ITTO20130968A1 - MICROPHONE WITH MATRIX AND METHOD OF ACQUISITION OF ACOUSTIC SIGNALS - Google Patents

Description

DESCRIPTION

"MATRIX MICROPHONE AND ACOUSTIC SIGNAL ACQUISITION METHOD"

The present invention relates to a matrix microphone and to a method for acquiring acoustic signals.

As is known, the spread of microelectromechanical microphones (or MEMS microphones, MicroElectroMechanical Systems) is becoming increasingly widespread in many sectors due to the numerous advantages they offer, especially the level of miniaturization and the excellent compromise between cost and performance.

Microelectromechanical microphones normally comprise a membrane defining an electrode of a capacitor. When the membrane deforms in response to an acoustic signal, the capacitance of the capacitor varies and the variations can be read and converted into an output signal indicative of the amplitude of the acoustic signal. Microelectromechanical microphones are therefore equipped with reading interfaces, which generally provide electrical output signals in PDM (Pulse Density Modulation) format, i.e. a series of high frequency pulses (typically from 1 MHz to 3 MHz), whose temporal density is indicative of the amplitude of the input signal. The electrical signals are then supplied to a processing unit for conversion to PCM (Pulse Code Modulation) format, which is the commonly used coding for audio signals.

High-frequency PDM format signals allow you to obtain PCM format signals with high accuracy.

Microelectromechanical microphones have proved to be particularly interesting, among other things, for applications in which arrays of microphones are used. In these cases, the signals provided by the individual microphones are collected by a single processing unit which, in addition to converting to PCM format, can combine and further process the received signals. In particular, it is possible to implement so-called beamforming algorithms, that is spatial filtering techniques that allow to selectively amplify the acoustic signals coming from a given direction, attenuating the other contributions. Beamforming algorithms are frequently used when directionality is important to improve reception quality, for example in the case of music recordings, for speech recognition, teleconferencing applications, web conferencing, and so on.

A very common beamforming algorithm is the so-called delay and sum beamforming.

In the one-dimensional case, the beamforming algorithm for delay and sum is applied to PCM signals collected by a plurality of microphones aligned along one direction and synchronized. In practice, the PCM signals are added after being delayed in order to compensate for the different acoustic path from the acoustic source to the respective microphones from which the PCM signals were produced. The acoustic signals reach the microphones in subsequent instants, with a delay that essentially depends on the angle between the direction of propagation of the acoustic signals and the direction along which the microphones are aligned. The compensation delay applied to each PCM signal allows to add coherently only the components due to signals coming from a specific direction, thus obtaining a selective amplification effect. The amount of delay applied determines the listening direction, that is the direction along which the coherent sum of the acoustic signals is obtained.

The beamforming algorithm for delay and sum usually requires interpolations between successive samples of the PCM signals, since, normally, the delay to be applied to the signals is not a multiple of the sampling frequency (which, for PCM signals, can be between 4 KHz and 48 KHz).

On the other hand, interpolation operations can degrade the quality of the audio signal produced and, in any case, require a considerable computational load. The resources needed to ensure sufficient processing capacity, in addition to representing a cost, also affect consumption and limit the autonomy of portable devices. The need is therefore felt to simplify the processing required to perform beamforming algorithms, in order to allow more effective exploitation of microelectromechanical microphone arrays.

The object of the present invention is to provide a matrix microphone and a method for acquiring acoustic signals which allow to overcome or at least mitigate the limitations described.

According to the present invention, a matrix microphone and an acoustic signal acquisition method are provided as defined in claims 1 and 12 respectively.

For a better understanding of the invention, some embodiments will now be described, purely by way of non-limiting example and with reference to the attached drawings, in which:

Figure 1 is a simplified block diagram of a matrix microphone according to an embodiment of the present invention;

Figure 2 is a more detailed block diagram of a part of the microphone of Figure 1;

Figure 3 is a graph illustrating quantities relating to the microphone of Figure 1;

Figure 4 is a schematic representation of a part of the microphone of Figure 1 in use;

Figure 5 is a simplified block diagram showing a variant of the microphone of Figure 1;

Figure 6 is a simplified block diagram of a matrix microphone according to a different embodiment of the present invention;

- figure 7 is a more detailed block diagram of a part of the microphone of figure 6; And

- Figure 8 is a simplified block diagram of a matrix microphone according to a further different embodiment of the present invention.

With reference to Figure 1, a matrix microphone is indicated as a whole with the number 1 and comprises a plurality of digital microphones 2 arranged in a matrix and a processing unit 3.

In one embodiment, the digital microphones 2 are aligned along an X direction so as to form a one-dimensional matrix and are spaced, for example, but not necessarily, in a uniform manner. In particular, adjacent digital microphones 2 are separated by a uniform distance D.

As shown in Figure 2, each digital microphone 2 comprises a microelectromechanical electroacoustic transducer 5, for example of the membrane capacitive type, and a converter 6. An electrical sensing property of the electroacoustic transducer 5 (for example the capacitance of a capacitor, of which the membrane, not shown, forms an electrode) is modified in response to interaction with an incident acoustic wave. Converter 6 reads changes in the electrical detection property and produces a BSJ native bit stream (for the generic J-th digital microphone 2) oversampled in Pulse Density Modulation (PDM) format. In one embodiment, in particular, the converter 6 is a sigma-delta converter and the supplied BSJ native bit stream has an FPDM sampling frequency between 1 MHz and 3 MHz and a corresponding sampling period τPDM (Figure 3).

With reference again to Figure 1, the processing unit 3 comprises delay elements 7 and an integrated processor 8, which can be, for example, a general-purpose microprocessor or a digital signal processor DSP (Digital Signal Processor). The processing unit 3 is configured to apply a beamforming algorithm on the signals received by the digital microphones 2, so as to allow the selection of a preferential reception direction.

The delay elements 7 are programmable and, in the embodiment of Figure 1, are defined by respective bistable elements controlled by the integrated processor 8.

Each delay element 7 is coupled to a respective digital microphone 2 to receive the corresponding native bit stream BS1, BS2,…, BSK. The delay elements 7 apply respective programmable delays τ1, τ2,…, τK to the received native bit streams BS1, BS2,…, BSK and produce corresponding delayed bit streams BS1 <*>

, BS2 <*>,…, BSK <*>. The relations between the delays τ1, τ2,…, τK are determined by the integrated processor 8 in the form of the number of PDM samples as explained in detail below. In practice, the delays τ1, τ2,…, τK are multiples of the sampling period τPDM.

The integrated processor 8 comprises a computing module 10, which determines and sets the delays τ1, τ2,…, τK for the delay elements 7, a PDM-PCM conversion stage 11 for each digital microphone 2 and an adder module 12.

The PDM-PCM conversion stages 11 receive respective delayed bit streams BS1 <*>

, BS2 <*>,…, BSK <*> in PDM format and provide respective audio signals S1, S2,…, SK by converting to PCM (Pulse Code Modulation) format, which is normally used for encoding digital audio signals. The audio signals S1, S2,…, SK can be represented for example by 16-bit integers with sampling frequency FPCM between 4 KHz and 48 KHz and corresponding sampling period τPCM (figure 3).

Each PDM-PCM 11 conversion stage includes a low-pass filter 13 and a decimator 15. The low-pass filter 13 reconstructs an oversampled waveform starting from the delayed bit stream BS1 <*>

, BS2 <*>, ..., BSK <*> received and decimator 15 eliminates redundant samples with a decimation factor M = FPDM / FPCM given by the ratio between the FPDM sampling frequency of the signals in PDM format and the FPCM sampling frequency of the signals in PCM format.

The audio signals S1, S2,…, SK are supplied to the summing module 12, where they are summed to form a synthetic audio signal SOUT.

In the synthetic audio signal SOUT, the contributions received by all the digital microphones 2 and due to the acoustic signals coming from a selected preferential reception direction are added together in a coherent way and then amplified, contrary to the contributions coming from different directions.

The selection of the preferential reception direction and the amplification are obtained thanks to the application of the delays τ1, τ2, ..., τKe respectively to the sum of the audio signals S1, S2, ..., SK.

Figure 4 shows how the same wave front WF of an acoustic signal, which travels with a speed C along a propagation direction Pθ inclined by an angle θ with respect to the normal to the direction of alignment X of the digital microphones 2, reaches the microphones digital 2 themselves in successive instants t1, t2, ..., tK, which for the generic J-th element of the microphone matrix are given by:

tJ = t1 + (J-1) τ [1]

with τ = (D / C) sin θ e 1 ≤ J ≤ K (C being the speed of sound in the air, equal to about 340 m / s).

In order to select a preferential reception direction, corresponding for example to the angle θ, the calculation module 10 determines the delays τ1, τ2,…, τK for each microphone, for example using the relation [1], so that in the module adder 12 samples corresponding to the same wave front traveling in the direction of propagation Pθ identified by the angle θ are added.

In particular, the generic delay τJ for the digital microphone 2 J-th is an integer given by:

τJ = [(K-J) τ / τPDM] [2]

where the operator [] indicates the integer part function and τPDM is the PDM sampling period, equal to 1 / FPDM (Figure 3). In practice, the delays τ1, τ2, ..., τK indicate the number of PDM samples from which it is necessary to translate the native bit streams BS1, BS2, ..., BSK to ensure that the audio signal samples S1, S2, ..., SK added into the adder module correspond to the reception of the same wave front traveling along the propagation direction Pθdefined by the angle θ. In other words, with the delays τ1, τ2,…, τK thus set, the direction of propagation Pθ is selected as the preferential reception direction. Note that, as defined here, the lags τ1, τ2,…, τK are integers. The corresponding effective durations are given by τ1τPDM, τ2τPDM,…, τKτPDM.

In principle, the preferential reception direction is determined exactly only if the delays τ1, τ2,…, τK are integer multiples of the sampling period τPDM. However, the FPDM sampling rate from which the samples of the native bit streams BS1, BS2, ..., BSK are produced is high enough for the error made using the approximation to the nearest sample (e.g. according to relation [2]) is completely negligible.

In fact, therefore, the introduction of the delays τ1, τ2, ..., τK before the conversion from the PDM format to the PCM format allows you to apply beamforming algorithms with satisfactory precision and angular resolution, without the need for interpolations.

In an embodiment, to which Figure 5 refers, the delay elements 7 are integrated in the digital microphones 2 and receive from the processing unit 3 control signals indicative of the delays τ1, τ2, ..., τK to be applied to the native bits BS1, BS2,…, BSK.

In some embodiments, it may be convenient to carry out the PDM-PCM conversion in multiple stages, for example for reasons of process optimization or acoustics. In these cases, there are multiple low-pass filters and decimators, each of which performs a partial decimation. The product of the partial decimation factors is equal to the overall decimation factor, which is determined by the ratio of the PDM sample rate to the PCM sample rate. Furthermore, in order to perform beamforming algorithms, it is possible to apply delays directly to the bitstream in PDM format upstream and, possibly, in intermediate stages of the PDM-PCM conversion without losing precision and angular resolution.

According to the embodiment illustrated in Figure 6, for example, a matrix microphone 100 comprises a plurality of matrix-arranged microelectromechanical digital microphones 102 and a processing unit 103, which is configured to perform PDM-PCM conversion in two stages and applying beamforming algorithms to select a preferential reception direction.

The microphones 102 are made substantially as already described with reference to Figure 2 and provide respective native bit streams BS1, BS2,…, BSK in PDM format.

The processing unit 103 comprises a calculation module 110, a respective processing line 111 for each digital microphone 102 in the matrix and an adder module 112.

The computation module 110 determines delays τ1, τ2,…, τK to be applied to the native bit streams BS1, BS2,…, BSK to select a preferential reception direction by means of a beamforming algorithm, substantially as previously described.

The processing lines 111 receive and convert respective native bit streams BS1, BS2,…, BSK into audio signals S1, S2,…, SK by means of a multi-stage conversion procedure.

The audio signals S1, S2, ..., SK, which are aligned according to the delays τ1, τ2, ..., τK determined by the computation module 110, are supplied to the summing module 112, which supplies a synthetic audio signal SOUT, in which the contributions due the acoustic signals coming from the preferential reception direction are amplified, unlike the contributions coming from other directions.

The processing lines 111 have the same structure and one of them, in particular the processing line J-th, is illustrated in Figure 8 by way of example. Each processing line 111 comprises a first delay element 107a, a first PDM-PCM conversion stage 111a, a second delay element 107b, a second PDM-PCM conversion stage 111b and a scrambling module 114.

The first PDM-PCM conversion stage 111a comprises, in cascade, a first low-pass filter 113a and a first decimator 115a, having a first decimation factor M1.

The second PDM-PCM conversion stage 111b is arranged in series with the first conversion stage and comprises, in cascade, a second low-pass filter 113b and a second decimator 115b, having a second decimation factor M2.

The product of the first decimation factor M1 and the second decimation factor M2 gives an overall decimation factor M of the processing lines 111:

M = M1 * M2

The overall decimation factor M is also equal to the ratio between the sampling frequency FPDM of the native bit streams BS1, BS2,…, BSK and the sampling frequency FPCM of the audio signals S1, S2,…, SK.

The first delay element 107a and the second delay element 107b are arranged respectively upstream of the first low-pass filter 113a and of the second low-pass filter 113b and are programmable to apply a first delay τJ1 and a second delay τJ2 respectively to the signals received in entrance.

The first delay τJ1 and the second delay τJ2 are determined by the decomposition module 114 so that the conditions are met:

τJ = τJ1 + M2 * τJ2

0 ≤ τJ1≤ M1-1

For this purpose, the decomposition module 114 can, for example, perform the following operations:

τJ1 = τJmod M1

τJ2 = (τJ- τJ1) / M1

where the mod operator provides the remainder of the division between integers.

However, it is understood that, in a different embodiment, the decomposition between first delays τ11, τ21, ..., τK1 and second delays τ12, τ22, ..., τK2 can be carried out directly by the calculation module 110 and the values provided respectively to the first delay elements 107a and the second delay elements 107b (for example, a decomposition module can be incorporated into the computation module 110).

Even with the solution described above, the angular resolution essentially depends on the FPDM sampling frequency and it is not necessary to resort to interpolation techniques.

It is particularly advantageous the case in which the first decimation factor M1 is equal to 8 and the first delays τ11, τ21, ..., τK1 are between 0 and 7. The data produced in this way, in fact, are always aligned by byte and it is therefore easy to implement the subsequent phases of delay, filtering and decimation by program.

Furthermore, as already mentioned previously, also in this case the first delay elements 107a can be integrated in the microelectromechanical microphones 112, which can therefore produce data aligned by byte.

In the embodiment illustrated in Figure 8, a matrix microphone 200 comprises a plurality of digital microphones 202a, 202b forming a two-dimensional matrix. In particular, first digital microphones 202a are aligned along a first direction X and provide first native bit streams BSX1, BSX2,…, BSXK with a PDM sampling rate. Second digital microphones 202b are aligned along a second Y direction, perpendicular to the X direction, and provide second native bit streams BSY1, BSV2, ..., BSVK with the same PDM sampling rate as the first native bit streams BSX1, BSX2, ..., BSXK .

The matrix microphone 200 also comprises a processing unit 203, having a computing module 210 and, for each digital microphone 202a, 202b, a respective processing line 211.

The computation module 210 determines delays τX1, τX2, ..., τXK in order to align samples of the first native bit streams BSX1, BSX2, ..., BSXK produced by the same wave front traveling in a direction of propagation and incident on the first digital microphones 202a; and delays τY1, τY2, ..., τYK in order to align samples of the second native bit streams BSY1, BSY2, ..., BSYK produced by the same wave front incident on the second digital microphones 202b. The wave front propagation direction is thus selected as the preferred reception direction for both the first digital microphones 202a and for the second digital microphones 202b.

Processing lines 211 apply the delays τX1, τX2, ..., τXK to the first native bit streams BSX1, BSX2, ..., BSXK and the delays τY1, τY2, ..., τYK to the second native bit streams BSY1, BSY2, ..., BSYKed perform the PDM-PCM conversion by filtering and decimation at one or more stages substantially as already described, thus providing first audio signals SX1, SX2,…, SXK and second audio signals SY1, SY2,…, SYK. The first audio signals SX1, SX2, ..., SXK and the second audio signals SY1, SY2, ..., SYK are all summed by the summing node 212, which provides a synthetic SOUT audio signal in which the contributions provided by both the first digital microphones 202a and by the second digital microphones 202b and due to acoustic signals coming from the preferential reception direction are amplified, contrary to the other contributions.

The described device and method make it possible to implement beamforming algorithms in a simple way and with extremely limited use of resources. At the same time, the signal detection is accurate and the angular resolution is high.

Thanks to the introduction of delay elements upstream of the PDM-PCM conversion, the execution of beamforming algorithms can be reduced to temporal translations and sums between integers, which require very little consumption in terms of both time and energy. In particular, it is possible to take advantage of the high PDM sampling rate to produce exact and correctly aligned samples to set the desired preferential reception direction without the need for interpolations.

Finally, it is clear that modifications and variations can be made to the described device and method, without departing from the scope of the present invention, as defined in the attached claims.

In particular, the delay elements, the low-pass filters and the decimators used for the time translations of the signals and for the PDM-PCM conversion can be realized by means of dedicated dedicated circuits or by program using, for example, a digital signal processor or a microprocessor.

In the case of multi-stage PDM-PCM conversion, delay elements may be present in each stage or only in some stages. Alternatively, it is also possible to use a single delay element for each digital microphone, upstream of the first PDM-PCM conversion stage.

Even in the case of multi-stage PDM-PCM conversion, the first delay elements can be integrated into the digital microphones.

Claims (17)

CLAIMS 1. Matrix microphone comprising: a plurality of digital microphones (2; 102; 202) arranged along at least one direction of alignment to form a matrix and configured to provide respective native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1 , BSV2,…, BSVK) in Pulse Density Modulation format with a first frequency (FPDM), in response to acoustic signals (WF); a processing unit (2; 103; 203), configured to convert native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) from Pulse Density format Modulation in audio signals (S1, S2, ..., SK; SX1, SX2, ..., SXK, SY1, SV2, ..., SVK) in Pulse Code Modulation format with a second frequency (FPCM), lower than the first frequency (FPDM), and to determine respective delays (τ1, τ2, ..., τK; τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) to be applied to native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) in order to align the audio signals (S1, S2, ..., SK; SX1, SX2, ..., SXK, SY1, SV2, ..., SVK) according to an algorithm of beamforming; a plurality of delay elements (7; 107a, 107b) configured to apply the respective delays (τ1, τ2, ..., τK; τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) to the native bit streams ( BS1, BS2,…, BSK; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK) before converting to Pulse Code Modulation format. Microphone according to claim 1, wherein the processing unit (3; 103; 203) comprises a plurality of PDM-PCM conversion stages (11; 111a, 111b; 211) and the delay elements (7; 107a , 107b) are arranged between respective digital microphones (2; 102) and respective PDM-PCM conversion stages (11; 111a, 11b; 211). Microphone according to claim 2, wherein the PDM-PCM conversion stages (11; 111a, 11b; 211) comprise respective decimators (15; 115a, 115b) and the delay elements (7; 107a, 107b) are arranged between respective digital microphones (2; 102) and respective decimators (15; 115a, 115b). Microphone according to claim 3, wherein the PDM-PCM conversion stages comprise first PDM-PCM conversion stages (111a) and second PDM-PCM conversion stages (111b) in cascade to respective first PDM-PCM conversion stages (111a); and in which the delay elements comprise first delay elements (107a), between respective digital microphones (102) and respective first PDM-PCM conversion stages (111a), and second delay elements (107b), between respective first stages of PDM-PCM conversion (111a) and respective second PDM-PCM conversion stages (111b). 5. Microphone according to claim 4, comprising a decomposition module (114) configured to decompose the delays (τX1, τX2,…, τXK, τY1, τY2,…, τYK) into first delays (τX1, τX2,…, τXK) and second delays (τY1, τY2, ..., τYK) and to provide the first delays (τX1, τX2, ..., τXK) to respective first delay elements (107a) and the second delays (τY1, τY2, ..., τYK) to respective second delay elements (107b). Microphone according to claim 4 or 5, wherein the decimators comprise first decimators (115a), having a first decimation factor (M1), in the first PDM-PCM conversion stages (111a) and second decimators (115b), having a second decimation factor (M2), in the second PDM-PCM conversion stages (111b). Microphone according to claim 6, wherein the first delays (τX1, τX2,…, τXK) are smaller than the first decimation factor (M1). Microphone according to claim 6 or 7, wherein the delay elements are configured to receive respective native bit streams (BS1, BS2,…, BSK; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK) and respective delays (τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) and to provide respective delayed bit streams (BS1 <*> , BS2 <*>,…, BSK <*>), delayed by a number of samples corresponding to the respective delays (τX1, τX2,…, τXK, τY1, τY2,…, τYK). 9. Microphone according to any one of the preceding claims, wherein the delays (τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) correspond to respective sample numbers of the native bit streams (BS1, BS2, ..., BSK ; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK). Microphone according to any one of the preceding claims, wherein at least some of the delay elements (7) are integrated in the digital microphones (2). Microphone according to any one of the preceding claims, wherein the processing unit (3; 103; 203) comprises an adder module (12; 112; 212) configured to sum the audio signals (S1, S2,…, SK; SX1, SX2,…, SXK, SY1, SV2,…, SVK) aligned. 12. A method of acquiring acoustic signals comprising: convert acoustic signals (WF) into native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) in Pulse Density Modulation format with a first frequency (FPDM), by a plurality of digital microphones (2; 102; 202) arranged along at least one direction of alignment to form a matrix; convert native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) from Pulse Density Modulation format into audio signals (S1, S2, ..., SK; SX1, SX2,…, SXK, SY1, SV2,…, SVK) in Pulse Code Modulation format with a second frequency (FPCM), lower than the first frequency (FPDM); determine respective delays (τ1, τ2, ..., τK; τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) to be applied to the native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2,…, BSVK) in order to align the audio signals (S1, S2,…, SK; SX1, SX2,…, SXK, SY1, SV2,…, SVK) according to a beamforming algorithm; apply the respective delays (τ1, τ2,…, τK; τX1, τX2,…, τXK, τY1, τY2,…, τYK) to the native bit streams (BS1, BS2,…, BSK; BSX1, BSX2,…, BSXK , BSY1, BSV2,…, BSVK) before converting native bit streams (BS1, BS2,…, BSK; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK) from Pulse Density Modulation format into audio signals (S1, S2,…, SK; SX1, SX2,…, SXK, SY1, SV2,…, SVK) in Pulse Code Modulation format. The method according to claim 12, wherein converting the native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) comprises decimating the native bit streams (BS1 , BS2,…, BSK; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK) and the respective delays (τ1, τ2,…, τK; τX1, τX2,…, τXK, τY1, τY2,…, τYK ) are applied to native bit streams (BS1, BS2,…, BSK; BSX1, BSX2,…, BSXK, BSY1, BSV2,…, BSVK) before decimating. The method according to claim 13, wherein converting the native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) comprises using a first PDM- conversion stage PCM (111a) and a second PDM-PCM conversion stage (111b); and in which to apply the respective delays (τ1, τ2, ..., τK; τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) to the native bit streams (BS1, BS2, ..., BSK; BSX1, BSX2, ..., BSXK, BSY1, BSV2, ..., BSVK) includes breaking down delays (τX1, τX2, ..., τXK, τY1, τY2, ..., τYK) into first delays (τX1, τX2, ..., τXK) and second delays (τY1 , τY2, ..., τYK) and use the first delays (τX1, τX2, ..., τXK) in the first PDM-PCM conversion stage (111a) and the second delays (τY1, τY2, ..., τYK) in the second PDM conversion stage -PCM (111b). Method according to claim 14, wherein decimating comprises performing a first decimation with a first decimation factor (M1) in the first PDM-PCM conversion stage (111a) and a second decimation with a second decimation factor (M2) in the second stage of PDM-PCM conversion (111b). Method according to claim 15, wherein the first delays (τX1, τX2,…, τXK) are smaller than the first decimation factor (M1). Method according to any one of claims 12 to 16, comprising adding the aligned audio signals (S1, S2,…, SK; SX1, SX2,…, SXK, SY1, SV2,…, SVK).