Signal processing method

Abstract

The invention relates to a signal processing method for converting X input signals into M output signals, comprising the following steps: a) applying a decorrelation filter to the X input signals to generate N decorrelated signals ei; b) for each decorrelated signal ei determined in step a), generating N delayed signals ei, k by applying a delay [tau] k and a gain gk to each decorrelated signal ei, selecting the delay [tau] k and the gain gk such that, for each decorrelated signal ei, the delayed signals ei, k simulate the propagation of the decorrelated signal in the virtual space, the virtual space comprises N virtual loudspeakers and virtual listening positions which are distributed according to a preset geometrical shape; c) for each k, summing the delayed signals ei, k (formula l): # imgabs0 # e) to determine M output signals, each output signal resulting from a linear combination of the sum Ek of the N delayed signals.

Description

Technical Field

[0001] The present invention relates to a signal processing method, in particular to a signal processing method for processing audio signals. Background Art

[0002] There are many music formats that define the structure of music content in computing terms. In particular, the music format is characterized by the number of channels. Therefore, depending on the music format, the music content is intended to be presented via a specific hardware configuration.

[0003] For example, music content in a stereo format comprising two channels is intended to be presented on two speakers, whereas music content in a 5.1 format comprising six channels is intended to be presented on six speakers.

[0004] When sound content in a particular format is not presented on the expected number of speakers, the result is a poor listening experience of the content. In particular, listening to stereo content on a home theater setup that traditionally includes 6 to 8 speakers does not allow a satisfactory immersive experience. The same applies to listening to 5.1, 7.1, surround sound or Dolby Digital content presented on headphones that include only two speakers. The case of music content in a MPEG-4 format (including four or more channels).

[0005] There are spreading algorithms that aim to output music content on multiple speakers. These algorithms make it possible, for example, to spread monophonic content on three or more speakers. However, even if these algorithms make it possible to obtain a greater effect of the spreading of the music content, the accuracy of the sound content is deteriorated due to the ambiguity added to the sound content.

[0006] The addition of reverberation is also a known method for outputting mono content on multiple speakers. This method consists in simulating listening to the sound content in a virtual room, taking into account the acoustic effects of the virtual room, in particular the echo caused by the virtual room. However, this method distorts the content by adding echoes that were not originally present.

[0007] US2021 / 352425 describes a method for processing a stereo signal, comprising forming a center channel signal. US2021 / 352425 describes a system for simulating an acoustic space, which aims to model the real acoustic effects of positioning.

[0008] The article "Virtual venues - an all-pass based time-variant artificial reverberation system for automotive applications" by Von Türckheim Friedrich et al. describes signal processing software for applications in the automotive sector. The article proposes a system that includes a simulated acoustic space, which aims to reproduce a real space.

[0009] US 2004 / 136554 discloses a method for processing a stereo signal intended to be output via two loudspeakers.

[0010] In general, it is not possible with known methods to define individually the number of channels of the sound content, the number of loudspeakers of the device intended to present the sound content, the echo, timbre and sound quality. Starting from a monophonic sound content, it is not possible to present this monophonic content via all of the loudspeakers of a 5.1 Home Cinema device, while still maintaining a precise and detailed perception of the positioning of the music content, while introducing a sense of sound immersion, and while maintaining the integrity of the sound content, i.e. the original artistic intention.

[0011] Therefore, there is a need to solve these problems. It is an object of the present invention to fully or partially solve this need. Summary of the invention

[0012] Therefore, a subject of the present invention is a signal processing method, which converts X input signals a p Converted into M output signals s q , p is an integer in the interval [1, X], q is an integer in the interval [2, M], the method comprises the following steps:

[0013] a) Apply the decorrelation filter to X input signals a p , to generate N decorrelated signals e i , i is an integer in the interval [1, N];

[0014] b) for each decorrelated signal e determined in step a) i , by adding the delay τ k and gain g k Applied to each decorrelated signal e i To generate N delayed signals e i,k , k is an integer in the interval [1, N], and the delay τ is selected k and gain g k , so that for each decorrelated signal e i , delayed signal e i,ksimulating the propagation of decorrelated signals in a virtual space, the virtual space comprising N virtual loudspeakers distributed according to a predetermined geometric shape and a virtual listening position,

[0015] c) For each integer value of k in the interval [1, N], the delayed signal e is calculated using the following formula i,k To sum:

[0016] [Mathematical formula 1]

[0017]

[0018] d) optionally, copying the delayed signal e determined in step b) i,k , and for each replica signal e i,k,j , j is an integer greater than or equal to 1, and the application delay ε i,k,j , then for each integer value of k in the interval [1, N], for each value of j, the delayed replica e is calculated using the following formula i,k (t-ε i,k,j ) to sum:

[0019] [Mathematical formula 2]

[0020]

[0021] e) Determine M output signals, each output signal s q The sum E of the N delayed signals from step c) k When executing step d), it is generated by the sum E of the N delayed signals from step c). k and the sum F of the j*N replica signals from step d) k,j The linear combination of .

[0022] The method according to the invention makes it possible in particular to present a music content comprising X input channels on a device comprising M physical loudspeakers. By means of the method according to the invention, M output signals can be obtained from the X input signals without these signals being altered, in particular without these signals being compressed or without any loss of accuracy.

[0023] The method according to the invention makes it possible to present a musical content in a format comprising X input channels on M physical loudspeakers, M being different from X, without compression or adding sound blur.

[0024] Preferably, no reflections or reverberations are added to the music content.

[0025] In one embodiment, M is greater than X.

[0026] In step a), the decorrelation filter may include an all-pass filter or a whitening filter, or a random delay may be applied to the input signal a p , for example by applying a relatively short delay, preferably less than 10 ms, or by applying a discrete cosine transform to the input signal a p .

[0027] The number of decorrelated signals generated in step a) may be between 1 and the larger of X and M; in other words, N may be an integer in the interval [1, max(M, X)].

[0028] In a preferred embodiment, N is equal to the maximum value of M and X, N=max(M,X).

[0029] A small number of decorrelated signals makes it possible to minimize the computational resources required to implement steps a) to e), while a large number of decorrelated signals makes it possible to improve the final listening quality. A number of decorrelated signals equal to the maximum of X and N advantageously minimizes the computational resources required while still having a satisfactory listening quality.

[0030] In step b), for each e i , determine the N delayed signals e i,k , for a fixed value of k, each delayed signal e i,k Able to communicate with virtual speakers through virtual space k Listen to the signal i , k is an integer in the interval [1; N]. For example, e 1,1 It can simulate listening to the decorrelated signals e1, e1 through the virtual speaker H1 in the virtual space. 1,1 Determined by applying the delay τ1 and gain g1 associated with the virtual loudspeaker H1.

[0031] In step b), each delay τ k Can be used with virtual speakers in virtual space k Similarly, each gain g k Can be used with virtual speakers in virtual space k Associated.

[0032] Each virtual speaker H in the virtual space can be k The distance from the virtual listening position determines the delay τ based on the speed of sound c in a vacuum k .

[0033] In one particular embodiment, the user may modify the number and / or positions of virtual speakers in the virtual space, and / or the virtual listening position.

[0034] The gain g can be calculated based on the distance between each virtual speaker in the virtual space and the virtual listening position. k The further a virtual speaker is from the virtual listening position, the lower the gain associated with the virtual speaker will be. The virtual listening position may be modified. In one embodiment, it is preset.

[0035] The virtual space may be an anechoic space.

[0036] In particular, the virtual space does not comprise any features intended to model real acoustic effects.In a preferred embodiment, the predetermined geometry of the virtual space is determined to minimize the Interaural Correlation Coefficient (hereinafter IACC).

[0037] IACC is an index of the similarity between the signals from the two ears, which can be used to model the spatial perception of the sound content. If the two signals from the two ears are called g(t) and d(t), then in the time domain, the interaural cross-correlation coefficient is defined as a function of the time offset δ by the following formula:

[0038] [Mathematical formula 3]

[0039]

[0040] The lower the IACC, the more immersed the user will feel when listening to the audio content.

[0041] As an alternative or in addition, the predetermined geometry is determined to minimize the variations in the frequency level caused by the frequency response of the delayed signal. Advantageously, such a geometry makes it possible to limit the variations that may be present in the delayed signal e. i,k Artifacts in the frequency response of .

[0042] Preferably, the virtual space is infinite or equivalent to an anechoic chamber, i.e. there is no echo in the virtual space; in other words, the virtual space is an anechoic space. In particular, the virtual space does not contain any obstacles that reflect signals. The absence of echoes makes it possible in particular to preserve the original artistic intention.

[0043] The method according to the invention may comprise, before step b), in particular before step a), selecting a predetermined geometry of the virtual space. Preferably, the predetermined geometry is selected by specifying the number of virtual loudspeakers of the virtual space.

[0044] Optional step d) has the advantage of increasing the sense of immersion by multiplying the signal amount around the delayed signal. The delayed signal is said to be thickened.

[0045] Preferably, J is equal to 1, thereby enhancing immersion while still limiting the computational power required to generate these replica signals.

[0046] In step e), the linear combination of the sum of the N delayed signals and the linear combination of the sum of the N delayed signals and the sum of the j*N replica signals when performing step d) may include fixed coefficients.

[0047] The coefficients may vary depending on the desired strength of each of the M output signals. In particular, the coefficients may depend on the actual device configuration comprising the M physical loudspeakers through which the M output signals are intended to be transmitted.

[0048] The coefficients may all be equal, so that the sound content is listened to via the real device and then a sense of omni-directional envelopment is given. Alternatively, the coefficients of the output signal intended to be sent via the physical speakers of the real device located in front of the user may be higher than the coefficients of the output signal intended to be sent via the physical speakers of the real device located behind the user, so that the user listens to the sound content via the real device and then a sense of frontal envelopment is given.

[0049] In general, the linear combination in step e) may include coefficients determined based on a desired sense of envelopment. The desired sense of envelopment may in particular be selected from the group consisting of lateral, frontal, rear and omnidirectional sense of envelopment, the list being non-limiting.

[0050] In a preferred embodiment, the coefficients are determined so that for a fixed value of k, for any value of j, the coefficients applied to E k The coefficient is equal to the coefficient applied to F k,j The coefficient of .

[0051] Step e) may include the following steps:

[0052] - For any value of k, for any value of j, E k and F k,j The sum is performed so that a vector of size (1, N) is obtained, which is represented by E k 、F k,j The sum of ;

[0053] - Multiply the vector by a matrix of size (N, M) determined based on the coefficients of the linear combination, which can be obtained by factoring k,q All coefficients are defined by k being an integer in the interval [1, N] and q being an integer in the interval [1, M].

[0054] The matrix can be defined by the following coefficients:

[0055] - For any value of k = q, α k,q is equal to 1,

[0056] -For any value of k different from q, α k,q Less than 1 and greater than or equal to 0.

[0057] When N=M, the matrix may be a unitary matrix.

[0058] When N = M, the matrix can be a circulant matrix, where α 1,1 is equal to 0, and for values ​​of q different from 1, α 1,q is less than 1 and greater than or equal to 0. In particular, the circulant matrix may include: α is less than 1 and greater than 0 1,2 , α less than 1 and greater than 0 1,q＝M , and α equal to 0 for values ​​of q different from 1, 2, and M 1,q , α 1,q＝M Preferably equal to α 1,2 .

[0059] For example, for N=M=5, the matrix can be equal to:

[0060] [Formula 4]

[0061] where g∈]0,1[

[0062] A factor of 1 can be replaced by applying a gain of 0 dB.

[0063] Coefficients less than 1 and greater than or equal to 0 may be replaced by applying a negative gain in decibels (dB).

[0064] The method according to the invention may comprise, before step e), in particular before step a), selecting the coefficients of the linear combination in step e), preferably by selecting a desired sense of envelopment.

[0065] When implemented, steps a), b), c), d) and e) may be computer-implemented.

[0066] In one embodiment, M is greater than or equal to 3 and X is equal to 2.

[0067] The method can be used to convert two input signals into at least three output signals, better at least four output signals, even better at least five output signals, even better at least six output signals. For presenting content in stereo format on loudspeakers in a motor vehicle or on loudspeakers in a concert hall or cinema or on loudspeakers in a home cinema installation, the use of a method for converting two input signals into three or more output signals is particularly advantageous.

[0068] In an alternative embodiment, M is equal to 2 and X is greater than or equal to 4.

[0069] The method can be used to convert at least four, at least five or preferably at least six input signals into two output signals. For content in (Dolby Atmos), ambisonic (surround sound), 5.1, 7.1 formats, it is particularly advantageous to use a method of converting four or more input signals into two output signals.

[0070] The invention also relates to the use of the method according to the invention for outputting:

[0071] - outputting a stereo music content comprising the two input signals via the loudspeakers of the motor vehicle, or

[0072] - Via home theater system or Dolby The system's speaker outputs include stereo music content from both input signals, the home theater system or Dolby The system comprises at least six loudspeakers, preferably at least eight loudspeakers, even more preferably at least ten loudspeakers, or

[0073] - outputting stereo music content comprising two input signals in a concert hall comprising four or more speakers, or

[0074] - outputting audio content comprising at least four input signals, preferably six input signals via headphones comprising two loudspeakers.

[0075] The invention also relates to a computer program comprising instructions which, when executed by a computer, cause the computer to implement the signal processing method according to the invention.

[0076] Preferably, the virtual space, in particular the predetermined geometry and / or gain g of the virtual space k and τ k The value is stored in the database.

[0077] A plurality of virtual space geometries may be stored in a database. Before step a) or before step b), a predetermined geometry of the virtual space for implementing step b) may be selected from the database.

[0078] A plurality of matrices determining the coefficients of the linear combination in step e) may be stored in a database. Prior to step e), in particular prior to step a), a matrix is ​​selected from the database for implementing step e) according to the invention.

[0079] The present invention also relates to a system, comprising:

[0080] - a computer program according to the invention,

[0081] - Optionally, a database comprising at least the geometry of the virtual space for implementing step b) and at least one matrix comprising coefficients of the linear combinations for implementing step e).

[0082] The system may include an interface that allows the user to set the geometry of the virtual space and / or the coefficients of the linear combination. In particular, the user may set the coefficients of the linear combination by selecting a desired sense of envelopment. In particular, the user may set the geometry of the virtual space by selecting the number of virtual speakers of the virtual space.

[0083] The system may include a microprocessor comprising a computer program according to the invention. The microprocessor may be embedded. The microprocessor may be contained in a phone, a television, a multimedia TV box, a car radio, a computer, a tablet computer, a smart watch. This list is not limiting.

[0084] The system according to the invention may also comprise M physical loudspeakers intended to send M output signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Other features and advantages of the present invention will become more apparent upon reading the following detailed description and studying the accompanying drawings, in which:

[0086] [ Figure 1 ] Figure 1 An exemplary implementation of the method according to the present invention is shown.

[0087] [ Figure 2a ] Figure 2a A plan view schematically showing an example of a predetermined geometric shape of virtual speakers and a virtual listening position in a virtual space including 8 virtual speakers,

[0088] [ Figure 2b ] Figure 2b Schematically shows the Figure 2a a side view of an example of a predetermined geometric shape,

[0089] [ Figure 3 ] Figure 3 Another exemplary implementation of the method according to the present invention is shown, and

[0090] [ Figure 4 ] Figure 4 A system for implementing the method according to the invention is schematically shown. DETAILED DESCRIPTION

[0091] Figure 1 An exemplary implementation of the signal processing method according to the present invention is shown, wherein the number of input signals is equal to 2 (X=2) and the number of output signals is equal to 4 (M=4).

[0092] Step a)

[0093] In step a), a decorrelation filter 10 is applied to the X input signals a1, a2 in order to generate N decorrelation signals e i , i is an integer in the interval [1, N].

[0094] Preferably, the number of decorrelated signals generated is equal to the maximum value of (X, M). Figure 1 In the example of , the number of decorrelated signals is 4, and N is equal to the maximum value of (X, M) = max(2, 4).

[0095] The decorrelation filter 10 may comprise an all-pass filter, a whitening filter, or may comprise applying a random delay to the input signals a1, a2, preferably less than 10 ms, or may comprise applying a discrete cosine transform to the input signals a1, a2. Preferably, the decorrelation filter 10 is an all-pass filter.

[0096] Step a) can be implemented by the computing module 10 .

[0097] Step b)

[0098] The decorrelated signals e1, e2, e3, e4 obtained at the end of step a) are then in step b) by applying a delay τ k and gain g k Delayed and amplified:

[0099] [Formula 5]

[0100]

[0101] exist Figure 1 In the example, for example, the following delayed signal e is obtained i,k :

[0102] For i=1:

[0103] e 1,1 (t) = g1*e1(t-τ1)

[0104] e 1,2 (t) = g2*e1(t-τ2)

[0105] e 1,3 (t) = g3*e1(t-τ3)

[0106] e 1,4 (t) = g4*e1(t-τ4)

[0107] For i=2:

[0108] e 2,1 (t) = g1*e2(t-τ1)

[0109] e 2,2 (t) = g2*e2(t-τ2)

[0110] e 2,3 (t) = g3*e2(t-τ3)

[0111] e 2,4 (t) = g4*e2(t-τ4);

[0112] For i=3:

[0113] e 3,1 (t) = g1*e3(t-τ1)

[0114] e 3,2 (t) = g2*e3(t-τ2)

[0115] e 3,3 (t) = g3*e3(t-τ3)

[0116] e 3,4 (t) = g4*e3(t-τ4)

[0117] And for i=4:

[0118] e 4,1 (t) = g1*e4(t-τ1)

[0119] e 4,2 (t) = g2*e4(t-τ2)

[0120] e 4,3 (t) = g3*e4(t-τ3)

[0121] e 4,4 (t) = g4*e4(t-τ4).

[0122] Each delay τ k Can be used with virtual speakers in virtual space k Similarly, each gain g k Can be used with virtual speakers in virtual space k Associated.

[0123] In other words, the delayed signal e 1,1 、e 2,1 、e 3,1 、e 4,1 The simulations come from listening to the decorrelated signals e1, e2, e3, e4 through virtual loudspeakers H1, H2, H3, H4, respectively, in the virtual space.

[0124] Delay τ k and gain g k Depends on the geometry of the virtual space. To determine the delay τ k and gain g k , so it is necessary to determine the geometry of the virtual space.

[0125] Virtual Space

[0126] The determination of the virtual space may result from the determination of the positions of virtual speakers and a virtual listening position of the virtual space.

[0127] The determination of the virtual space may be determined to minimize the IACC.

[0128] The virtual space can be determined to minimize the delay caused by the signal e i,k The frequency response causes the frequency level changes.

[0129] In a preferred embodiment, the virtual space is determined to minimize IACC and the delay signal e i,k The frequency response causes the frequency level changes.

[0130] The virtual space is preferably an anechoic space, ie the virtual space does not contain any obstacles that reflect signals. In particular, the virtual space is advantageously not a partially enclosed or closed virtual room. Among other things, the absence of echoes allows the original artistic intent to be preserved.

[0131] In one particular embodiment, the user may modify the number and / or positions of virtual speakers in the virtual space, and / or the virtual listening position.

[0132] According to the desired number of virtual speakers, a plurality of virtual spaces are preferably preset.

[0133] Before step b), in particular before step a), the user may determine the geometry of the virtual space, for example by selecting a predetermined geometry from a database.

[0134] Depending on the number N of decorrelated signals, a predetermined virtual space is selected from the set of virtual spaces based on the number of virtual speakers of the virtual spaces in the set of virtual spaces, each virtual space in the set of virtual spaces including a single number of virtual speakers.

[0135] Delay

[0136] Delay τ k It can be determined based on the distance between each virtual speaker and the virtual listening position in the virtual space and based on the speed of sound c in a vacuum, such as Figure 2a and Figure 2bshown.

[0137] Figure 2a and Figure 2b A virtual space 3 comprising eight virtual loudspeakers 32 and a virtual listening position 34 is schematically shown. Figure 2a shows a plan view of the virtual space 3, Figure 2b A side view of the virtual space 3 is shown. k The value of τ can be determined as follows: k =d k / c, where c is the speed of sound in vacuum.

[0138] for Figure 2a and Figure 2b In the example of , the distances d1, d2, d3, and d4 between the virtual listening position 34 and the virtual loudspeakers H1, H2, H3, and H4 are defined based on the parameter y, respectively. k The following values:

[0139] [Mathematical formula 6]

[0140]

[0141] τ1 is associated with the virtual speaker H1 and is calculated based on the position of the virtual speaker H1 relative to the virtual listening position 34, τ2 is associated with the virtual speaker H2 and is calculated based on the position of the virtual speaker H2 relative to the virtual listening position 34, τ3 is associated with the virtual speaker H3 and is calculated based on the position of the virtual speaker H3 relative to the virtual listening position 34, and τ4 is associated with the virtual speaker H4 and is calculated based on the position of the virtual speaker H4 relative to the virtual listening position 34.

[0142] Gain

[0143] In step b), the gain g may be calculated based on the distance between each virtual speaker in the virtual space and the virtual listening position. k .

[0144] The gain may be determined according to the following criteria: the further a virtual speaker is from the virtual listening position, the lower the gain associated with the virtual speaker will be.

[0145] The gain may be inversely proportional to the distance between the virtual speaker and the virtual listening position.

[0146] A gain of 0 dB may be associated with the virtual speaker closest to the virtual listening position.

[0147] The gains may be determined by applying an affine function based on the distance between each virtual speaker and the virtual listening position.

[0148] The gain g can be calculated by applying a function f that satisfies, for example, the following relationship: k : For any distance d, f(2d)=f(d)-6dB. Preferably, a gain of 0dB is fixed for the distance between the virtual speaker closest to the virtual listening position and the virtual listening position.

[0149] At the end of step b), this gives a delayed signal that is N times the number of virtual loudspeakers in the virtual space. Figure 1 This gives 16 delayed signals: 1,1 、e 1,2 、e 1,3 、e 1,4 、e 2,1 、e 2,2 、e 2,3 、e 2,4 、e 3,1 、e 3,2 、e 3,3 、e 3,4 、e 4,1 、e 4,2 、e 4,3 、e 4,4 .

[0150] Step b) may be implemented by the computing module 12 .

[0151] Preferably, the geometry of the virtual space and / or the gains and delays resulting from this geometry are stored in a database.

[0152] Step c)

[0153] In step c), the delayed signal e is calculated using the following formula: i,k To sum:

[0154] [Mathematical formula 7]

[0155]

[0156] Each E k Corresponding to the virtual speaker H through the virtual space k Listen to the decorrelated signal i The sum of the delayed signals generated by the simulation.

[0157] Step c) may be implemented by the computing module 14 .

[0158] Step d)

[0159] In a preferred embodiment, the method according to the present invention comprises an optional step d). The optional step d) can be performed in parallel with step c). The optional step d) is performed after step b).

[0160] Optional step d) allows to enhance the immersion by multiplying the number of signals around the delayed signal. The larger J is, the greater the immersion.

[0161] Optional step d) comprises applying a delay ε i,k,j For the delayed signal e i,k J copies are made, where J is an integer greater than or equal to 1. In one embodiment, J is equal to 1, which enhances immersion while still limiting the computational power required to generate these copies.

[0162] By applying a delay ε to the delayed signal i,k,j The obtained replica signal oscillates around the delayed signal. The applied delay may be of the order of about 100 μs.

[0163] The replicated signal is obtained by applying the following formula:

[0164] [Mathematical formula 8]

[0165] f i,k,j (t) = e i,k (t-ε i,k,j ),j is an integer, j≥1.

[0166] Optional step d) further comprises replicating the delay signal e i,k , use the following formula to calculate the copied signal f i,k,j To sum:

[0167] [Mathematical formula 9]

[0168]

[0169] The optional step d) may be likened to adding J virtual speakers around the virtual speaker in the virtual space.

[0170] A gain can be applied to the replica signal f i,k,j and / or applied to the sum of the replica signals F k,j .

[0171] In one embodiment, a single gain is applied to the replica signal f depending on the desired level of the replica signal relative to the delayed signal, in particular the desired level of the sum of the replica signals relative to the sum of the delayed signals. i,k,j and / or the sum of the replica signals F k,j .

[0172] When implementing step d), the user may determine the number of copies J to be performed in step d) before step b), in particular before step a).

[0173] The optional step d) may be implemented by the calculation module 16 .

[0174] Step e)

[0175] In step e), the sum E of the signals respectively generated by step c) and optionally step d) is combined by linear combination k and optionally F k,j .

[0176] The coefficients of the linear combinations can be determined manually by the user.

[0177] Preferably, the coefficients are predetermined according to the required strength for each of the M output signals.The coefficients may depend on the actual installation configuration comprising the M physical loudspeakers through which the M output signals are intended to be sent.

[0178] In one embodiment, the user selects a desired sense of envelopment, and this selection of a desired sense of envelopment enables the coefficients to be fixed.

[0179] For example, when the coefficients are all equal, an all-round sense of envelopment is given. Alternatively, the coefficients of the output signals intended to be sent via the physical speakers of the real device located in front of the user may be higher than the coefficients of the output signals intended to be sent via the physical speakers of the real device located behind the user, thereby giving a frontal sense of envelopment. In particular, for a frontal sense of envelopment, a gain of 0 dB may be applied to the output signals intended to be sent via the physical speakers of the real device located in front of the user, a gain of -6 dB may be applied to the output signals intended to be sent via the physical speakers of the real device located on the sides, and a gain of -12 dB may be applied to the output signals intended to be sent via the physical speakers of the real device located behind the user.

[0180] The desired sense of enclosure can be selected from side, front, back and all-round enclosure.

[0181] The coefficients can be pre-selected, which defines the coefficients that are selected by default if no selection is made by the user.

[0182] Before step e), in particular before step a), the user may determine the coefficients of the linear combination for implementing step e), for example by selecting the coefficients from a database. Preferably, the coefficients are selected by selecting the desired sense of envelopment, for example from a database.

[0183] Step e) may be implemented by the computing module 18 .

[0184] The coefficients may be stored in a database 200 .

[0185] Figure 3Another exemplary implementation of the signal processing method according to the present invention is shown, wherein the number of input signals is equal to 6 (X=6) and the number of output signals is equal to 2 (M=2).

[0186] These steps are similar to those for Figure 1 The examples describe those steps.

[0187] In this example, N is also equal to the maximum of M and X, which is 6 here.

[0188] In this example, J is equal to 1, that is, the delayed signal is copied once.

[0189] Figure 4 A system for implementing the method according to the invention is schematically shown.

[0190] System 2 includes:

[0191] - a computer 100 comprising a computer program comprising instructions which, when executed by a computer, cause the computer to implement steps a), b), c) and e) of the signal processing method according to the invention, preferably to implement steps a), b), c), d) and e) of the signal processing method according to the invention,

[0192] - a database 200 containing at least one virtual space 3 geometry for implementing step b) of the method according to the invention and / or at least one matrix determining the coefficients of a linear combination for implementing step e) of the method according to the invention;

[0193] - Optionally, M physical loudspeakers 300, intended to render M output signals.

[0194] The database 200 may be integrated into the computer 100 .

[0195] The system may include communication means 400 allowing the database to communicate with the computer 100 .

[0196] A computer is understood to be any computing device for executing instructions.

[0197] The method according to the invention can advantageously be used to output:

[0198] - outputting a stereo music content comprising the two input signals via the loudspeakers of the motor vehicle, or

[0199] - Via home theater system or Dolby The system's speaker outputs include stereo music content from both input signals, the home theater system or Dolby The system includes at least six speakers, preferably eight speakers, and even more preferably ten speakers, or

[0200] - outputting stereo music content comprising the two input signals in a concert hall preferably comprising four or more loudspeakers, or

[0201] - outputting audio content comprising at least four input signals, preferably six input signals via headphones comprising two loudspeakers.

[0202] These usage examples are not limiting.

[0203] Of course, the invention is not limited to the exemplary embodiments that have just been described.

[0204] In particular, the number of input signals and the number of output signals may vary. They advantageously depend on the desired use.

[0205] The invention makes it possible in particular to ensure an improved spreading of the X input signals, that is to say to preserve the integrity of the signals and to limit the spreading of artefacts.

[0206] Unlike the known prior art, the purpose of the present invention is not to reproduce the acoustics of a room or real positioning. In fact, the addition of early reflections and / or late reverberation, although having an impact on the immersion, modifies the original artistic intention. In particular, early reflections add a corresponding colour to the geometry of the room and the acoustic quality of the walls of said room. The sound to which such reflections are added may then become muffled or, on the contrary, very clear.

[0207] The process of adding delay reverb modifies the perceived sound depth and echo. Short, percussive sounds to which delay reverb is added are perceived as long and drowned in the reverb.

[0208] In addition, pre-processing operations and / or post-processing operations can be applied to the input and / or output signals, respectively. For example, an equalizer filter, a finite impulse response filter, or a second-order infinite impulse response filter can be applied to the input signal a. p and / or output signals q , to modify the input signal a respectively p and / or output signals q timbre.

Claims

1. A signal processing method, wherein X input signals a p Converted into M output signals s q , p is an integer belonging to the interval [1, X], q is an integer belonging to the interval [2, M], M is different from X, the method comprises the following steps: a) For the X input signals a p Apply the decorrelation filter (10) to generate N decorrelation signals e i , i is an integer in the interval [1, N]; b) for each decorrelated signal e determined in step a) i , by adding the delay τ k and gain g k Applied to each decorrelated signal e i To generate N delayed signals e i,k , k is an integer belonging to the interval [1, N], and the delay τ is selected k and the gain g k , so that for each decorrelated signal e i , the delayed signal e i,k Simulating the propagation of the decorrelated signal in a virtual space (3), wherein the virtual space includes N virtual loudspeakers (32) distributed according to a predetermined geometric shape and a virtual listening position (34), and the virtual space is an anechoic space; c) For each integer value of k belonging to the interval [1, N], the delayed signal e is calculated using the following formula: i,k To sum: d) copying the delayed signal e determined in step b) i,k , and for each replica signal e i,k,j , j is an integer greater than or equal to 1, and the application delay ε i,k,j , then for each integer value of k in the interval [1, N], for each value of j, the delayed replica signal e is calculated using the following formula i,k (t-ε i,k,j ) to sum: e) Determine M output signals, each output signal s q The sum E of the N delayed signals from step c) k When executing step d), it is generated by the sum E of the N delayed signals from step c). k and the sum F of the j*N replica signals from step d) k,j The linear combination of .

2. The method according to the preceding claim, the virtual space (3) being an anechoic space.

3. A method according to any one of the preceding claims, wherein: The predetermined geometry of the virtual space (3) is determined so as to minimize the inter-aural correlation coefficient.

4. A method according to any one of the preceding claims, wherein: The predetermined geometry of the virtual space (3) is determined to minimize the change in frequency level caused by the frequency response of the delayed signal.

5. A method according to any one of the preceding claims, wherein: The decorrelation filter comprises an all-pass filter or a whitening filter, or p Apply random delay, or to the input signal a p Apply a discrete cosine transform.

6. A method according to any one of the preceding claims, wherein: The linear combination in step e) includes coefficients determined based on a desired sense of envelopment, wherein the desired sense of envelopment is selected from the group consisting of frontal, side, rear and omnidirectional sense of envelopment.

7. A method according to any one of the preceding claims, wherein: N is an integer in the interval [1, maximum(X, M)], preferably N is equal to the maximum value of X and M.

8. Method according to any of the preceding claims, comprising, before step a), selecting a predetermined geometry of the virtual space (3), preferably by specifying a number of virtual loudspeakers (32) of the virtual space.

9. The method according to any one of the preceding claims comprises, before step a), preferably selecting the coefficients of the linear combination in step e) by selecting a desired sense of envelopment, wherein the linear combination in step e) comprises coefficients determined based on the desired sense of envelopment, wherein the desired sense of envelopment is selected from the front, side, back and all-round sense of envelopment.

10. Use of the method according to any one of the preceding claims for outputting: - outputting a stereo music content comprising the two input signals via the loudspeakers of the motor vehicle, or - Via home theater system or Dolby The system's speaker outputs include stereo music content from both input signals, the home theater system or Dolby The system comprises at least six loudspeakers, preferably at least eight loudspeakers, even more preferably at least ten loudspeakers, or - outputting stereo music content comprising two input signals in a concert hall comprising four or more speakers, or - outputting audio content comprising at least four input signals, preferably six input signals via headphones comprising two loudspeakers.

Images