Multi-channel audio data processing apparatus for a motor vehicle

Abstract

The invention relates to a multi-channel audio data processing device for a motor vehicle, said device comprising a data memory (4) receiving multi-channel audio data, said multi-channel audio data comprising rank audio data comprising a first audio channel and a second audio channel, and for each rank a first rank digital filter and a second rank digital filter. The first rank digital filter and the second rank digital filter of a given rank are determined from a transformation into a rank correction data digital filter for each loudspeaker of the rank. The rank correction data results from phase measurements on two signals, on the one hand resulting from a reproduction of the first audio channel and on the other hand a reproduction of the second audio channel, each time with one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, these phase measurements being made for each position. The correction data is obtained according to the formula C i = D i + P i + SF i where D is a balance factor, P is a phase inversion factor P, SF is a correction position adjustment factor. A computer (120) applies each first rank digital filter to all or a portion of the first audio channel of the rank audio data and each second rank digital filter to all or a portion of the second audio channel of the rank audio data.

Description

[0001] The sound quality inside a vehicle is an ever-evolving field. For decades, content has largely been stereo, and now includes left and right channels (and sometimes a center channel). The number of speakers inside a vehicle has increased over time, and now there are speakers in each row of seats. Therefore, these channels are typically associated with a given row—either the front or rear in a conventional vehicle—but there may be more rows in a van.

[0002] Improvements in sound quality are accompanied by a continuous increase in the computing power required to achieve the desired audio processing. Typically, this processing is part of a set of in-vehicle processes that also consume computing power. Therefore, in recent years, the pursuit of audio quality has been accompanied by the need to find trade-offs in terms of available computing power, ensuring that all audio processing falls within the limits of available computing capabilities.

[0003] Sound field localization, i.e. the position of the instruments and their timbre, is currently being corrected in ways that the applicant finds unsatisfactory: bass booming, spatial blurring of vocals and instruments, energy gaps in sound field localization, vocal splitting, etc.

[0004] These problems are primarily related to phase and timing issues in automotive cabin acoustics, which have not yet been corrected. Furthermore, correcting these issues for each individual car seat (in the presence of multiple listening schemes) adds an additional, extremely difficult challenge.

[0005] Hardware solutions exist, which will not be detailed here, but are not relevant to the context of this invention, as this invention relates to software solutions. The applicant has developed an algorithm called Stage, which is capable of addressing part of the aforementioned problem by modifying the phase of a specific frequency band. However, the applicant believes that this solution can be further improved.

[0006] It is also possible to use a manually configured infinite impulse response (IIR) full-pass filter. However, this type of solution is approximate and has very limited accuracy.

[0007] The present invention improves upon this situation. To this end, the present invention provides a multi-channel audio data processing apparatus for a motor vehicle, comprising a data storage unit arranged to receive multi-channel audio data, and a first row digital filter and a second row digital filter for each row, the multi-channel audio data including row audio data associated with a corresponding speaker row, the multi-channel audio data being associated with at least one row, and the row audio data associated with a given speaker row including a first audio channel and a second audio channel. The first and second row digital filters for a given row are determined based on correction data including row correction data for each row of speakers associated with the multichannel audio data. The row correction data associated with the given speaker row originates from phase measurements of two signals, one originating from the reproduction of a first audio channel in the given row of speakers, and the other from the reproduction of a second audio channel in the given row of speakers, each using one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz. These phase measurements are performed, on the one hand, at a first position associated with the speakers of the given row and for determining a first set of row phase differences, where each element is associated with the phase difference of a measurement obtained at the first position for the first and second audio channels at the same corresponding frequency; and on the other hand, at a second position associated with the speakers of the given row and for determining a second set of row phase differences, where each element is associated with the phase difference of a measurement obtained at the second position for the first and second audio channels at the same corresponding frequency. Each element of the row correction data is associated with one of the plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and each element C i Associated with one of the plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and according to Formula C i = D i +P i + S F i We obtain, where i is the index of a given frequency among the plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and D i According to the formula Calculate, where R1 i It is the value of the element at index i in the first group of phase differences, R2 i It is the value of the element at index i in the second group of phase differences, P i It is a phase correction value equal to 0 or π, and depends on cos R1 i + cos R2 i The sum of values, S is the distribution factor between -1 and 1, and F i According to formula (R1) i R2 i + π)mod 2π π calculation. The first row digital filter and the second row digital filter are obtained by transforming the row correction data into a finite impulse response filter or an infinite impulse response filter. The apparatus also includes a computer arranged to apply each of the first row digital filters of a given row to all or a portion of the first audio channel of the row audio data associated with the given row, and to apply each of the second row digital filters of a given row to all or a portion of the second audio channel of the row audio data associated with the given row.

[0008] This device is particularly advantageous because it can correct phase alignment between different frequencies for each row of a vehicle under limited computing power. Unlike the Stage algorithm or manually setting an all-pass filter, this solution is metrologically based and utilizes a measurement system to measure and correct phase problems in the car. Furthermore, the correction can be customized and balanced between a first and a second position, or centered on either the first or the second position, at the expense of the other.

[0009] According to various embodiments, the present invention can have one or more of the following features:

[0010] - The first and second row digital filters are determined based on row correction data, P i The value is calculated according to the following formula:

[0011] Where h is a lag factor selected from the range [0; 0.2]. - The computer was also configured as follows: On one hand, it receives input audio data as input, in order to separate it into low-frequency audio data with frequencies below a threshold between 1.5 kHz and 5 kHz, and high-frequency audio data with frequencies above a threshold between 1.5 kHz and 5 kHz. On the other hand, the output audio data is returned, which includes: a first audio channel, which combines the signal from the first audio channel of the high-frequency audio data and the signal from the first audio channel of the low-frequency audio data corrected by the first row digital filter; and a second audio channel, which combines the signal from the second audio channel of the high-frequency audio data and the signal from the second audio channel of the low-frequency audio data corrected by the second row digital filter. The computer is configured to undersample the low-frequency audio data in order to, on the one hand, apply a first row of digital filters to the first audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal originating from the first audio channel of the low-frequency audio data corrected by the first row of digital filters; and on the other hand, apply a second row of digital filters to the second audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal originating from the second audio channel of the low-frequency audio data corrected by the second row of digital filters. - A computer is configured to generate a signal from a first audio channel derived from high-frequency row audio data and a signal from a second audio channel derived from high-frequency row audio data, the signals having a delay selected according to row correction data, and - The first row of digital filters and the second row of digital filters are obtained by transforming them into finite impulse response filters according to the allocation of row correction data, such that C i (First audio channel) = a C i And C i (Second audio channel) = (1 a) C i , where a is a real number between 0 and 1.

[0012] The present invention also relates to a multi-channel audio signal processing method for a motor vehicle, comprising the following operations: a) receiving multi-channel audio data, the multi-channel audio data including row audio data associated with a corresponding speaker row, the multi-channel audio data being associated with at least one row, and the row audio data associated with a given speaker row including a first audio channel and a second audio channel; b) obtaining a first row digital filter and a second row digital filter for each row, the first row digital filter and the second row digital filter for a given row being determined based on correction data including row correction data of each row of speakers associated with the multi-channel audio data, the row correction data associated with the speakers of the given row being derived from phase measurements of two signals, the two signals being derived, on the one hand, from the reproduction of the first audio channel in the speakers of the given row, and on the other hand, from the reproduction of the second audio channel in the speakers of the given row, each time using a frequency of less than 1.5 kHz to 5 kHz. The phase measurements are performed, on the one hand, at a first position associated with a loudspeaker in a given row, and are used to determine a first set of row phase differences, wherein each element is associated with the phase difference of a measurement obtained at the first position for the first and second audio channels at the same corresponding frequency, and on the other hand, at a second position associated with a loudspeaker in a given row, and are used to determine a second set of row phase differences, wherein each element is associated with the phase difference of a measurement obtained at the second position for the first and second audio channels at the same corresponding frequency, and each element of the row correction data is associated with one of the plurality of frequencies between a threshold of less than 1.5 kHz and 5 kHz, and each element C i Associated with one of the plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and according to Formula C i = D i + P i + S F i We obtain, where i is the index of a given frequency among the plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and D i According to the formula Calculate, where R1 i It is the value of the element at index i in the first group of phase differences, R2 i It is the value of the element at index i in the second group of phase differences, P i It is a phase correction value equal to 0 or π and depends on cos R1 i + cosR2 i The sum of values, S is the distribution factor between -1 and 1, and F i According to formula F i = (R1 i R2 i + π) mod 2π π calculation, the first row digital filter and the second row digital filter are obtained by transforming the row correction data into a finite impulse response filter or an infinite impulse response filter, and c) applying each of the first row digital filters of a given row to all or part of the first audio channel of the row audio data associated with the given row, and applying each of the second row digital filters of a given row to all or part of the second audio channel of the row audio data associated with the given row.

[0013] According to various implementation schemes, the method can have one or more of the following features: - The first and second row digital filters in operation b) are determined based on the row correction data, P i The value is calculated according to the following formula:

[0014] Where h is a lag factor selected from the range [0; 0.2]. Operation a) includes separating the row audio data into low-frequency row audio data with frequencies below the threshold between 1.5 kHz and 5 kHz and high-frequency audio data with frequencies above the threshold between 1.5 kHz and 5 kHz, and operation c) includes returning the row audio data, which includes: a first audio channel that combines signals from the first audio channel of the high-frequency row audio data and signals from the first audio channel of the low-frequency row audio data corrected by a first row digital filter; and a second audio channel that combines signals from the second audio channel of the high-frequency row audio data and signals from the second audio channel of the low-frequency audio data corrected by a second row digital filter. - Operation c) includes: c1) undersampling the low-frequency audio data to, on the one hand, apply a first row digital filter to the first audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal from the first audio channel of the low-frequency audio data corrected by the first row digital filter; and on the other hand, apply a second row digital filter to the second audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal from the second audio channel of the low-frequency audio data corrected by the second row digital filter; and c2) combining the signal from the first audio channel of the low-frequency audio data corrected by the first row digital filter with the first audio channel of the high-frequency audio signal; and on the other hand, combining the signal from the second audio channel of the low-frequency audio data corrected by the second row digital filter with the second audio channel of the high-frequency audio signal to generate the returned audio data. - Operation c) includes generating a signal from a first audio channel derived from high-frequency stacked audio data and a signal from a second audio channel derived from high-frequency stacked audio data, the signals having a delay selected according to stacked correction data, and - The first row of digital filters and the second row of digital filters are obtained by transforming them into finite impulse response filters according to the allocation of row correction data, such that C i (First audio channel) = a C i And C i (Second audio channel) = (1 a) C i , where a is a real number between 0 and 1.

[0015] The present invention also relates to a computer program implemented by a computer, comprising instructions for performing the methods of the present invention, and to a data storage medium on which such a computer program is recorded.

[0016] Other features and advantages of the invention will become clearer from the following description and with reference to the examples and drawings given for illustrative and non-limiting purposes, wherein: - Figure 1 A schematic top view of a vehicle implementing the device of the present invention is shown; and - Figure 2 It shows Figure 1 A general block diagram of the devices used.

[0017] The accompanying drawings and descriptions below largely contain elements with specific properties. Therefore, they can not only be used to better understand the invention, but also, where appropriate, help to define the invention.

[0018] Figure 1 A schematic top view of a vehicle implementing the apparatus of the present invention is shown. As shown in the figure, the vehicle 2 has two rows of seats: a front row 4 and a rear row 6. The front row 4 includes a driver's seat 8 and a passenger seat 10. The rear row 6 includes a left rear seat 12 and a right rear seat 14. Alternatively, the rear row 6 may have more seats, and the vehicle may have a third or more rows. As another alternative, there may be only one row of seats.

[0019] In the example described herein, the front row 4 includes a left front speaker 16 and a right front speaker 18, while the rear row 6 includes a left rear speaker 20 and a right rear speaker 22.

[0020] The applicant, in seeking to improve the Stage algorithm, became interested in measurements taken inside a vehicle. This is because obtaining a broad and consistent sound field at a given location largely depends on the fact that, for a given signal, the frequencies it contains arrive at that location simultaneously. However, due to sound processing operations and the unique shape of the vehicle interior, frequencies arrive at the ear in different ways and propagate in different manner. This is particularly evident in the channel-alike portions of multichannel audio signals, and these portions are often referred to as the "mono-channel portion" or "correlated portion."

[0021] In this process, the applicant found that processing the phase row by row was beneficial because the seats in a given row primarily receive the sound path from the speakers in that row. Therefore, the applicant first measured a linear frequency sweep for each audio channel of each row at each position.

[0022] Therefore, vectors representing the signal phase arriving at each location at a large number of frequencies can be obtained. More specifically, these vectors can be named by specifying the measurement location, the row to which that location belongs, and the channel from which it originates.

[0023] For example, with position 1 in the first row (i.e. Figure 1 The vector associated with the driver's seat 8 in the first row and the first audio channel (e.g., the left channel) in the first row is denoted as FL1, and the vector associated with position 1 in the first row (i.e., Figure 1 The vector associated with the driver's seat 8 in the first row and the second audio channel (e.g., the right channel) in the first row is denoted as FR1, and the vector associated with position 2 in the first row (i.e., Figure 1 The vector associated with passenger seat 10 in the first row and the first audio channel (e.g., the left channel) of the first row is denoted as FL2, and the vector associated with position 2 in the first row (i.e., Figure 1 The vector associated with the passenger seat 10 and the second audio channel (e.g., the right channel) of the first row is denoted as FR2. For the second row, the naming convention is similar, but the first letter of the vector will be changed to indicate that row (e.g., "G" or any other letter).

[0024] Each vector includes phase measurements for multiple frequencies below a selected threshold. The applicant found that a threshold between 1.5 kHz and 5 kHz yielded the best results.

[0025] In fact, the phase alignment effect is particularly noticeable at the lowest frequencies, but becomes less pronounced above 5 kHz: perceiving the phase effect at this frequency requires almost perfect hearing, while keeping the head within 5 cm of its maximum position—and the higher the frequency, the lower the accuracy of the human ear, and the smaller the permissible head movement. Keeping the head within 5 cm while driving in a motor vehicle is already practically impossible.

[0026] As shown below, the filter corresponding to the correction calculated using phase vector measurement can be implemented using a finite impulse response filter (“FIR”) or an infinite impulse response filter (“IIR”).

[0027] The threshold determines, on the one hand, the frequency at which phase alignment is performed, and on the other hand, the filter size required to achieve this phase alignment, according to the Nyquist condition. However, for comparable correction accuracy, higher frequencies require more coefficients in the filter and greater computational power to implement it. Lower thresholds require less computational power but result in poorer correction. The applicant determined that correction no longer possesses favorable quality when the threshold is below 1.5 kHz. Therefore, the range between 1.5 kHz and 5 kHz was determined to be the optimal trade-off for implementing the invention. Preferably, the threshold range can be between 2 kHz and 3 kHz, with a preferred value of 3 kHz, as this represents the optimal trade-off in terms of computational power / correction quality in the applicant's tests. As shown below, correction is implemented using a digital filter. The applicant found that, preferably, the resolution of the phase measurement used for correction is greater than or equal to the resolution of the filter.

[0028] After measuring the phase vectors, the applicant investigated a solution for correcting phase alignment at each position in a given row. Their research showed that processing the phase difference at each position is advantageous. Therefore, according to the formulas R1 = FL1 - FR1 and R2 = FL2 - FR2, vectors FL1, FR1, FL2, and FR2 are combined into two phase difference vectors R1 and R2. This means that vector R1 contains the phase difference between the first and second audio channels at position 1 in the first row. Similarly, vector R2 contains the phase difference between the first and second audio channels at position 2 in the first row.

[0029] Subsequently, the phase alignment of each vector can be corrected by applying corrections based on three correction factors: a) a phase inversion factor P, which includes adding a phase correction of π for each frequency if the sum of the cosines of the phase differences between the two positions is less than 0; b) a balance adjustment factor D; and c) a correction position adjustment factor S. F.

[0030] The applicant discovered that the phase reversal factor P can be determined in two different ways: with hysteresis and without hysteresis. The correction without hysteresis is based on the assumption that the measurement is perfectly reliable. In this case, each element P of the phase reversal vector P... i The value is determined according to the following formula: - If cosR1 i + cosR2 i If ≤ 0, then P i= π; - If cos R1 i + cos R2 i >0, then P i = 0.

[0031] However, the phase measurement and the resulting phase difference are inevitably subject to noise. This means that in cosR1 i +cosR2 i When the sum is close to 0, the P of adjacent frequencies i The value may vary significantly, solely due to measurement noise, which can introduce artifacts during calibration. Even if the measurement is perfect, if cosR1 i + cosR2 i The sum changes very rapidly near zero, which can cause problems in digital filter design. This is because continuously changing a digital filter in such a fast manner is complex, or even impossible.

[0032] Therefore, the applicant determines the lag correction according to the following formula:

[0033] This means that, near 0 and within the width range h, P i The value of remains unchanged to maintain the stability of the phase correction. Therefore, the absolute value of cosR1 exceeding h... i + cosR2 i The value of h must be changed from 0 correction to π correction, or from π correction to 0. Therefore, some frequencies that should be corrected will not be corrected, but artifacts caused by measurement noise can be avoided. The applicant determined that h values ​​selected from the range [0; 0.2] yielded optimal results. Incidentally, h equal to 0 corresponds to a formula without hysteresis.

[0034] The balance adjustment factor D is determined according to the following formula:

[0035] The adjustment factor D ensures symmetry in performance between the left and right seats: reconstructing the right and left channels in both seats will be identical. Perceptually, this makes the sound field quality and frequency response as similar as possible across different locations in the car.

[0036] Finally, the position adjustment factor S is corrected. F enables adjustment of phase correction, which is beneficial for correcting specific positions in the row. First, this factor is obtained by determining the vector F, where the element F... i According to formula F i = (R1 i R2i + π)mod 2π π is calculated, and the adjustment scalar S is applied to vector F.

[0037] If S is 0, the perceived phase correction is the same for both the first and second positions. If S is 1, the phase correction will be optimal for position 1. If S is -1, the phase correction will be optimal for position 2. Therefore, this factor allows for adjustment of the perceived phase correction quality for a given position. Thus, using an S value of 1 would be highly advantageous when driving alone. Similarly, switching between multiple filters corresponding to different S values ​​can be advantageous depending on the situation: for example, in cases where audio programs are interrupted to provide information (such as traffic information on the radio or interruption notices in GPS navigation), using a filter with S equal to 1 to facilitate driver comprehension might be beneficial.

[0038] Finally, the phase correction vector can be defined as the sum of three vectors: the phase reversal factor, the balance adjustment factor D, and the correction position adjustment factor, i.e., the correction vector C, where each element C... i According to formula C i = D i + P i + S F i get.

[0039] Once the correction vector C is determined for each row and each desired S value, it is converted into a digital filter for motor vehicles. For this purpose, in the case of finite impulse response filters, a least mean square (LMS) based design or an integral and windowed design can be used.

[0040] In the case of infinite impulse response (IR) filters, the design can be based on the least squares method. However, the divergence characteristic of IR filters must then be handled very carefully. Therefore, although finite impulse response (FIR) filters require more computational resources, the applicant prefers FIR filters.

[0041] This results in one or more digital filters for each row and for each desired S value. These filters are applied only to the first or second audio channel.

[0042] Alternatively, the applicant found it advantageous to generate a digital filter for each audio channel, which allows for the allocation of corrections, thereby limiting the risk of filter saturation. In this case, a correction factor a is used, according to formula C. i (First audio channel) = a C i And Ci (Second audio channel) = (1 a) C i Define a correction filter for the first channel and a correction filter for the second channel, where 'a' is selected from the range of 0 to 1. A value of 0.5 allows for optimization of saturation risk because in the resulting digital filter, half of the correction is handled by each audio channel.

[0043] Upon completion of this research, the applicant will be able to achieve the following: Figure 2 The multi-channel audio data processing device 100 for motor vehicles shown is illustrated.

[0044] The device 100 includes a data storage 110 and a computer 120.

[0045] The data storage 100 receives two types of data: one is multi-channel audio data 104 to be processed, and the other is data 108 that defines each row of phase-corrected digital filters.

[0046] The data storage device 4 can be any type of data storage device capable of receiving digital data: hard disk, hard disk with flash memory, any form of flash memory, random access memory, disk, local distributed storage or cloud storage, etc.

[0047] Computer (PC) 120 may include one or more processors (P) 122. In some embodiments, computer 120 includes a programmable processor that can provide a computer program (CPP) 124 (or "computer program product") to implement phase correction processing. Computer program CPP 124 stores computer program (CP) 126, which includes computer-readable instructions (CRI) 128. Computer program CPP 124 may be stored on computer-readable medium (CRM) 130, which may be a non-transitory computer-readable medium, such as magnetic media (hard disk, SSD, magnetic tape, etc.), optical media (CD, DVD, Blu-ray disc, etc.), memory (RAM, flash memory, etc.), distributed storage, or cloud storage, etc. Computer-readable medium (CRM) 130 may also be stored in data storage 100.

[0048] The processor P122 can be any processor suitable for performing the calculations described below. Such a processor can be implemented in any known manner, such as a microprocessor in a personal computer, laptop, tablet, or smartphone; a dedicated signal processing processor (“DSP”); a special-purpose chip of the FPGA or SoC type; computing resources in a grid or cloud; a graphics processing unit (“GPU”) cluster; a microcontroller; or any other type capable of providing the computing power required to perform the calculations described below. One or more of these elements can also be implemented in the form of special-purpose electronic circuitry such as an ASIC. Combinations of processors and electronic circuitry are also contemplated. Special-purpose machine learning processors are also considered.

[0049] Because the correction filter is limited to frequencies below a threshold between 1.5 kHz and 5 kHz, the computer 120 filters the input multichannel audio signal, separating it into high-frequency signals with frequencies above the threshold and low-frequency signals with frequencies below the threshold. This means that each channel in each row of the multichannel audio signal is divided into low-frequency and high-frequency components.

[0050] Then, the computer 120 applies a corresponding digital filter to each channel of each row of the low-frequency signal, and then recombines the resulting signal with the high-frequency signal of each channel to produce a corrected audio signal.

[0051] Advantageously, the applicant found that two processing operations could be applied to further improve the correction and / or the computational power consumed: Computer 120 can be configured to undersample low-frequency signals, apply digital filters to them, and then oversample them before recombining them with high-frequency signals.

[0052] This makes it possible to use lower-cost digital filters.

[0053] Computer 120 is able to determine the delay factor of the high-frequency signal based on the vector C determined for the correlation array. This is because the vector C will have the effect of delaying certain frequencies in the low-frequency signal, and it is desirable to align the high-frequency signal as closely as possible with these delays and the delays caused by the digital filter design.

Claims

1. A multi-channel audio data processing device for motor vehicles, comprising: - Data storage (110), which is arranged to receive: Multichannel audio data, which includes row audio data associated with a corresponding speaker row, the multichannel audio data being associated with at least one row, and the row audio data associated with a given speaker row including a first audio channel and a second audio channel; For the first and second row digital filters of each row; The first and second row digital filters for a given row are determined based on correction data including row correction data for each row of speakers associated with the multichannel audio data. The row correction data associated with the speakers of the given row originates from phase measurements of two signals, one originating from the reproduction of a first audio channel in the speakers of the given row, and the other from the reproduction of a second audio channel in the speakers of the given row, each using one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz. These phase measurements are performed, on the one hand, at a first position associated with the speakers of the given row and used to determine a first set of row phase differences, where each element is associated with the phase difference of a measurement obtained at the first position for the first and second audio channels at the same corresponding frequency; and on the other hand, at a second position associated with the speakers of the given row and used to determine a second set of row phase differences, where each element is associated with the phase difference of a measurement obtained at the second position for the first and second audio channels at the same corresponding frequency. Each element of the correction data is associated with one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and each element C i Associated with one of several frequencies below a threshold between 1.5 kHz and 5 kHz, and according to formula C i = D i + P i + S F i We obtain, of which: i is the index of a given frequency among a range of frequencies below a threshold between 1.5 kHz and 5 kHz. D i According to the formula Calculate, where R1 i It is the value of the element with index i in the first group of row phase differences, R2 i It is the value of the element with index i in the second set of row phase differences. P i It is a phase correction value equal to 0 or π, and depends on cos R1 i + cos R2 i The sum of values, S is the allocation factor between -1 and 1, and F i According to formula (R1) i R2 i + π) mod 2π π calculation The first row of digital filters and the second row of digital filters are obtained by transforming the row correction data into finite impulse response filters or infinite impulse response filters. The apparatus (100) further includes a computer (120) configured to apply each first row digital filter of a given row to all or part of a first audio channel of row audio data associated with the given row, and to apply each second row digital filter of a given row to all or part of a second audio channel of row audio data associated with the given row.

2. The apparatus of claim 1, wherein the first row of digital filters and the second row of digital filters are determined based on row correction data, P i The value is calculated according to the following formula: Where h is a lag factor selected from the range [0; 0.2].

3. The apparatus of any of the preceding claims, wherein the computer (120) is further arranged to: On one hand, it receives input audio data as input, in order to separate it into low-frequency audio data with frequencies below a threshold between 1.5 kHz and 5 kHz, and high-frequency audio data with frequencies above a threshold between 1.5 kHz and 5 kHz. On the other hand, the output audio data is returned, which includes: A first audio channel, which combines a signal from the first audio channel derived from high-frequency audio data and a signal from the first audio channel derived from low-frequency audio data corrected by a first row of digital filters, and a second audio channel, which combines a signal from the second audio channel derived from high-frequency audio data and a signal from the second audio channel derived from low-frequency audio data corrected by a second row of digital filters.

4. The apparatus of claim 3, wherein the computer (120) is arranged to undersample the low-frequency audio data to, on the one hand, apply a first row of digital filters to a first audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal originating from the first audio channel of the low-frequency audio data corrected by the first row of digital filters, and on the other hand, apply a second row of digital filters to a second audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal originating from the second audio channel of the low-frequency audio data corrected by the second row of digital filters.

5. The apparatus of claim 3 or 4, wherein the computer (120) is arranged to generate a signal from a first audio channel derived from high-frequency row audio data and a signal from a second audio channel derived from high-frequency row audio data, the signals having a delay selected according to row correction data.

6. The apparatus of any preceding claim, wherein the first row of digital filters and the second row of digital filters are obtained by transforming them into finite impulse response filters according to the allocation of row correction data, such that C i (First audio channel) = a C i And C i (Second audio channel) = (1 a) C i , where a is a real number between 0 and 1.

7. A method for processing multi-channel audio signals for motor vehicles, comprising the following operations: a) Receive multi-channel audio data, which includes row audio data associated with a corresponding speaker row, the multi-channel audio data being associated with at least one row, and the row audio data associated with a given speaker row including a first audio channel and a second audio channel. (b) For each row, a first row digital filter and a second row digital filter are obtained. The first row digital filter and the second row digital filter for a given row are determined based on correction data including row correction data of each row of speakers associated with the multichannel audio data. The row correction data associated with the speakers of a given row originates from phase measurements of two signals, one originating from the reproduction of a first audio channel in the speakers of the given row, and the other from the reproduction of a second audio channel in the speakers of the given row, each using one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz. These phase measurements are performed, on the one hand, at a first position associated with the speakers of the given row and used to determine a first set of row phase differences, wherein each element is associated with the phase difference of the measurement obtained at the first position for the first and second audio channels at the same corresponding frequency, and on the other hand, at a second position associated with the speakers of the given row and used to determine a second set of row phase differences, wherein each element is associated with the phase difference of the measurement obtained at the second position for the first and second audio channels at the same corresponding frequency. Each element of the correction data is associated with one of a plurality of frequencies below a threshold between 1.5 kHz and 5 kHz, and each element C i Associated with one of several frequencies below a threshold between 1.5 kHz and 5 kHz, and according to formula C i = D i + P i + S F i We obtain, of which: i is the index of a given frequency among a range of frequencies below a threshold between 1.5 kHz and 5 kHz. D i According to the formula Calculate, where R1 i It is the value of the element with index i in the first group of row phase differences, R2 i It is the value of the element with index i in the second set of row phase differences. P i It is a phase correction value equal to 0 or π, and depends on cos R1 i + cos R2 i The sum of; S is the allocation factor between -1 and 1, and F i According to formula F i = (R1 i R2 i + π) mod 2π π calculation The first row of digital filters and the second row of digital filters are obtained by transforming the row correction data into finite impulse response filters or infinite impulse response filters. c) Applying each first row digital filter of a given row to all or part of the first audio channel of the row audio data associated with the given row, and applying each second row digital filter of a given row to all or part of the second audio channel of the row audio data associated with the given row.

8. The method of claim 7, wherein the first row digital filter and the second row digital filter in operation b) are determined based on row correction data, P i The value is calculated according to the following formula: Where h is a lag factor selected from the range [0; 0.2].

9. The method of claim 7 or 8, wherein operation a) comprises separating the sub-audio data into low-frequency sub-audio data with frequencies below the threshold between 1.5 kHz and 5 kHz and high-frequency audio data with frequencies above the threshold between 1.5 kHz and 5 kHz, and operation c) comprises returning the sub-audio data, the sub-audio data comprising: A first audio channel, which combines a signal from the first audio channel derived from high-frequency audio data and a signal from the first audio channel derived from low-frequency audio data corrected by a first row of digital filters, and a second audio channel, which combines a signal from the second audio channel derived from high-frequency audio data and a signal from the second audio channel derived from low-frequency audio data corrected by a second row of digital filters.

10. The method of claim 9, wherein operation c) comprises: c1) Undersample the low-frequency audio data to apply a first row of digital filters to the first audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal from the first audio channel of the low-frequency audio data corrected by the first row of digital filters; and apply a second row of digital filters to the second audio channel of the undersampled low-frequency audio data and oversample the resulting signal to generate a signal from the second audio channel of the low-frequency audio data corrected by the second row of digital filters; and c2) Combine the signal from the first audio channel of the low-frequency audio data corrected by the first row of digital filters with the first audio channel of the high-frequency audio signal; and combine the signal from the second audio channel of the low-frequency audio data corrected by the second row of digital filters with the second audio channel of the high-frequency audio signal to generate returned audio data.

11. The method of claim 9 or 10, wherein operation c) includes generating a signal from a first audio channel derived from high-frequency row audio data and a signal from a second audio channel derived from high-frequency row audio data, the signals having a delay selected according to row correction data.

12. The method of any one of claims 7 to 11, wherein the first row of digital filters and the second row of digital filters are obtained by transforming them into finite impulse response filters according to the allocation of row correction data, such that C i (First audio channel) = a C i And C i (Second audio channel) = (1 a) C i , where a is a real number between 0 and 1.

13. A computer program implemented by a computer, comprising instructions for performing the method of any one of claims 7 to 12.

14. A data storage medium having the computer program of claim 13 recorded thereon.

Images