorporates an audio device 14 including several speakers 15.

[0050] The audio device 14 of the decoder box 10 is integrated into a multi-channel audio system 16 which, in addition to the decoder box 10, includes the satellite speakers 12.

[0051] The decoder box 10 receives an audio-video stream and transmits the video stream to the TV 11 and the audio stream to the audio devices of the audio system 16. Here, the TV 11 is therefore not used to reproduce the audio stream (but the invention could be implemented by also using the speakers of the TV 11).

[0052] The audio-video stream can originate from any source, such as a broadcast network (satellite television network, internet connection, digital terrestrial television (DTT) network, cable television network, etc.), or other equipment connected to the set-top box (a CD, DVD, or other player). BlueRay, a smartphone, a tablet, etc.), or even a storage device (for example a USB key or a memory card connected to the decoder box).

[0053] The decoder box 10 also plays the role of audio playback equipment in the audio system 16: its speakers 15 reproduce part of the audio stream.

[0054] The speakers 15 of the audio device 14 of the set-top box 10 include a left speaker, a right speaker and a center speaker, which allow the set-top box to "physically" reproduce the channels Left (L), Right (R) and Center (C).

[0055] The audio device 14 of the decoder box 10 also includes audio components 17 (including audio amplifiers) which shape the audio signals reproduced by the speakers 15 of the decoder box 10.

[0056] The audio device 14 of the decoder box 10 also includes a processing unit 18.

[0057] The processing unit 18 (electronic and software) comprises at least one processing component 19, which is, for example, a "general-purpose" processor, a processor specialized in signal processing (or DSP, for Digital Signal Processor ), a specialized processor for artificial intelligence algorithms (of the NPU type, for Neural Processing Unit ), a microcontroller, or a programmable logic circuit such as an FPGA (for Field Programmable Gate Arrays ) or an ASIC (for Application Specific Integrated Circuit ) .

[0058] The processing unit 18 also includes one or more memories 20, connected to or integrated into the processing component 19. At least one of these memories 20 forms a computer-readable recording medium, on which is recorded at least one computer program comprising instructions which lead the processing component 19 to execute at least some of the steps of the processing procedure which will be described.

[0059] The processing unit 18 manages the multichannel audio playback of the audio stream by the audio system 16 and, in particular, implements a virtualization process to give the illusion that audio channels not physically present are (virtually) present. The system is configured in 5.1.2, and the channels Left top middle And Right top middle are virtualized. The channels Left surround (L s ), Right surround (R s ) are reproduced here by the 12 speakers, and the channels Left (L), Right (R) and Center (C) are reproduced by the speakers 15 of the decoder box 10. Here, the LFE channel is reinjected onto the different channels, but it would also be possible to have a subwoofer to reproduce it.

[0060] With reference to the figure 3 The processing unit 18 implements a processing chain 22 comprising a module upmixer 23, an extraction module; 24, a processing module; 25, an injection module; 26, a module downmixer 27 and a bass generation module 28 which produces and sums the low-frequency components to produce the LFE channel. These modules are arranged in this order from upstream to downstream. By "upstream," we mean the side of the input E of the processing chain 22, and by "downstream," we mean the side of the output S of the processing chain 22.

[0061] The processing chain 22 also includes a selection module 30 which cooperates with the extraction module 24.

[0062] The audio stream contains input audio channels Ce, typically between two and eight input channels. Here, in this case, the audio stream is in stereo format and includes two input channels Ce, which are the R and L channels.

[0063] The module upmixer 23 acquires the input audio channels C e and creates first audio channels C 1 from the input audio channels. The first audio channels C 1 are the L / R / C / L s / R s / L tm / R tm channels.

[0064] The extraction module 24 performs an extraction of frequency components belonging to a predefined frequency band. The first audio channels C1, derived from the input audio channels Ce, are applied to the input of the extraction module 24.

[0065] The predefined frequency band includes low frequencies below a predefined threshold. The frequency components extracted by the extraction module 24 are therefore the low-frequency (LF) components below the predefined threshold, for example, those between 1 Hz and the predefined threshold. The predefined threshold is equal to the frequency of crossover, which is, for example, equal to 500 Hz.

[0066] The selection module 30 selects at least one selected audio channel C s from among the first audio channels C 1, commands the extraction module 24 so that it extracts said frequency components only from at least one selected audio channel C s, and not from at least one unselected audio channel C ns.

[0067] Therefore, the extraction is not performed on all audio channels. The purpose of this operation will be explained later.

[0068] The processing module 25 performs at least one processing that induces an inter-channel delay on at least one selected audio channel C sp deprived of low-frequency components and on at least one unselected audio channel C ns (therefore not deprived of said low-frequency components), so as to produce processed audio channels Ct. The at least one processing here includes a virtualization.

[0069] For example, virtualization applies a "head-to-signal" transfer function to selected audio channels C sp deprived of low-frequency components (L and R channels deprived of low-frequency components) and to unselected audio channels C ns (C / L s / R s / L tm / R tm channels).

[0070] The injection module 26 sums the low frequency components BF on each processed audio channel Ct so as to produce second audio channels C 2 intended to be used to produce output channels C so themselves intended to be reproduced by the multichannel audio system.

[0071] The injection module 26 therefore reinjects, onto the treated channels Ct, the low frequency components BF previously extracted.

[0072] The module downmixer Module 27 acquires the second audio channels C2 and performs matrixing on the second audio channels C2. downmixer 27 produces third audio channels C 3 from second audio channels C 2. The output audio channels C so include the third audio channels C 3.

[0073] Matrixing is performed according to a predetermined distribution matrix M, also stored in memory 20.

[0074] The second audio channels C 2 are therefore allocated to the physical channels L, R, C of the decoder box 10 and to the physical channels L s , R s of the satellite speakers 12 according to the predetermined distribution matrix M. The matrix is ​​for example similar to the matrix in the Annex, where each value is coded on a 16-bit integer, i.e. 16385 possible values.

[0075] Alternatively, if the speakers 12 are not present, it is possible to assign the second audio signals C 2 only to the physical channels L, R, C of the decoder box 10.

[0076] The low-frequency creation module 28 then produces the LFE channel by extracting low frequencies (down to the frequency de crossover ) of the third audio channels C 3 generated previously, and summation of these low frequencies from these channels (e.g. summation of all low frequencies extracted up to 500 Hz).

[0077] The C output channels intended to be reproduced by the audio system 16, therefore include the third audio channels C 3 and the LFE channel.

[0078] The advantages of the invention are now explained.

[0079] First, we are interested in a degraded implementation of audio chain 22 of the figure 2 , which would not include the selection module 30.

[0080] The extraction module 24 therefore extracts in this case all the low-frequency components (below 500 Hz) from all the channels L, R, C, Ls, Rs, Ltm, Rtm. This extraction is performed indiscriminately for all channels, including channel C. This results in a set of filtered channels designated by L', R', C', L's, R's, L'tm, R'tm.

[0081] The processing module 25 implements virtualization on all these channels.

[0082] All low-frequency (LF) components are therefore preserved from phase shifts related to virtualization, and the power level is not affected. The audio device 14 solves the power reduction problem encountered in the prior art corresponding to the figure 1 , which is due to the inter-channel delay which generates a phase shift between the channels.

[0083] We can see on the figure 4 that the amplitude in the frequency domain of the C1 curve corresponding to the LFE channel at the output of the processing chain 22 is greater than the amplitude of the C2 curve corresponding to the LFE channel at the output of the processing chain 1. The improvement is very clear for this channel.

[0084] However, a problem arises in the case of a source with more than two channels.

[0085] We know that channel C is primarily dedicated to human voices. The low-frequency (LF) components extracted by extraction module 24 from channel C therefore include low-frequency components of human voices.

[0086] The reinjection of these low-frequency components, performed by the injection module 26 across all channels, therefore reinjects the low-frequency components of human voices onto all channels. This reinjection of the low-frequency components of human voices onto all channels degrades the audio quality, impairs voice intelligibility, and generates an unpleasant echo effect for the user. This echo effect is due to the user's perception that the human voices are coming from multiple audio sources because the low-frequency components of human voices are present in all physically present channels, thus giving the impression of multiple audio sources for the "low-frequency" reproduction of human voices. Indeed, the low-frequency portion of the voice is thus diffused across all channels and not just on channel C.

[0087] In the case of signals rendered by heterogeneous speakers, the low-frequency part of the voice would then be treated with a different tone, or even a different sound level.

[0088] Similarly, a similar problem occurs with the Ls / Rs channels, where the low frequencies are broadcast on the L, R, and C channels, thus breaking the multichannel effect. For example, a "rumble" normally reproduced on the Ls channel is heard by the user on all channels.

[0089] Thus, even if the audio rendering is significantly improved compared to previous art, the user experience is degraded and the reproduced signal is distorted compared to the original intention of the content creator.

[0090] Using the selection module 30 solves the problem just described.

[0091] Indeed, thanks to the selection module 30, the extraction of low frequencies is selective in the sense that it is carried out only on a subset of channels (e.g. L and R) selected by the selection module 30.

[0092] Here, for example, the selection module 30 excludes the C channel dedicated to human voices and / or the L s / R s channels, so that the low frequency components of the "excluded" channels (e.g. C; L s / R s) do not end up in the other channels (e.g. L and R) when the bass is reinjected.

[0093] Here, the at least one audio channel selected C s by the selection module 30 includes at least one predefined audio channel, statically selected by the selection module 30.

[0094] For example, the selection module 30 selects only the L, R audio channels on which the extraction module 24 must perform the extraction of the low frequency components.

[0095] The selected audio channel(s) C are pre-defined in a static configuration file stored in memory 20 of the processing unit 18. These audio channels can also be defined directly by the user (i.e., user selection). For example, a mask in .json format is used for selecting audio channels: {bass_extraction_channel_mask :[true, true, false, false, false, false, false]}

[0096] In this example, it is stated that: Audio channels L and R are considered for low-frequency component extraction (selected channels): "true"; and audio channels C, LFE, L s, R s, L tm, R tm are not considered for bass extraction (unselected channels): "false".

[0097] This selective extraction has the effect of improving the audio rendering, in particular by strengthening the intelligibility of human voices in the broadcast audio stream and limiting echoes, while preserving the multichannel effect.

[0098] On the figures 5 à 7 , the input signal is a 5.1 stream having a 100 Hz component on the C channel and a 200 Hz component on the L channel.

[0099] We can see on the figure 5 curve C3 of the Left channel at the input of chain 22 (on which the 200 Hz component can be distinguished), curve C4 of the Left channel at the output of chain 22 without the selection module 30 and curve C5 of the Left channel at the output of chain 22 with the selection module 30.

[0100] We can see on the figure 6 the C6 curve of the Center channel at the input of chain 22 (on which the 100 Hz component can be distinguished), the C7 curve of the Center channel at the output of chain 22 without the selection module 30 and the C8 curve of the Center channel at the output of chain 22 with the selection module 30.

[0101] We can see on the figure 7 the C9 curve of the LFE channel at the input of chain 22, the C10 curve of the LFE channel at the output of chain 22 without the selection module 30 and the C11 curve of the LFE channel at the output of the chain with the selection module 30.

[0102] The C channel was not selected by the selection module 30, and the 100 Hz component is not present on the C11 curve while the 200 Hz component, from the L channel, is present.

[0103] Thus, the measurement results clearly show that the invention has made it possible to completely exclude the low-frequency component of the C channel from the LFE channel.

[0104] Of course, the invention is not limited to the embodiments described but encompasses any variant falling within the scope of the invention as defined by the claims.

[0105] The processing that induces an inter-channel delay, performed by the processing module, is not necessarily virtualization. It could be filtering, or the creation of an effect (reverb for example).

[0106] The predefined frequency band, to which the frequency components extracted by the extraction module belong, is not necessarily a low-frequency band. It could be, for example, a band between 3 kHz and 4 kHz.

[0107] The processing chain does not necessarily include a module upmixer And downmixer.

[0108] We have described here that the audio channel(s) selected by the selection module are predefined channels, statically selected by the module. However, the processing chain could analyze audio channels upstream of the extraction module, with at least one audio channel selected by the selection module then being chosen based on this analysis. Upstream audio channels are, for example, the input audio channels or the first audio channels. The analysis, performed for example by the selection module, could consist of detecting a predetermined sound, for example, defined by a certain level in a certain frequency band. The selection module can also select the audio channel(s) whose frequency components are extracted, depending on the module's current configurations. upmixer and / or the module downmixer. For example, if the downmix The current version is 3.1, the L and R channels could be selected and therefore taken into account for bass extraction.

[0109] With reference to the figure 8 , in a second particular embodiment called "hybrid selection", the selection module 30 is configured to jointly select a first subset of channels statically as described above (and therefore for example by accessing a configuration file stored in memory 20) and a second subset of channels according to the analysis carried out on channels (for example the first audio channels C 1 ) present upstream of the extraction module as described above.

[0110] By "subset of channels", we mean one or more channels.

[0111] This hybrid channel selection method has the advantage of adapting in real time to a wide variety of audio formats, while ensuring excellent immersive sound quality. In particular, it allows for a good compromise between processing speed (by limiting the number of channels to analyze) and adaptability to different audio configurations. Annexe distribution matrix:

[0112] Output L R C LFE Ls Rs Input L 16384 0 0 0 0 0 R 0 16384 0 0 0 0 C 0 0 16384 0 0 0 LFE 0 0 0 0 0 0 Ls 0 0 0 0 16384 0 Rs 0 0 0 0 0 16384 Tls 16384 0 0 0 0 0 Trs 0 16384 0 0 0 0

Claims

1. Audio device (14) comprising a processing chain (22) arranged to acquire input audio channels (C e ) and to produce audio output channels (C so ) intended to be reproduced by a multichannel audio system (16), the processing chain comprising: - a frequency component extraction module (24) belonging to a predefined frequency band, to the input of which are applied first audio channels (C1) from the input audio channels; - a selection module (30) arranged to select at least one selected audio channel (C s ) among the first audio channels (C1), and to control the extraction module (24) so ​​that it extracts said frequency components only from at least one selected audio channel, and not from at least one unselected audio channel (C ns) ; - a processing module (25) arranged to perform at least one processing operation inducing an inter-channel delay on at least one selected audio channel (C sp ) deprived of said frequency components and on at least one unselected audio channel (C ns ), so as to produce processed audio channels (Ct); - an injection module (26) arranged to sum said frequency components on each processed audio channel so as to produce second audio channels (C2) intended to be used to produce the output audio channels (C so ).

2. Audio device according to claim 1, wherein at least one selected audio channel (C s ) by the selection module (30) includes at least one predefined audio channel, statically selected by the selection module.

3. Audio device according to claim 2, wherein at least one selected audio channel (C s) includes a Left channel and a Right channel, and at least one unselected audio channel (C ns ) includes a Centre channel.

4. Audio device according to any one of the preceding claims, the processing chain (22) being arranged to analyze audio channels present upstream of the extraction module (24), the at least one selected audio channel (C s ) by the selection module (30) comprising at least one channel selected according to this analysis.

5. Audio device according to any one of the preceding claims, wherein the predefined frequency band includes low frequencies below a predefined threshold.

6. Audio device according to any one of the preceding claims, wherein the processing module (25) implements virtualization.

7. Audio device according to any one of the preceding claims, the processing chain (22) further comprising a module upmixer (23) arranged to produce the first audio channels (C1) from the input audio channels (C e ).

8. Audio device according to any one of the preceding claims, the processing chain (22) further comprising a module downmixer (27) arranged to produce third audio channels (C3) from second audio channels (C2), the output audio channels comprising the third audio channels (C3).

9. Equipment comprising an audio device according to one of the preceding claims.

10. Equipment according to claim 9, the equipment being a decoder box (10).

11. Processing method, implemented in the audio device (14) according to any one of claims 1 to 8, and comprising the steps of: - selecting at least one selected audio channel (C s) among the first audio channels, and control the extraction module (24) so ​​that it extracts the frequency components only from at least one selected audio channel, and not from at least one unselected audio channel (C ns ) ; - perform at least one processing operation that induces an inter-channel delay on at least one selected audio channel (C sp ) deprived of said frequency components and on at least one unselected audio channel (C ns ), so as to produce processed audio channels (Ct); - sum the said frequency components on each processed audio channel so as to produce second audio channels (C2) intended to be used to produce the output audio channels (C so ).

12. Computer program comprising program code instructions for executing the steps of the process according to claim 11 when said program is executed on a computer.

13. Computer-readable recording medium on which the computer program according to claim 12 is recorded.

Images