Method for controlling an acoustic field, associated control system and associated computer program

Abstract

The invention relates to a method for controlling an acoustic field, comprising a calibration phase comprising the steps of: - emitting a sound wave by a set of loudspeakers (12, 14) on the basis of a predetermined signal; - measuring, for the or each loudspeaker (12, 14), respectively, a response of the sound wave by a plurality of sensors (16), so as to obtain initial measurements, each initial measurement being associated with one of the loudspeakers (12, 14) and being further associated with one of the sensors (16); - defining a subset comprising at least one loudspeaker, referred to as at least one emitter loudspeaker (12), from among the set of loudspeakers (12, 14), according to a predetermined criterion; and - determining a set of transform filters for reproducing an audio signal (50) by the set of loudspeakers (12, 14).

Description

[0001] DESCRIPTION

[0002] TITLE: Method for controlling an acoustic field, associated control system and computer program

[0003] The present invention relates to a method for controlling an acoustic field.

[0004] The invention further relates to a system for controlling an acoustic field. The invention also relates to an associated computer program.

[0005] The present invention relates to the field of controlling an acoustic field.

[0006] Sound reproduction systems are known per se. These systems are intended, for example, for use by professionals, such as people in mixing studios, or for use by individuals in home cinemas. Such systems are also used, for example, to control an acoustic field for a large audience, such as in movie theaters or concert halls. These systems integrate, for example, loudspeakers capable of reproducing only mid-range or high-frequency frequencies and loudspeakers capable of also reproducing low frequencies and / or loudspeakers dedicated to reproducing low frequencies, called subwoofers or subwoofers.

[0007] Generally, the reproduction of the sound field is done in a listening area inside a room whose walls are the source of many reflections and resonances that amplify, attenuate or cancel certain frequencies of the sound field in this listening area. In addition, the listening area may have obstacles generating other reflections and resonances.

[0008] These changes in the sound field are different depending on the different positions in the space. To achieve a satisfactory listening experience for users positioned at different points in the listening area covered by the sound field, it is necessary to control the sound waves emitted by the respective loudspeakers. In particular, it is necessary to control and modify the emitted sound waves, to obtain a controlled sound field at each point in the listening area. For example, in the absence of control, interference may disrupt the listening experience.

[0009] Such control depends on a very large number of parameters, such as for example the number, position and orientation of the loudspeakers, the shape of the listening room and the materials of the walls and the positions and types of obstacles in the room, in which the sound field is reproduced, etc.

[0010] Also, it is often difficult, if not impossible, to implement optimized control for each point of the listening area, given the great complexity of the different parameters and factors influencing the control. Thus, known solutions for controlling acoustic fields can still be improved.

[0011] One aim of the invention is to obtain a method for controlling an acoustic field and an associated control system, which make it possible to obtain a high quality of hearing experience, in particular at each position of a listening area. Also, another aim of the invention is to obtain a control method which is easy to implement.

[0012] For this purpose, an object of the invention is a method for controlling an acoustic field, comprising a calibration phase comprising steps of:

[0013] - emission of a sound wave by a set of loudspeakers according to a predetermined signal;

[0014] - measurement, respectively for the or each loudspeaker, of a response of the sound wave by a plurality of sensors, to obtain initial measurements, each initial measurement being associated with one of the loudspeaker(s) and being further associated with one of the sensors;

[0015] - definition of a subset comprising at least one loudspeaker, called at least one transmitting loudspeaker, among the set of loudspeakers, according to a predetermined criterion;

[0016] - processing, comprising the application of a predetermined function to each initial measurement associated with the at least one transmitting loudspeaker to obtain, for each initial measurement associated with the at least one transmitting loudspeaker, a processed measurement;

[0017] - determining a set of transformation filters for reproduction of an audio signal by the set of loudspeakers, the set of transformation filters being determined at least as a function of each initial measurement and as a function of a target acoustic field obtained from each processed measurement.

[0018] According to other advantageous aspects of the invention, the control method comprises one or more of the following characteristics, taken in isolation or in all technically possible combinations:

[0019] - during the emission step, a plurality of loudspeakers of the loudspeaker set emits the sound wave according to said predetermined signal;

[0020] - the loudspeakers emit the sound wave consecutively according to said predetermined signal; - the control method further comprises a phase of reproducing the audio signal, comprising steps of:

[0021] + filtering, comprising applying, for each loudspeaker, a transformation filter from the set of transformation filters to the audio signal to obtain a control signal for each loudspeaker, and

[0022] + reproduction of the audio signal by the speaker set according to each corresponding control signal;

[0023] - the predetermined function modifies an amplitude of a part of each initial measurement to obtain each processed measurement;

[0024] - the predetermined function defines at least one time envelope applied to each initial measurement;

[0025] - the time envelope selects only a start having a predetermined duration of each initial measure, other parts of each initial measure being eliminated, and / or

[0026] - the time envelope decreases the amplitude of each initial measurement as a function of time, and / or

[0027] - the time envelope harmonizes a decay of the sound wave response according to each initial measurement, and / or

[0028] - the time envelope increases the amplitude of each initial measurement as a function of time;

[0029] - the predetermined function includes a frequency filter modifying the amplitude of each initial measurement as a function of a frequency of the initial measurement;

[0030] - the frequency filter is a time-frequency filter modifying the amplitude of each initial measurement furthermore as a function of time;

[0031] - the calibration phase further comprises a step of combining the processed measurements to obtain, for each sensor, a target measurement defining said target acoustic field;

[0032] - during the determination step, the set of transformation filters is determined by minimizing a cost function, parameters of the cost function being at least each initial measurement and the target acoustic field;

[0033] - the cost function determines at least the sum of the least squares errors, for each sensor, between the sum of the initial measurements for each loudspeaker, filtered by the corresponding transformation filter, and the target measurement;

[0034] - the cost function includes at least the following expression:

[0035] | |H - g - d | | 2 where: g is a vector comprising the transformation filters of said set of transformation filters for each loudspeaker;

[0036] H is a global convolution matrix comprising each initial measurement of each loudspeaker measured by each sensor; d is a vector comprising each target measurement of the target sound field at a respective position of each sensor;

[0037] - all the sensors are arranged in such a way as to obtain an unambiguous measurement of the acoustic field;

[0038] - each sensor is arranged at a maximum distance from a nearest neighboring sensor strictly less than half of a smallest wavelength of the acoustic field;

[0039] - the subassembly comprises several transmitting loudspeakers, the transmitting loudspeakers being arranged so as to produce a first wavefront in a predetermined listening area having interference below a predetermined threshold between elementary wavefronts respectively emitted by the transmitting loudspeakers, for a predetermined frequency band, at each instant and at each position of the listening area.

[0040] The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement a control method as described above.

[0041] The invention also relates to a system for controlling an acoustic field comprising a set of loudspeakers, a plurality of sensors and at least one controller, the set of loudspeakers being configured to emit a sound wave according to a predetermined signal, the plurality of sensors being configured to measure, respectively for the or each loudspeaker, a response of the sound wave, to obtain initial measurements, each initial measurement being associated with one of the loudspeaker(s) and being further associated with one of the sensors, the control system being further configured to:

[0042] - define a subset comprising at least one loudspeaker, called at least one transmitting loudspeaker, among the set of loudspeakers, according to a predetermined criterion;

[0043] - applying a predetermined function to each initial measurement associated with the at least one transmitting loudspeaker to obtain, for each initial measurement associated with the at least one transmitting loudspeaker, a processed measurement;

[0044] - determining a set of transformation filters for reproduction of an audio signal by the set of loudspeakers, the set of transformation filters being determined at least as a function of each initial measurement and as a function of a target acoustic field obtained from each processed measurement.

[0045] The invention also relates to a method for controlling an acoustic field, comprising a calibration phase comprising steps of:

[0046] - emission of a sound wave by a set of loudspeakers according to a predetermined signal;

[0047] - measurement, respectively for the or each loudspeaker, of a response of the sound wave by a plurality of sensors, to obtain initial measurements, each initial measurement being associated with one of the loudspeaker(s) and being further associated with one of the sensors;

[0048] - definition of a subset of the loudspeakers according to a predetermined criterion, each loudspeaker of the subset being called an emitting loudspeaker;

[0049] - processing, comprising determining a set of calculated responses of the transmitting loudspeakers, each calculated response corresponding to an acoustic response of a respective transmitting loudspeaker of the subset at a position of a respective sensor;

[0050] - combination of the calculated responses to obtain, for each position, a target response, the target responses defining a target acoustic field,

[0051] - determining a set of transformation filters for reproduction of an audio signal by the set of loudspeakers, the set of transformation filters being determined at least as a function of each initial measurement and as a function of said target acoustic field obtained from each calculated response.

[0052] These characteristics and advantages of the invention will appear on reading the description which follows, given solely by way of non-limiting example, and made with reference to the appended drawings, in which:

[0053] [Fig 1] Figure 1 is a schematic view of a system for controlling an acoustic field according to the invention;

[0054] [Fig 2] Figure 2 is a flowchart of a control method implemented by the control system of Figure 1;

[0055] [Fig 3] Figure 3 is a first example of a loudspeaker arrangement of the control system of Figure 1;

[0056] [Fig 4] Figure 4 is a second example of a loudspeaker arrangement of the control system of Figure 1; [Fig 5] Figure 5 comprises a schematic perspective view of a portion of the control system of Figure 1, Figure 5 further comprising two views along planes of the schematic perspective view;

[0057] [Fig 6] [Fig 7] [Fig 8] [Fig 9] [Fig 10] [Fig 11] Figures 6 to 11 are schematic views of exemplary loudspeaker arrangements of the control system of Figure 1;

[0058] [Fig 12] Figure 12 is an amplitude diagram, according to different frequencies and different instants, of an acoustic signal at a given position of an acoustic field, in the absence of filtering of the acoustic signal by transformation filters according to the control method of Figure 2, and

[0059] [Fig 13] Figure 13 is a diagram similar to the diagram of Figure 12, of the acoustic signal filtered by the transformation filters according to the control method of Figure 2.

[0060] With reference to FIG. 1, a system 10 for controlling an acoustic field comprises a set of loudspeakers 12, 14, a plurality of sensors 16 and at least one controller 18.

[0061] The control system 10 is configured to control the acoustic field in a listening area 19. The listening area 19 is notably located in a part of a room 20, such as a movie theater, a concert hall or a living room for example. According to one example, the listening area 19 is identical to the dimensions of the room 20.

[0062] The set of loudspeakers comprises for example at least one emitting loudspeaker 12 and at least one non-emitting loudspeaker 14. According to one example, the emitting and non-emitting loudspeakers 12, 14 have the same technical characteristics, and in particular have different functions for generating the acoustic field during an operation of the control system 10.

[0063] In particular, each loudspeaker 12, 14 of the loudspeaker assembly is connected to the controller 18 by wired or non-wired connections for control by said controller 18. For reasons of visibility, only the connections between the controller 18 and the transmitting loudspeakers 12 are illustrated in FIG. 1.

[0064] In particular, the emitting loudspeakers 12 are arranged and / or controlled so as to be capable of generating a first wavefront having a predetermined spatial and temporal homogeneity throughout the listening area 19. Preferably, the non-emitting loudspeakers 14 do not contribute to this first wavefront having the predetermined spatial and temporal homogeneity throughout the listening area 19. By "predetermined spatial and temporal homogeneity", it is in particular understood that interference between the emitting loudspeakers 12, for a predetermined frequency band, is less than a predetermined threshold, at each instant and at each position of the listening area 19.

[0065] By "first wavefront" is meant in particular the set of positions in space as well as the times at which a wave emitted by the corresponding loudspeaker(s) is established for the first time. At a given time, the first wavefront corresponds to the positions in space reached for the first time by the wave. At a given position in space the first wavefront corresponds to the time at which the wave reaches this point for the first time. In particular, the first wavefront is contained in the initial part, of predetermined duration, of a sound wave emitted by the corresponding emitting loudspeaker(s) 12.

[0066] The set of loudspeakers 12, 14 is configured to correct, for example attenuate, a portion of the sound waves emitted by the emitting loudspeakers 12. In particular, both the emitting loudspeakers 12 and the non-emitting loudspeakers 14 are configured to correct, for example attenuate, a corresponding portion of the sound waves.

[0067] With reference to FIG. 1, the control system 10 preferably comprises several transmitting loudspeakers 12, such as for example four, and several non-transmitting loudspeakers 14, such as for example four. According to an example not shown, the control system 10 comprises a single transmitting loudspeaker 12.

[0068] The sensors 16 are, for example, microphones. Each sensor 16 defines a measurement position, preferably a three-dimensional measurement position. Each sensor 16 is configured to measure sound waves generated by the loudspeakers 12, 14 at the measurement position. The sensors 16 are preferably omnidirectional microphones whose sensitivity does not depend, or very little, on a direction of incidence of the wave on the microphone.

[0069] Controller 18 is for example a calculator.

[0070] The controller 18 comprises for example a reception module 21, a processing module 22 and a transmission module 24.

[0071] In the example of Figure 1, the reception module 21, the processing module 22, the transmission module 24 are each produced at least partially in the form of software, or a software brick, for example stored in a memory 30 of the controller 18 and executable by a processor 32 of the controller 18. The memory 30 of the controller 18 is capable of storing reception software, processing software and transmission software. The processor 32 of the controller 18 is capable of executing the reception software, the processing software and the transmission software.

[0072] The operation of the control system 10 is now described with reference to FIG. 2 which is a flowchart of a method 100 for controlling the acoustic field according to a first embodiment.

[0073] The control method 100 according to the first embodiment notably comprises a preliminary phase 102, a calibration phase 104 and a reproduction phase 106 of an audio signal 50 to be reproduced.

[0074] In the preliminary phase 102, a number and / or an arrangement of the set of loudspeakers 12, 14 and / or the sensors 16 is determined. The preliminary phase 102 is for example implemented by the controller 18 or by a computer separate from the controller 18.

[0075] The preliminary phase 102 comprises, for example, a first preliminary determination step 110 and a second preliminary determination step 112.

[0076] During the first preliminary determination step 1 10, a number and an arrangement of the set of loudspeakers 12, 14 is determined, in particular so as to obtain the first wavefront emitted by the or jointly by each emitting loudspeaker which has the predetermined spatial and temporal homogeneity.

[0077] Several examples of the arrangement of the transmitting loudspeakers during the first preliminary determination step 1 10 are described in the following.

[0078] According to a first example, a single emitting loudspeaker 12 is chosen for the control system 10 during step 110. In this case, the emitting loudspeaker 12 emits, during the operation, a wave having a spherical propagation shape in first approximation. The other loudspeakers are non-emitting loudspeakers 14. All the loudspeakers 12, 14 together correct, for example, reflections and resonances of the sound wave emitted by the emitting loudspeaker 12 in the listening area 19 during an operation of the control system 10.

[0079] According to a second example, with reference to FIG. 3, two transmitting loudspeakers 12 are chosen for the control system 10 during step 110.

[0080] In the example, the transmitting loudspeakers 12 are arranged along a straight line, for example parallel to a wall of the room 20.

[0081] Preferably, the listening zone 19 corresponds in this example to a sector, defined from a midpoint of a segment connecting the two emitting loudspeakers 12, the angular aperture of which is inversely proportional to the distance which separates the two emitting loudspeakers 12 and / or is inversely proportional to the reproduced frequency. This sector thus defines the listening zone 19. In the listening zone 19 according to this example, the first wavefront, designated by the general reference FR in FIG. 3, has a propagation direction illustrated by arrows 40.

[0082] For example, a wavefront emitted by the transmitting loudspeakers 12 comprises destructive interference whose attenuation is less than or equal to -3 dB as long as elementary wavefronts of the transmitting loudspeakers 12 overlap to within 0.25 of a wavelength with the wavefront of its or its nearest neighbors.

[0083] In particular, the first wavefront is homogeneous and / or substantially free of interference. In particular, moreover, this first wavefront is directional and directed towards the listening area 19 while the first wavefront in the direction of the side walls is attenuated by destructive interference, thus producing less unwanted room reflections and resonances.

[0084] According to a third example, with reference to FIG. 4, four transmitting loudspeakers 12 are chosen for the control system 10 during step 110, namely the transmitting loudspeakers 12 arranged inside the room 20. As in the second example, the transmitting loudspeakers 12 are preferably arranged along a straight line, for example parallel to a wall of the room 20.

[0085] For example, the 12 transmitting loudspeakers are arranged in a configuration called a "Double Bass Array". In this case, for example, the reflections of the four 12 transmitting loudspeakers against the side walls of the room contribute to the first wavefront.

[0086] According to the third example, the transmitting loudspeakers 12 are coupled so as to obtain a first wave front substantially forming a plane wave 42, as illustrated in particular in FIG. 4. The plane wave is in particular the most advanced case of a directional wave.

[0087] In the general case of the invention, the transmitting loudspeakers are preferably placed such that when coupled, they produce a directional wave and / or in the extreme case a unidirectional wave or plane wave.

[0088] Of course, in particular the acoustic field develops in the three dimensions of space and the emitting loudspeakers 12 arranged in space to form a first uniform wavefront in the listening area in 3D. Thus, Figures 3 and 4 are interpreted as both top views and side views. This is illustrated in Figure 5 for the example of Figure 3. For example, with reference to Figure 5, the control system comprises four emitting loudspeakers 12 which are arranged in the same plane. According to a first plane A, for example corresponding to a top view, two of the four emitting loudspeakers 12 are visible, the other two being superimposed. According to a second plane B, for example corresponding to a side view, also two of the four emitting loudspeakers 12 are visible. Thus, the arrangement illustrated by Figures 3 and 4 applies to both the arrangement in the first plane A and in the second plane B.

[0089] Examples of arrangement of the transmitting loudspeakers 12 of the control system 10 are illustrated with reference to FIGS. 6 to 11. The transmitting loudspeakers 12 according to each of these example arrangements are arranged and / or controlled so as to be capable of generating the first wavefront having the predetermined spatial and temporal homogeneity throughout the listening area 19, in particular a first uniform wavefront in 3D.

[0090] For example, the transmitting loudspeakers are arranged in the same plane, for example in a vertical plane, visible in figures 6 to 11.

[0091] For example, with reference to FIG. 6, the control system 10 comprises six transmitting loudspeakers 12, three of which are respectively arranged along a straight line, the straight lines being parallel. With reference to FIG. 10, the control system 10 comprises three transmitting loudspeakers 12 arranged along a straight line. With reference to FIG. 11, the control system 10 comprises two transmitting loudspeakers 12 arranged along a straight line. With reference to FIG. 7, the control system 10 comprises three transmitting loudspeakers 12 arranged in a triangle. With reference to FIG. 8, the control system 10 comprises five transmitting loudspeakers 12 arranged in a substantially random manner while respecting a criterion of maximum distance between neighboring loudspeakers.

[0092] In particular, the maximum distance between two adjacent emitting loudspeakers is such that the waves emitted by each of the two adjacent loudspeakers overlap to within 0.25 wavelengths in the angular sector of the listening area. This distance is for example estimated by the following approximate calculation: where: d-adj. max is the maximum distance between two adjacent transmitting loudspeakers 12;

[0093] / minfound is the smallest wavelength to be reproduced; and

[0094] "listening is the angular sector of the listening area 19. For example, for a listening area with an angular sector of 60° and for a minimum wavelength of 3.44 m, corresponding to a maximum frequency of 100 Hz, the maximum distance between 2 loudspeakers is 1.72 m.

[0095] Referring to Figure 9, the control system 10 comprises four transmitting loudspeakers 12 arranged in a rectangle.

[0096] In the second preliminary determination step 112, a number and arrangement of the sensors 16 is determined. For example, the measurement position of each sensor 16 is determined to obtain the arrangement of the sensors 16.

[0097] In particular, the set of sensors 16 is arranged so as to obtain an unambiguous measurement of the acoustic field by the measurement positions of the sensors 16.

[0098] By "unambiguous" it is understood that the signals from the sensors 16 represent, up to a predetermined maximum frequency, called the spatial aliasing frequency, at least the amplitude of the acoustic field at each position of the listening area 19 and not only at the positions of the sensors 16. The continuous acoustic field in space is in particular correctly represented by the finite number of sensors, without loss of information. In particular, for example, the addition of an additional sensor 16 would not add additional information on the acoustic field, such as a value of the amplitude for a given frequency at a given time and for a given point in the listening area 19, compared to the information already included in the signals from the sensors 16 without this additional sensor.

[0099] For example, the sensors 16 are arranged according to an irregular three-dimensional mesh. By "irregular mesh", it is in particular understood that each distance, in the three dimensions, may differ from each other distance between the sensors 16, and / or in particular that at least two neighboring sensors 16 have a distance from each other that is different from a distance between two other neighboring sensors 16.

[0100] Preferably, the distance chosen between two adjacent sensors 16 is less than half of the shortest wavelength to be monitored in the acoustic field, so as to satisfy the conditions of the Nyquist-Shannon theorem. In other words, for example, the arrangement of the sensors 16 is determined by arranging each sensor 16 at a maximum distance from a nearest neighboring sensor 16 strictly less than half of the smallest wavelength of the acoustic field.

[0101] For example, for a maximum frequency of 100 Hz, the maximum distance between two adjacent sensors 16 is 1.72 meters. Preferably, the arrangement of the sensors 16 is non-coplanar. By "non-coplanar" it is meant that no plane, in a three-dimensional space, includes all the sensors 16. For example, at least one sensor 16 is arranged outside such a plane.

[0102] The number of sensors 16 chosen during step 112 depends in particular on the maximum frequency of the acoustic field to be controlled and the three-dimensional dimensions of the listening zone 19. For example, when the distance chosen between two adjacent sensors 16 is less than half of the shortest wavelength to be controlled in the acoustic field, the three-dimensional dimensions of the listening zone 19 make it possible to obtain the number of sensors 16.

[0103] According to one example, the sensors 16 are arranged in a 3D tetrahedral mesh or in any other regular polyhedron or in a 3D mesh formed of several rows and layers of sensors arranged in a staggered pattern. Preferably, the mesh is chosen to be irregular. Alternatively, a regular mesh can be made irregular by applying a random displacement to the position of each sensor.

[0104] For example, the sensors 16 are arranged to satisfy the conditions of the Nyquist-Shannon theorem.

[0105] During the calibration phase 104, the acoustic field is calibrated in particular according to measurements obtained from the sensors 16.

[0106] Preferably, the calibration phase 104 is implemented at least once before an implementation of the reproduction phase 106.

[0107] The calibration phase 104 comprises for example an emission step 120, a measurement step 122, a definition step 123, a processing step 124, a combination step 126 and a determination step 128.

[0108] In the transmission step 120, the or each loudspeaker 12, 14 receives a predetermined signal and emits a resulting sound wave according to the predetermined signal.

[0109] Preferably, the loudspeakers 12, 14 receive the predetermined signal and emit the sound wave consecutively. Thus, one of the loudspeakers 12, 14 emits the sound wave, then another of the loudspeakers 12, 14 emits it, etc. so that all the loudspeakers 12, 14 of the control system 10 are measured.

[0110] The predetermined signal is, for example, a spectrally dense and / or deterministic excitation signal. For example, the predetermined signal comprises several frequencies in a predetermined frequency range, or an excitation signal whose amplitude and / or frequency is modified over time. For example, the predetermined signal comprises frequencies, and preferably consists of frequencies, which are less than or equal to a predetermined maximum threshold, such as, for example, 100 Hz or 200 Hz.

[0111] During the measuring step 122, each sensor 16 measures a response of the sound wave respectively for the or each loudspeaker 12, 14, to obtain initial measurements. Each initial measurement is associated with one of the loudspeaker(s) 12, 14 and is further associated with one of the sensors 16. Thus, in particular, each initial measurement corresponds to the response of the sound wave emitted by one of the loudspeakers 12, 14 alone or considered separately, and measured by one of the sensors 16.

[0112] During the definition step 123, the controller 18 defines a subset comprising at least one loudspeaker which is the emitting loudspeaker 12, among the set of loudspeakers 12, 14. The controller 18 defines the or each emitting loudspeaker 12 according to a predetermined criterion. The predetermined criterion is in particular that the emitting loudspeaker(s) 12 is / are capable of generating the first wavefront having the predetermined spatial and temporal homogeneity throughout the listening area 19.

[0113] According to one example, the subset comprising the or each transmitting loudspeaker 12 forms a parameter received by the controller 18.

[0114] Preferably, the predetermined criterion defines that the emitting loudspeakers 12 are those which are arranged so as to produce the first wavefront in the listening zone 19 having interferences lower than a predetermined threshold between elementary wavefronts emitted respectively by the emitting loudspeakers 12, for a predetermined frequency band, at each instant and at each position of the listening zone 19.

[0115] In particular, the transmitting loudspeakers 12 defined according to the predetermined criterion are arranged according to the arrangement examples of FIGS. 3 to 11.

[0116] During the processing step 124, the controller 18, and in particular the processing module 22, applies a predetermined function to each initial measurement associated in particular with each transmitting loudspeaker 12 to obtain, for each of these initial measurements, a processed measurement. In particular, the controller 18 applies the same predetermined function to each initial measurement. In particular, the controller 18 applies the predetermined function to each initial measurement which is associated with a transmitting loudspeaker 12 among the transmitting loudspeaker(s) 12, to obtain a processed measurement for each initial measurement which is associated with the corresponding transmitting loudspeaker 12.

[0117] Preferably, the predetermined function modifies an amplitude of a portion of each initial measurement to obtain each processed measurement. In one example, the predetermined function includes a frequency filter modifying the amplitude of each initial measurement as a function of a frequency. For example, some frequency bands are attenuated, and other frequency bands are increased by the frequency filter.

[0118] In one example, the frequency filter is a time-frequency filter that modifies the amplitude of each initial measurement further as a function of time.

[0119] For example, the predetermined function defines at least one time envelope applied to each initial measure.

[0120] The time envelope selects, for example, only a start having a predetermined duration from each initial measurement, with other parts of each initial measurement being eliminated. This corresponds in particular to a truncation of the initial measurement. For example, only the first wavefront of the acoustic field as measured according to each initial measurement is extracted, and in particular the successive sound waves are eliminated. This makes it possible, for example, to eliminate, or at least reduce, the response of the room 20 contained in the initial measurements and present in the listening area 19.

[0121] In one example, the time envelope decreases or increases the amplitude of each initial measurement as a function of time. This makes it possible, in particular, to reduce the duration of the response of room 20, without eliminating it entirely, or to increase the duration of this response.

[0122] In one example, the time envelope has values ​​greater than 1 for certain instants, thus amplifying the amplitude of each initial measurement at that instant.

[0123] According to one example, the time envelope has values ​​greater than 1 and increasing as a function of time, which in particular makes it possible to lengthen the response of room 20, for example in the case where the response would be too short or the effect of the room should be amplified.

[0124] In one example, the time envelope harmonizes a decay (from the English term "time decay" literally meaning "damping time") of the response of the sound wave according to the initial measurement. The decay depends in particular on a frequency of the response of the sound wave according to the initial measurement. For example, the time envelope in this example reduces the decay for some frequencies and increases it for others. In another example, the time envelope adjusts the decay to obtain a decay that is uniform according to the frequency.

[0125] In the combination step 126, the controller 18, and in particular the processing module 22, combines the processed measurements to obtain, for each sensor 16, a target measurement defining a target acoustic field in this position. In particular, the controller 18 separately combines the processed measurements of the measurement position of the corresponding sensor 16, to obtain the target measurement for this measurement position. For example, for a given sensor 16, the controller 18 combines the processed measurement obtained from the initial measurement of a first of the emitting loudspeakers 12 with the processed measurement obtained from the initial measurement of a second of the emitting loudspeakers 12, and this for all the emitting loudspeakers 12. Thus, in particular, each target measurement is associated with the measurement position of the corresponding sensor 16, as for example determined during step 112.

[0126] When the measurement positions of the sensors 16 are arranged so as to obtain an unambiguous measurement of the acoustic field, the target measurements at these positions thus substantially represent the target acoustic field in the listening zone 19.

[0127] In one example, combining the processed metrics includes calculating a sum of the processed metrics to obtain each target metric.

[0128] In this case, for example, the target acoustic field corresponds, in particular when the initial measurements are processed so as to extract only the first wavefront, for example by truncation, to the first uniform and / or interference-free wavefront, and / or directed towards the listening area 19. Preferably, in this case, a wavefront directed outside the listening area 19, for example towards the walls of the room 20, is attenuated by destructive interference, in particular so that resonances and reflections from the room 20 are very little excited, or even absent.

[0129] According to another example, or in addition, the combination of the processed measurements comprises the addition of a specific gain and / or delay, for example before the addition of the processed measurements. This makes it possible, in particular, on the one hand, to adjust a width and / or a shape of the first wavefront and in particular to control the way in which the room is excited, and, on the other hand, makes it possible to compensate for imprecise positioning of the transmitting loudspeakers 12.

[0130] According to another example, or in addition, the combination of the processed measurements comprises the application of a filter modifying the amplitude, the phase and / or the delay of the sound wave of the respective processed measurement. For example, this filter separately modifies the amplitude, the phase and / or the delay, and this for the processed measurement of each emitting loudspeaker 12. This filter is for example constructed by beamforming techniques (literally meaning “beam formation”) in particular making it possible to orient and / or control the width of the first wavefront generated by the combined action of the emitting loudspeakers 12.

[0131] The combining step 126 is an optional step. Preferably, the combining step 126 is implemented in the presence of several transmitting loudspeakers 12.

[0132] For example, the combining step 126 is implemented when the control system 10 comprises several transmitting loudspeakers 12.

[0133] During the determination step 128, the controller 18 determines a set of transformation filters for a transformation of an audio signal 50 by the or each loudspeaker 12, 14, at least as a function of each initial measurement and as a function of the target acoustic field obtained from each processed measurement.

[0134] For example, the controller 18 determines the transformation filters by minimizing a cost function. Parameters of the cost function being at least each initial measurement, in particular considered individually for each measurement position and each loudspeaker 12, 14, and the target acoustic field for each measurement position.

[0135] For example, the cost function determines at least the sum of the least squares errors, for each sensor 16, between the sum of the initial measurements, filtered by the corresponding transformation filter, and the target measurement.

[0136] For example, the cost function is defined as follows:

[0137] 7(5) = HH - g - d | | 2 ,

[0138] J g) is the cost function; g is a vector comprising the transformation filters of said set of transformation filters for each loudspeaker 12, 14;

[0139] H is a global convolution matrix comprising each initial measurement of each loudspeaker 12, 14 measured by each sensor 16, and d is a vector comprising each target measurement of the target sound field at a respective position of each sensor 16, in particular at the respective measurement position.

[0140] For example, let p be the index of one of the P measuring positions of the sensors 16, and let s be the index of one of the S loudspeakers 12, 14.

[0141] In particular, the vector g is defined as follows:

[0142] 9 being the number of speakers.

[0143] Each element gs is a vector of length T g comprising the impulse response of the transformation filter associated with the loudspeaker s for each instant t from f=1 to t=T g :

[0144] The global convolution matrix H is for example a multi-channel convolution matrix for the S loudspeakers 12, 14 and the P measuring positions of the sensors 16, for example defined per block as follows: hyy ... h ls ... h 1S '

[0145] H = h p y ... h ps ... h pS

[0146] -hpy ■" h Ps ■" h PS . with for example h ps being the convolution matrix associated with the initial measure of length T h of the sensor at position p of the response of the sound wave emitted by the loudspeaker s.

[0147] The convolution matrix h ps is for example defined as follows:

[0148] For example, the convolution matrix h ps is a Toeplitz matrix.

[0149] For example, when two speakers from the set of speakers are chosen as speakers 12, 14, and the control system 10 includes two sensors 16, the overall convolution matrix is ​​defined as follows in blocks:

[0150] H = [Æ, , h 12 ; h 21 h 22 ] where the “;” denotes a vertical concatenation.

[0151] The target response c / of the target acoustic field is for example defined as follows: d , with P being the number of sensors 16.

[0152] Each element of p is in particular a vector of size T g comprising the impulse response of the target acoustic field at point p and for each instant t from t=1 to t=T g . d p(l) - rfp (t dp (Tg). As another example, or in addition, the cost function is defined as follows: g) = HH - g - d \ \ 2 + A - \ \B - g\ \ 2 , Or :

[0153] A is a weighting factor, and

[0154] B is a regularization or correction matrix.

[0155] According to examples, the regularization matrix B is obtained by tests or simulations, or is predetermined.

[0156] According to another example, or in addition, the cost function includes a set of N regularization terms Reg n (g) each depending at least on g and advantageously on other parameters r including possibly one or more regularization matrix(ies):

[0157] / (5 1 ) = \\H ■ g - d | | 2 + A^RegKg.F) + ••• + A N . Reg N (g, r)

[0158] According to examples, the vector g is for example defined as follows: or as follows: g = H T H + A ■ B T B 1 ■ (H T d) or as follows: g = F(H,d,r) with F being a predetermined filter calculation operator and r being a set of parameters, including possibly one or more regularization matrix(ies).

[0159] The reproduction phase 106 of the audio signal 50 comprises for example a filtering step 140 and a reproduction step 142.

[0160] During the filtering step 140, the controller 18, and in particular the processing module 22, applies, for each loudspeaker 12, 14, the transformation filter associated with this loudspeaker 12, 14 to the audio signal 50 to obtain a control signal for this loudspeaker 12, 14.

[0161] In the reproduction step 142, the set of loudspeakers 12, 14 reproduces the audio signal 50 according to each corresponding control signal.

[0162] When the reproduction phase 106 comprises the reproduction of the audio signal 50 comprising a bass frequency range, for example less than 100 Hz, according to one example the reproduction phase 106 comprises a preprocessing step, comprising the extraction of the parts of the signal having frequency ranges less than 100 Hz. In this case, for example, only these parts are applied to the transformation filters during the filtering step 140. Preferably, each loudspeaker 12, 14, and in particular each emitting loudspeaker 12, emits sound waves consisting of frequencies less than or equal to a predetermined maximum threshold, such as for example 100 Hz or 200 Hz. This preferably concerns each phase 102, 104, 106 of the method 100. In particular, each loudspeaker 12, 14 is configured to emit only sound waves of frequencies less than the maximum threshold.

[0163] A second embodiment of the control method 100 is described in the following. The control method 100 according to the second embodiment comprises at least some of the features of the first embodiment, except for the differences described below.

[0164] The calibration phase 104 of the control method 100 according to the second embodiment comprises for example the emission step 120, the measurement step 122, the definition step 123, the processing step 124, the combination step 126 and the determination step 128.

[0165] According to one example, the definition step 123 comprises defining the subset comprising several transmitting loudspeakers 12 among the set of loudspeakers 12, 14, according to the predetermined criterion.

[0166] The processing step 124 according to the second embodiment comprises, for example, the determination of a set of calculated responses of the transmitting loudspeakers 12.

[0167] Each calculated response corresponds in particular to an acoustic response of a respective emitting loudspeaker 12 of the subassembly at a position of a respective sensor 16, in particular at the position called the measuring position in the first embodiment. Each calculated response forms in particular a synthetic response.

[0168] For example, each calculated response corresponds to a pulse positioned on the first wavefront emitted by the corresponding emitting loudspeaker 12. The pulse of each calculated response depends in particular on a propagation time between the position of the corresponding emitting loudspeaker 12 and the measurement position. In particular, the set of calculated responses represents an idealized wavefront of an omnidirectional source.

[0169] In the combination step 126, the controller 18 combines calculated responses to obtain, for each sensor 16, and in particular for each measurement position, a target response defining a target acoustic field at this position. For example, the controller 18 combines the calculated responses in an identical manner as the combination of the measurements processed according to the first embodiment. In this case, the controller 18 obtains in particular the target responses instead of the target measurements defined in the first embodiment.

[0170] For example, combining the calculated responses involves calculating a sum of the calculated responses, respectively at each measurement position, to obtain each target response at that position.

[0171] In the determination step 128, the controller 18 determines the set of transformation filters for the reproduction of the audio signal 50 by the set of loudspeakers 12, 14. According to the second embodiment, the set of transformation filters is determined at least as a function of each initial measurement and as a function of the target acoustic field obtained from each calculated response. In particular, in this case the target acoustic field is defined by the target responses.

[0172] In particular, the control method 100 according to the second embodiment comprises determining the calculated response at the location of the measurement processed according to the first embodiment. The control method 100 according to the second embodiment comprises, for example, determining the target response at the location of the target measurement.

[0173] In particular, the control method 100 according to the second embodiment comprises some or all of the characteristics of the method according to the first embodiment, by replacing the target measurement with the target response.

[0174] According to one example, the control method 100 according to the second embodiment comprises applying a predetermined function to obtain the calculated response.

[0175] The preliminary phase 102 and / or the reproduction phase 106 according to the second embodiment are preferably identical to the first embodiment, in particular by applying the aforementioned differences of the second embodiment.

[0176] Figures 12 and 13 are diagrams comprising amplitudes A in dB according to different frequencies F in Hz, of an acoustic signal at a given position of an acoustic field, illustrated over time t in ms. Figure 13 represents the amplitude after the application of the transformation filters, and Figure 12 represents the amplitude without the application of the transformation filters. It is understood that the amplitude A is attenuated in the case of the application of the transformation filters from t = 100 ms, thus corresponding to an example of truncation of the sound wave, making it possible in particular to reduce or eliminate reflections and resonances of the sound wave. In the case of the application of the transformation filters, the duration of the decay of the listening room is greatly reduced. In particular, thanks to the control method, the attenuation of the amplitude A and therefore the reduction of the decay is obtained at each position of the listening zone 19.It is understood that the control method 100 and the control system 10 have a large number of advantages.

[0177] Using initial measurements and processing them, and possibly combining them, and / or determining the set of calculated responses and combining them, to obtain the transformation filters, makes it possible to obtain a high quality of hearing experience, in particular in the entire listening area 19. According to preferred examples, this is also valid outside of the sole measurement positions of the sensors 16 thanks to the unambiguous representation of the acoustic field. Also, thanks to the measurements, the method makes it possible to easily adapt a control of the loudspeakers 12, 14 to a respective listening area 19.

[0178] In particular, the control method 100 is very robust to potential measurement noise. Indeed, low-frequency measurements, such as for example equal to or less than 100 Hz, are very noisy, but since the same noise is found in the target acoustic field by construction, the noise is not directly reflected in the transformation filters. This in particular avoids resorting to high levels of regularization, at least in the useful bandwidth, and thus facilitates the method.

[0179] Also, according to examples, the control method 100 does not require the determination of propagation delays which is particularly imprecise in the low frequencies which greatly compromises the performance in the case of an idealized response.

[0180] Furthermore, the combination step 126 makes it possible to obtain an easily achievable target acoustic field since it corresponds to a combination of measurements or responses from the transmitting loudspeakers 12. This is therefore, by construction, a result that can be achieved at least by the transmitting loudspeakers 12. During the determination step 128, the transformation filters are therefore calculated for a target close to the initial measurements and the filters that minimize the cost function are consequently simple filters. These filters are in particular very simple in their initial part and devoid of audible artifacts such as significant amplifications and long response times. The sound quality of the result is greatly improved.In particular, the combination step 126 makes it possible not to use a global response defining the target acoustic field which would be further from the initial measurements and would produce a greater error in the sense of the cost function and would produce more complex filters. In particular, the combination of responses of the transmitting loudspeakers 12 applies both to initial filtered measurements, namely the processed measurements, and to synthetic responses, namely the calculated responses, for each transmitting loudspeaker 12. The control method 100 makes it possible in particular, over the entire predetermined range of frequencies, such as between 20 Hz and 100 Hz, to obtain a very small variation in the amplitude of the acoustic field obtained during the reproduction of the acoustic signal 50 compared to a reproduction of the acoustic signal 50 without the application of the transformation filters.In particular, the application of the transformation filters obtained during the calibration phase 104 makes it possible to reduce the variation in the amplitude of the sound wave at each position of the listening zone 19.

[0181] The control method 100 makes it possible in particular, over the entire predetermined range of frequencies, such as between 20 Hz and 100 Hz, to obtain an attenuation of the reflections and resonances and in particular a reduced decay time during the reproduction of the acoustic signal 50 compared to a reproduction of the acoustic signal 50 without the application of the transformation filters. In particular, the application of the transformation filters obtained during the calibration phase 104 makes it possible to reduce the variation in the decay time of the listening room at each position of the listening zone 19.

Claims

AMENDED CLAIMS received by the International Bureau on June 18, 2024 (18.06.2024) 1. Method for controlling (100) an acoustic field, comprising a calibration phase (104) comprising steps of: - emission (120) of a sound wave by a set of loudspeakers (12, 14) as a function of a predetermined signal; - measurement (122), respectively for the or each loudspeaker (12, 14), of a response of the sound wave by a plurality of sensors (16), to obtain initial measurements, each initial measurement being associated with one of the or each loudspeaker(s) (12, 14) and being further associated with one of the sensors (16); - definition (123) of a subset comprising at least one loudspeaker, called at least one transmitting loudspeaker (12), among the set of loudspeakers (12, 14), according to a predetermined criterion; - processing (124), comprising the application of a predetermined function to each initial measurement associated with the at least one transmitting loudspeaker (12) to obtain, for each initial measurement associated with the at least one transmitting loudspeaker (12), a processed measurement; - determining (128) a set of transformation filters for a reproduction of an audio signal (50) by the set of loudspeakers (12, 14), the set of transformation filters being determined at least as a function of each initial measurement and as a function of a target acoustic field obtained from each processed measurement.

2. Control method (100) according to claim 1, wherein, during the emission step (120), a plurality of loudspeakers (12, 14) of the loudspeaker assembly emits the sound wave according to said predetermined signal, preferably the loudspeakers (12, 14) emitting the sound wave consecutively according to said predetermined signal.

3. Control method (100) according to claim 1 or claim 2, further comprising a phase of reproducing (106) the audio signal (50), comprising steps of: - filtering (140), comprising applying, for each loudspeaker (12, 14), a transformation filter from the set of transformation filters to the audio signal (50) to obtain a control signal for each loudspeaker (12, 14), and - reproduction (142) of the audio signal (50) by the set of loudspeakers (12, 14) as a function of each corresponding control signal.

4. A control method (100) according to any preceding claim, wherein the predetermined function modifies an amplitude of a portion of each initial measurement to obtain each processed measurement.

5. A control method (100) according to claim 4, wherein the predetermined function defines at least one time envelope applied to each initial measurement, the time envelope selecting only a start having a predetermined duration of each initial measurement, other parts of each initial measurement being eliminated, and / or the time envelope decreases the amplitude of each initial measurement as a function of time, and / or the time envelope harmonizes a damping time of the response of the sound wave according to each initial measurement, and / or the time envelope increases the amplitude of each initial measurement as a function of time.

6. A control method (100) according to claim 4 or claim 5, wherein the predetermined function comprises a frequency filter modifying the amplitude of each initial measurement as a function of a frequency of the initial measurement, preferably the frequency filter being a time-frequency filter modifying the amplitude of each initial measurement further as a function of time.

7. Control method (100) according to any one of the preceding claims, in which the calibration phase (104) further comprises a step of combining (126) the processed measurements to obtain, for each sensor (16), a target measurement defining said target acoustic field.

8. Control method (100) according to claim 7, wherein, during the determining step (128), the set of transformation filters is determined by minimizing a cost function, parameters of the cost function being at least each initial measurement and the target acoustic field.

9. Control method (100) according to claim 8, wherein the cost function determines at least the sum of the errors in the least squares sense, for each sensor (16), between the sum of the initial measurements for each loudspeaker, filtered by the corresponding transformation filter, and the target measurement, preferably the cost function comprises at least the following expression: HH - g - d \ \ 2 where: g is a vector comprising the transformation filters of said set of transformation filters for each loudspeaker (12, 14); H is a global convolution matrix comprising each initial measurement of each loudspeaker (12, 14) measured by each sensor (16); d is a vector comprising each target measurement of the target acoustic field at a respective position of each sensor (16).

10. Control method (100) according to any one of the preceding claims, in which the set of sensors (16) is arranged so as to obtain an unambiguous measurement of the acoustic field, preferably each sensor (16) being arranged at a maximum distance from a nearest neighboring sensor (16) strictly less than half of a smallest wavelength of the acoustic field.

11. Control method (100) according to any one of the preceding claims, in which the subassembly comprises several transmitting loudspeakers (12), the transmitting loudspeakers (12) being arranged so as to produce a first wavefront in a predetermined listening zone (19) having interferences lower than a predetermined threshold between elementary wavefronts respectively emitted by the transmitting loudspeakers (12), for a predetermined frequency band, at each instant and at each position of the listening zone (19).

12. Method for controlling (100) an acoustic field, comprising a calibration phase (104) comprising steps of: - emission (120) of a sound wave by a set of loudspeakers (12, 14) as a function of a predetermined signal; - measurement (122), respectively for the or each loudspeaker (12, 14), of a response of the sound wave by a plurality of sensors (16), to obtain measurements initials, each initial measurement being associated with one of the speaker(s) (12, 14) and being further associated with one of the sensors (16); - definition (123) of a subset of the loudspeakers (12, 14) according to a predetermined criterion, each loudspeaker of the subset being called an emitting loudspeaker (12); - processing (124), comprising determining a set of calculated responses of the transmitting loudspeakers (12), each calculated response corresponding to an acoustic response of a respective transmitting loudspeaker (12) of the subset at a position of a respective sensor (16); - combination (126) of the calculated responses to obtain, for each position, a target response, the target responses defining a target acoustic field, - determining (128) a set of transformation filters for a reproduction of an audio signal (50) by the set of loudspeakers (12, 14), the set of transformation filters being determined at least as a function of each initial measurement and as a function of said target acoustic field obtained from each calculated response.

13. A method (100) for controlling an acoustic field according to any one of the preceding claims, the acoustic field being controlled in a listening area (19) by means of a control system (10) comprising the set of loudspeakers (12, 14), a plurality of sensors (16) and at least one controller (18), the set of loudspeakers comprising at least one emitting loudspeaker (12) and at least one non-emitting loudspeaker (14), the emitting loudspeaker(s) (12) being arranged and controlled so as to be capable of generating a first wavefront having a predetermined spatial and temporal homogeneity throughout the listening area (19) and the non-emitting loudspeaker(s) (14) not contributing to this first wavefront, all the loudspeakers (12,14) of the loudspeaker assembly correcting reflections and resonances of the sound wave emitted by the emitting loudspeaker(s) (12) in the listening area (19) the method comprising, in addition to the calibration phase (104), a preliminary phase (102), the preliminary phase (102) comprising the determination of a number and / or an arrangement of the loudspeakers of the loudspeaker assembly (12, 14) and / or of the sensors of the plurality of sensors (16), so as to obtain the first wavefront emitted by the emitting loudspeaker(s) which has the predetermined spatial and temporal homogeneity., 14. Method for controlling (100) an acoustic field according to claim 13, comprising, in addition to the preliminary phase (102) and the calibration phase (104), a reproduction phase (106) of an audio signal (50) to be reproduced, the reproduction phase (106) comprising a filtering step (140) and a reproduction step (142), in the filtering step (140), the controller (18) applies, for each loudspeaker (12, 14), the transformation filter associated with the loudspeaker (12, 14) considered to the audio signal (50) to be reproduced to obtain a control signal for the loudspeaker (12, 14) considered, in the reproduction step (142), the set of loudspeakers (12, 14) reproduces the audio signal (50) to be reproduced as a function of each corresponding control signal.

15. A method according to claim 13 or claim 14, wherein the transmitting loudspeakers (12) are arranged so that the maximum distance between two adjacent transmitting loudspeakers is such that the waves emitted by each of the two adjacent loudspeakers overlap to within 0.25 wavelengths in the angular sector of the listening area (19).

16. Method according to any one of claims 13 to 15, in which the transmitting loudspeakers (12) are arranged: - according to a configuration called “double bass network”; - in the same plane; - according to a rectangle; or, - along a straight line, preferably parallel to a wall of a room (20) inside which the listening area (19) is located.

17. A method according to any one of claims 13 to 16, wherein the combination for defining the target field comprises applying a filter modifying the amplitude, phase and / or delay of the sound wave of either the processed measurement or the calculated response of each emitting loudspeaker (12), said filter preferably being constructed by "beamforming" techniques making it possible to orient and / or control a width of the first wavefront generated by the combined action of the emitting loudspeakers (12).

18. A computer program comprising software instructions which, when executed by a computer connected to a set of speakers and a plurality of sensors, implement a control method (100) according to any one of the preceding claims.

19. Control system (10) of an acoustic field comprising a set of loudspeakers (12, 14), a plurality of sensors (16) and at least one controller (18), adapted for implementing a control method according to any one of claims 1 to 17.

Images