EP2901717A1

EP2901717A1 - Method and device for generating audio signals to be delivered to a sound reproduction system

Info

Publication number: EP2901717A1
Application number: EP13779301.4A
Authority: EP
Inventors: Etienne Corteel; Khoa-Van NGUYEN
Original assignee: Sonic Emotion Labs
Current assignee: Sennheiser Electronic GmbH and Co KG
Priority date: 2012-09-27
Filing date: 2013-09-25
Publication date: 2015-08-05
Anticipated expiration: 2033-09-25
Also published as: FR2996095A1; EP2901717B1; EP2901717B2; WO2014049268A1; FR2996095B1

Abstract

The invention relates to a method for generating audio signals, from an input audio signal, to be delivered to a sound reproduction system comprising a first set of loud-speakers at a reference height, and a second set of loud-speakers at a higher height than the reference height, that are vertically more directive than the loud-speakers of the first set. Said method comprises: the identification of a position of at least one sound object of the input audio signal; the generation of first audio signals for feeding the first set, by applying, to the input audio signal, at least one delay and/or a gain and a sound spatialisation technique taking into account the identified position and a position of the loud-speakers of the first set; and the generation of second audio signals for feeding the second set, by applying, to the input audio signal, at least one delay and/or a gain and a sound spatialisation technique taking into account the identified position and a position of the loud-speakers of the second set; at least one sound spatialisation technique being a holophonic technique and the delays and/or gains being selected in such a way that the combination of the first signals restituted by the first set and the second signals restituted by the second set verifies a defined sound restitution effect at at least one reference spatial position.

Description

Method and device for generating audio signals for delivery to a sound reproduction system

Background of the invention

The invention relates to the general field of acoustic treatments.

It relates more particularly to a technique of sound reinforcement. In known manner, sound reinforcement techniques are commonly used for the sound of extended rooms, such as for example a concert hall or an auditorium, in order to obtain a uniform sound level signal over the entire room.

These techniques must verify a double constraint, namely to ensure a sufficiently high level of sound for the listeners located at the back of the room without unduly increasing this sound level, in particular to limit the discomfort of the listeners located near the stage or speaker.

For this purpose, known sound reinforcement systems rely on Public Address (PA) sound broadcasting systems making it possible to transmit to listeners placed at the back of the room at a distance from the stage, possibly combined with high lines. low-power back-up speakers ("front fill") placed at the edge of the stage and intended for listeners located in front of and in the axis of the stage.

These systems, however, suffer from various disadvantages.

Thus, in particular, the placement of loudspeakers on the stage can be problematic, in order to respect certain practical and aesthetic constraints (limited space of the stage, visual rendering for the listener, etc.).

Moreover, these systems do not generally offer the possibility of having a spatialized sound reproduction of the signals in the room.

There are also in the state of the art, sound reinforcement systems that rely on holophonic techniques known as Wavefront Synthesis or Wave Field Synthesis (WFS) in English, and offer a spatialized sound reproduction of signals in the room.

These systems use horizontal lines of omnidirectional loudspeakers, placed on the stage, which make it possible to restore the sound objects of the signals to various precise positions in the room (ie good horizontal location of the sound objects), so as to create among the listeners a perception of envelopment in the sound stage, as well as a coherence between the visual and auditory information perceived by these listeners, and a better intelligibility of the sound objects.

Such a system is presented for example in document EP 2 206 365. However, if the use of such systems for sound reinforcement allows a good horizontal and vertical location of the sound objects of the signals, it nevertheless encounters certain limits. Indeed, the attenuation of the sound level of the restored signals as a function of the distance to the scene is relatively fast and therefore requires a substantial increase in the sound level restored by the line of speakers, which can cause discomfort for the listeners close to the stage.

By way of example, FIG. 1A illustrates in full line the variations of the sound level expressed in decibel, of a signal rendered using a WFS type technique by a set of 16 omnidirectional loudspeakers positioned on a stage, these high speakers having a sensitivity of 92 dB and being powered by amplifiers of 250 W (watts).

In addition, such a system generally requires the use of a large number of speakers, which can be problematic for practical and aesthetic reasons, as mentioned above.

One solution to overcome some of these disadvantages is to place horizontal lines of omnidirectional loudspeakers above the stage.

As illustrated in broken lines in FIG. 1A for a line of loudspeakers consisting of 9 loudspeakers placed at a height of 6 meters from the stage (sensitivity of 97 dB and amplifiers of 250 W), such a solution makes it possible to restore a better sound level for the listeners located at the back of the room while limiting the sound level diffused near the listeners located near the stage.

However, this solution has the disadvantage of a poor perception by the listener of the vertical location of sound objects of the restored signal, especially when it is close to the scene.

FIG. 1B illustrates by way of example (on the basis of assumptions similar to those described above for FIG. 1A):

- in full line, the elevation (also called site) perceived as a function of distance by a listener for a row of omnidirectional loudspeakers placed at the stage level; and

- in broken lines, the distance perceived by a listener for a line of omnidirectional loudspeakers placed at a height of 6m above the level of the stage.

There is therefore a need for a sound reinforcement technique and device that does not have the various aforementioned disadvantages.

Object and summary of the invention

The invention responds in particular to this need by proposing a method of generating, from an audio input signal, audio signals intended to be supplied to a sound reproduction system, this sound reproduction system comprising a first set of loudspeakers positioned at a reference height, and a second set of loudspeakers positioned at a height greater than the reference height, the loudspeakers of the second together being more directional vertically than the speakers of the first set. This generation method comprises:

A step of identifying a position of at least one sound object of the input audio signal;

A step of generating a plurality of first audio signals intended to feed the first set of loudspeakers, by applying to the input audio signal:

o at least one delay and / or gain; and

a sound spatialization technique taking into account the identified position of said at least one sound object of the input audio signal and a position of the speakers of the first set;

A step of generating a plurality of second audio signals intended to feed the second set of loudspeakers, by applying to the input audio signal:

o at least one delay and / or gain; and

a sound spatialization technique taking into account the identified position of said at least one sound object of the input audio signal and a position of the speakers of the second set;

at least one of the sound spatialization techniques applied during the generation of the first and second audio signals being a holophonic technique; and

the delays and / or gains applied during the generation of the first and second audio signals being chosen so that the combination of the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers; speakers verifies a determined sound reproduction effect at at least one reference spatial position defined for the first and second sets of speakers.

Correlatively, the invention also relates to a device for generating audio signals intended for a sound reproduction system from an input audio signal, this sound reproduction system comprising a first set of loudspeakers positioned at a height of reference, and a second set of speakers positioned at a height greater than the reference height, the speakers of the second set being more vertically oriented than the speakers of the first set. This generation device comprises:

Means for identifying a position of at least one sound object of the input audio signal;

Means for generating a plurality of first audio signals intended to feed the first set of loudspeakers, able to apply to the input audio signal:

o at least one delay and / or gain; and

Means for generating a plurality of second audio signals intended to feed the second set of loudspeakers, able to apply to the input audio signal: o at least one delay and / or gain; and

at least one of the sound spatialization techniques applied by the means for generating the first and second audio signals being a holophonic technique; and

the gains and / or delays applied by the means for generating the first and second audio signals being chosen so that the combination of the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers; speakers verifies a determined sound reproduction effect at at least one reference spatial position defined for the first and second sets of speakers.

In other words, the invention proposes, in order to optimize the reproduction of the audio input signal in a room or a room, to combine two advantageous techniques, namely:

- a sound reinforcement technique, based on a system of restitution consisting of two sets of separate loudspeakers, positioned at different heights and having different vertical directivities; and

A sound spatialization technique used to generate, from the input audio signal, the audio signals intended to supply these two sets of loudspeakers.

The invention thus applies in a preferred but non-limiting manner to the restitution of audio signals in a large room, such as a theater or an auditorium with a stage.

Preferably, to optimize the rendering of the input audio signal, the first set of speakers is placed on or in front of the stage of the room, while the second set of speakers is placed above the stage (aligned vertically or slightly offset from the first set of speakers).

In accordance with the invention, the loudspeakers of the highest set (second set) are more directional (on average) than the speakers of the lowest set (first set).

It is therefore possible to vertically orient the second set of speakers differently from the first set, so that the sound reproduced by these two sets of speakers covers the room evenly in terms of sound level.

More specifically, the invention provides the possibility for this purpose to vertically orient the speakers of the second set towards the back of the room, while the speakers of the first set can be directed to the part of the room located at near the scene. Thus it is possible to broadcast the input audio signal with a uniform sound level, even at a great distance from the scene, without increasing too much this sound level. In other words, the vertical directivity of the speakers of the second set and its positioning in height limits the sound level restored close to the scene while ensuring a good sound level for listeners placed in the back of the room.

The speakers of the second set therefore acts as a booster speaker for the return of the input audio signal.

In addition, the presence of directional booster speakers above the stage advantageously reduces the power of the speakers used for the first set. Therefore, one can use for the first set of speakers less bulky because of their reduced power or more limited number. The invention makes it possible to overcome more easily practical and aesthetic constraints related to the positioning of the speakers in the room to restore the signals.

Moreover, the limitation of the sound level reproduced by the second set of speakers in the vicinity of the scene authorized by the invention makes it possible to favor the vertical localization of the sound objects of the signal at the level of the first set of loudspeakers, benefiting from them. an effect known as "Haas effect".

This Haas effect, which concerns the perception of sounds by the human brain, is described in particular in the J. Blauert document entitled "Spatial Hearing - Revised Edition: The Psychophysics of Human Sound Localization", MIT Press, Rev. Sub. Edition, October 2, 1996.

According to this effect, the location of a sound object is given by the direction from which the sound comes first to the ear (or first wavefront), even if the intensity of the sounds arriving in a second ear delay from another direction is greater than the intensity of the first wavefront. We speak of "effect of precedence". This effect, however, is not absolute and works all the better as the sound level of the following wavefronts is low.

In addition, the use of two sets of loudspeakers for reproducing the input audio signal and the directivity of the loudspeakers of the second set combined with the use of a sound spatialization technique are taken into account, in accordance with the invention, by applying delays and / or gains on the input audio signal during the generation of signals for powering the first and second sets of speakers.

These delays and gains are applied independently for each set of speakers, and a spatial sound space is envisioned for each set of speakers separately.

However, the delays and the gains applied in accordance with the invention are advantageously chosen so that the combination of the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers verifies a signal effect. sound reproduction determined at a reference spatial position defined in the room for the first and second sets of speakers. This effect of sound restitution determined is for example the coherence (of phase) of the signals restored by the first and second sets of loudspeakers at the reference spatial position. For this purpose, the delays applied during the generation of the first and second audio signals are chosen so that the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers arrive simultaneously. at the reference spatial position.

As a variant, it is possible to choose to accentuate the natural Haas effect previously described, caused by the height positioning and the vertical directivity of the second set of speakers, by applying during the generation of the first and second audio signals. , delays chosen so that the second signals output by the second set of speakers arrive with a predetermined offset to the reference spatial position with respect to the first signals output by the first set.

This time shift between the signals is then chosen short enough to be able to benefit from the Haas effect (that is to say typically less than 30-35 ms, for example of the order of 15 to 20 ms), and not not generate a feeling of echoes for the audience of the room.

Thanks to the Haas effect thus accentuated, a listener will hear in the first place the signals rendered by the loudspeakers of the first set placed at the reference height (for example on stage), which makes it possible to further improve the vertical location of the sound objects perceived by the listener.

In addition, emphasizing the Haas effect through the choice of delays applied to the first and second signals offers the possibility of further reducing the power of the loudspeakers of the first set, and thus of using loudspeakers. less bulky speakers or even fewer.

The gains, in one or the other of the variants envisaged for the adjustment of the delays, are chosen so as to guarantee a sufficient sound level at the bottom of the room while optimizing the vertical localization with the reference position.

The invention therefore makes it possible, on the one hand, to improve the quality of the signal reproduction compared to the state of the art, and to respond more effectively to the aesthetic and practical constraints imposed when positioning the loudspeakers in the room. .

In a particular embodiment, at least one of the sound spatialization techniques applied during the generation of the first and second audio signals is:

A holophonic sound spatialization technique of the WFS type, such as, for example, the sound spatialization technique described in document EP 2 206 365, or the technique described in the E. Corteel document titled Equalization in an Extended Area using

Multichannel Inversion and Wave Field Synthesis, J. Audio Eng. Soc. Flight. 54, No. 12, December 2006: such a technique advantageously makes it possible to envelop the audience of the room in the sound stage and offer a good quality of reproduction of the audio signal; or - a dynamic matrix mixing technique. This technique can take the form of, for example, a stereophonic loudspeaker panning on the loudspeakers, relatively easy to implement, or alternatively a VBAP (Vector Based Amplitude Panning) or DBAP type of intensity panning. (Distance Based Amplitude Panning) known per se.

As a variant, it is of course possible to envisage other sound spatialization techniques.

In a particular embodiment, the input audio signal is a multichannel audio signal and, during at least one of the generation steps, the same delay and / or the same gain is applied to each channel of the signal input audio.

This embodiment facilitates the implementation of the invention, in particular by limiting the complexity associated with the choice of gains and / or delays to be applied.

Although preferentially during at least one of the generation steps, said at least one delay and / or gain is applied to the input signal before the sound spatialization technique, it is also possible to invert these two treatments.

In this case, and when the input audio signal is a multichannel signal comprising a number of channels smaller than the number of loudspeakers of the first and / or second set, it will be preferred to apply the same delay and / or of the same gain to all channels of the input audio signal (that is, the delay and gain applied to the input audio signal is independent of the channel and the position of the sound objects of the input audio signal).

Moreover, when the application of said at least one gain and / or delay is implemented after the sound spatialization technique, the use of a gain and / or delay independent of the signal channels (and therefore identical for all channels) allows to preserve the distribution of the gains and delays affected between the loudspeakers by the sound spatialization technique.

It should be noted that when the sound spatialization techniques applied during the generation of the first and second loudspeaker supply signals are linear (this is the case, for example, of WFS type techniques or more generally of techniques based on on the use of linear filtering) and that the gains and / or delays applied to the input signal are independent of the channels of this input signal, the first and the second signals generated by applying the gains and the delays and the sound spatialization technique are equivalent to the first and second signals generated by first applying the sound spatialization technique and then the gains and delays.

In another embodiment, it can be envisaged that during at least one of the generation steps, a delay and / or a distinct gain is applied to at least two distinct channels of the input audio signal.

In a particular embodiment, the sound reproduction system further comprises at least a third set of loudspeakers positioned at a height greater than the reference height, and the generation method comprises a step of generating a loudspeaker. plurality of third audio signals for powering the third set of speakers by applying to the input audio signal:

At least one delay and / or gain; and

A sound spatialization technique taking into account the identified position of said at least one sound object of the input audio signal and a position of the loudspeakers of the third set;

said at least one delay and / or gain being chosen so that a combination of the third signals output by the third set of speakers and signals output by one of the other speaker sets of the playback system checks a sound reproduction effect determined at a reference spatial position defined for this other set and for the third set of speakers.

This embodiment proposes to add a set of additional booster speakers to enhance the sound reinforcement of the room.

Such an embodiment can be very relevant for rooms of very large dimensions or having elements acoustically masking the signals returned by the loudspeakers (eg presence of balconies in a concert hall).

It should be noted that the invention is not limited to a system comprising three sets of loudspeakers but it is also possible to add a fourth, a fifth, etc., set of loudspeakers, the choice of delays and / or gains applied. during the generation of the signals intended to supply these sets of loudspeakers being preferentially carried out by considering the sets of loudspeakers in pairs (eg second and third sets, third and fourth sets, etc.).

In another embodiment, it is conceivable to add also to the restitution system another set of speakers of the type subwoofers (or "subwoofers" in English), to add an additional effect during the sound reproduction of the input audio signal. Similarly, in this embodiment, the generating method comprises a step of generating a plurality of audio signals for supplying the set of subwoofer loudspeakers by applying to the input audio signal:

At least one delay and / or gain; and

A low-pass type filtering;

said at least one delay and / or gain being selected so that the combination of the signals output by the set of subwoofer type speakers and signals output by one of the other sets of speakers of the system of restitution verifies a determined sound reproduction effect at a reference spatial position defined for this other set and for the set of speakers of the subwoofer type.

In a particular embodiment, the different steps of the generation method according to the invention are determined by instructions of computer programs. Accordingly, the invention also relates to a program on an information medium, this program being capable of being implemented in a generation device or more generally in a computer, this program comprising instructions adapted to the implementation steps of a generation method as described above.

This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape.

The invention also relates to a computer-readable or microprocessor-readable information medium, and comprising instructions of a program as mentioned above.

The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a floppy disk or a disk. hard.

On the other hand, the information medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network.

Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question.

The invention also relates to a set of reproduction of an input audio signal comprising:

A sound reproduction system comprising a first set of loudspeakers positioned at a reference height, and a second set of loudspeakers positioned at a height greater than the reference height, the loudspeakers of the second set being more directive vertically as the speakers of the first set; and

A device for generating audio signals according to the invention for supplying the loudspeakers of the first set and the loudspeakers of the second set of the sound reproduction system.

The reproduction unit has the same advantages as those mentioned above for the device and the generation method.

It can also be envisaged, in other embodiments, that the generation method, the generating device and the restitution assembly according to the invention present in combination all or part of the aforementioned characteristics. Brief description of the drawings

Other features and advantages of the present invention will be described in the description below, with reference to the accompanying drawings which illustrate embodiments having no limiting character. In the figures:

FIGS. 1A and 1B, already described, respectively illustrate the disadvantages of sound reinforcement systems of the state of the art;

FIG. 2 schematically represents a reproduction assembly and a generating device according to the invention, in a first embodiment;

FIG. 3 illustrates an example of positioning of the two sets of loudspeakers of the restitution assembly of FIG. 2 in a room;

FIG. 4 schematically illustrates the hardware architecture of the generation device of FIG. 2, in the first embodiment;

FIG. 5 represents, in the form of a flow chart, the main steps of a generation method according to the invention as implemented in the first embodiment, by the generation device of FIG. 2;

FIG. 6 shows the variations in the propagation time of the signals emitted by omnidirectional loudspeakers situated on and above the stage, as a function of the distance of a listener with respect to the scene;

FIG. 7 illustrates various advantages provided by the invention, in terms of variations in the sound level of the signals as a function of the distance of a listener with respect to the scene;

FIG. 8 schematically represents a reproduction assembly and a generation device according to the invention, in a second embodiment;

FIG. 9 illustrates an example of positioning of the three sets of loudspeakers of the rendering assembly of FIG. 8 in a room; and

FIG. 10 represents, in the form of a flow chart, the main steps of a generation method according to the invention as implemented in the second embodiment, by the generation device of FIG. 8.

Detailed description of the invention

FIG. 2 represents, in its environment, a set 1 of sound reproduction of an input audio signal S, according to the invention, in a first embodiment.

In the example considered here, it is assumed by way of illustration that the audio signal S is a multichannel signal resulting from a sound capture (ie recording) of a signal played on a scene SCENE of a theater R , performed by microphones placed on the SCENE stage or distributed in various places of this room, in a manner known per se.

This assumption is not, however, limiting, the audio signal S possibly being a pre-recorded audio signal or having prerecorded elements. It is furthermore assumed that the audio signal S is intended to be reproduced in the theater by using the sound reproduction assembly 1 according to the invention.

No limitation is strictly attached to the format of the S signal: thus, it may be a signal composed of individual sockets (eg for a theater, from a microphone per musician, singer or actor for live sound recording), a stereo signal, a 5.1, 7.1, or 10.2 format signal, or a signal in HOA (Higher Order Ambisonics) format, etc., or finally a combination of these techniques.

According to the invention, the set 1 of sound reproduction used for the reproduction of the audio signal S comprises:

A sound reproduction system 2 comprising a plurality of loudspeakers; and

A device 3 for generating audio signals intended to supply the loudspeakers of the sound reproduction system 2.

More specifically, in the first embodiment illustrated in FIG. 2, the sound reproduction system 2 comprises:

A first set E1 of P loudspeakers denoted HP1-1, HP1-2, HP1-P, P being a positive integer greater than 1; and

A second set E2 of N loudspeakers, denoted HP2-1, HP2-2,..., HP2-N, N being a positive integer greater than 1.

The integers P and N can be equal or distinct.

As illustrated in FIG. 3, the loudspeakers HP1-1, HP1-2, HP1-P of the first set El are positioned at a height h1 referred to as reference.

More precisely, in the example of FIG. 2, the loudspeakers HP1-1, HP1-2, HP1-P of the first set El are arranged so as to form a horizontal line of loudspeakers positioned on the SCENE stage of R. The reference height hl corresponding here to the height of the SCENE scene.

In accordance with the invention, the loudspeakers HP2-1, HP2-2,..., HP2-N of the second set E2 are positioned at a height h2 greater than the reference position h1. In the example of Figure 2, they are arranged to form a horizontal line of speakers positioned at height h2.

It should be noted that the loudspeakers of the second set E2 are not necessarily vertically aligned with (i.e. in the same vertical plane as) the speakers of the first set El, with respect to the scene SCENE. They can be shifted in particular with respect to the speakers of the first set El in the direction of the AUDIT part of the auditorium R where the audience is.

Moreover, no limitation is attached strictly speaking to the arrangement of the speakers in the first set El and the second set E2. The invention thus also applies to other configurations of loudspeakers, such as for example to matrices speakers, or configurations distinct from a linear configuration, such as a configuration adapted to the shape of the scene (eg in an arc).

Similarly, the positioning of the loudspeakers of each set, relative to each other, can be varied and varied: it will be possible to consider speakers distributed uniformly within the set of speakers, or distributed according to several groups placed at the ends and at the optical center of the scene, etc.

Several examples of speaker arrangement will be described later to illustrate the invention.

It should be noted that the invention advantageously offers the possibility of having a number N of loudspeakers of the second set E2 smaller than the number P of loudspeakers of the first set El.

Indeed, in a manner known per se, the use of finite size speaker assemblies induces a phenomenon of aliasing ("aliasing" in English) of spatial spectrum. This phenomenon occurs from a folding frequency or "aliasing" frequency which, as mentioned in EP 2 206 365, increases as a function of the distance of the listener from the scene.

The number of speakers required for each set of speakers can therefore be determined by setting the desired aliasing frequency and the position of a reference listener in the room relative to the scene. Since the loudspeakers of the second set E2 are further away from the reference listener than the loudspeakers of the first set E1, because of their positioning in height, it is possible by this means to set a more limited number of loudspeakers. speakers for the set E2 that for the set El.

According to the invention, the speakers HP2-1, HP2-2, ..., HP2-N of the second set E2 are more vertically oriented than the speakers HP1-1, HP1-2, HP1-N of the first set El.

For the purposes of the invention, the "speakers HP2-1, HP2-2, ..., HP2-N of the second set E2 are more vertically oriented than the speakers HP1-1, HP1-2 , HP1-N of the first set El "that the average directivity of the speakers of the second set is greater than the average directivity of the speakers of the first set.

Thus, by way of illustration, it is possible to use omnidirectional loudspeakers for the first set E1, while the loudspeakers of the second set E2 are chosen with a vertical opening of between 10 ° and 60 °.

As mentioned above, the speakers HP2-1, HP2-2, ..., HP2-N of the second set E2 are primarily intended to provide the sound coverage of the part of the theater, or at least the audience, the farthest from the stage. To this end, they are oriented vertically towards the EXT end of the audience (which corresponds in Figure 3 at the back of the theater for the sake of simplification). In contrast, the speakers HP1-1, HP1-2, HP1-P of the first set El are primarily intended to provide the sound coverage of the part of the theater, or at least the audience, placed at near the stage. To this end, they are mounted on the stage or just in front of the stage.

They can be indifferently omnidirectional or directional vertically. When the speakers of the first set El are directional vertically, they are oriented towards the front of the AUDIT part of the room, that is to say, towards the part located near the scene.

The different loudspeakers of the sets E1 and E2 are powered by multichannel audio signals, denoted respectively Sl-1, Sl-2,..., Sl-P and S2-1, S2-2, S2-N, generated according to to the invention by the generation device 3 from the input audio signal S.

To generate these signals, in the first embodiment described here, the generation device 3 relies on various software modules, namely:

A module 3A, able to identify the positions of the different objects OBJ (or sources) sound included in the input signal S;

Modules 3B1 and 3B2 capable of independently applying on the input signal S, a delay τ1 and a gain G1, and a delay τ2 and a gain G2 respectively. The delays τΐ and τ2 can be positive, negative or zero, and the gains G1 and G2 can be lower, greater than or equal to 1; and

Modules 3C1 and 3C2 capable of independently applying on the signals coming respectively from the modules 3B1 and 3B2, sound spatialization techniques noted respectively T1 and T2.

The characteristics of these modules will be described in more detail later. The generation device 3 here has the hardware architecture of a computer, as schematically illustrated in FIG. 4.

It comprises in particular a processor (or microprocessor) 4, a random access memory 5, a read-only memory 6, a non-volatile flash memory 7 as well as communication means 8 able to transmit and receive audio signals.

The communication means 8 comprise, on the one hand, means for receiving the input audio signal S, and, on the other hand, an interface (wired or wireless) with the loudspeakers HP1-1, HP1-2, HP1-P of the set El and with the loudspeakers HP2-1, HP2-2, ..., HP2-N of the set E2 to provide them with the audio signals Sl-1, Sl-2, .. ., Sl-P and S2-1, S2-2, S2-N. These means are known to those skilled in the art and will not be described further here.

The read-only memory 6 of the generation device 3 constitutes a recording medium in accordance with the invention, readable by the (micro) processor 4 and on which is recorded a computer program according to the invention, comprising instructions for performing the steps of a generation method described later with reference to Figure 5. It should be noted that no limitation is attached to the nature itself of the generation device 3.

Thus, in particular, the generation device 3 may be in the form of a computer as illustrated in FIG. 4, or alternatively of an electronic chip or of an integrated circuit, in which the computer program comprising the instructions for carrying out the generation method according to the invention is incorporated.

Similarly, the modules 3A, 3B1, 3B2, 3C1 and 3C2 on which the generation device 3 is based can alternatively be DSP (Digital Signal Processor) or FPGA (Field Programmable Gate Array) modules.

We will now describe, with reference to FIG. 5, the main steps of a generation method according to the invention, as they are implemented, in the first embodiment, by the generation device 3, on receiving the input audio signal S, for generating and supplying the signals S1-1, S1-2, ..., S1-P and S2-1, S2-2, S2-N to the high-end sets E1 and E2; speakers respectively.

It is therefore assumed here that the generation device 3 receives the multichannel input audio signal S via its communication means 8 (step E10). As mentioned above, no limitation is attached to the way the audio signal S is generated nor to its actual format.

In the first embodiment described here, it is furthermore assumed that the generating device 3 is supplied, via a suitable interface, with the POS positions (OBJ) of the sound objects OBJ included in the signal S and that we want to restore. These positions are identified in a horizontal plane, namely in the horizontal plane in which the speakers of the first set are located (ie the scene plane SCENE here, S being obtained by capturing the played signals on the scene S), in terms of angle and distance with respect to a predetermined and predetermined point of origin.

This interface is for example a graphical interface developed for this purpose allowing a user to interact with the device 3, or a communication interface developed with a mixer operating on the signal S.

Obtaining these positions by the generation device 3 constitutes an identification step (step E20) of the position of at least one sound object of the input audio signal S within the meaning of the invention.

In another embodiment, the POS (OBJ) positions of the OBJ sound objects are identified directly by the generation device 3 and more specifically by the module 3A.

To this end, the module 3A can use a position tracking device (or "tracking" in English) of sound objects present on the scene, known per se, and operating with optical and / or electromagnetic signals, and / or acoustic, etc. In yet another embodiment, in order to determine the position of the sound objects of the signal S, the module 3A of the generation device 3 implements a method of spatial analysis known per se, based on a decomposition of the signal S in sub-phase. frequency bands.

Various spatial analysis methods can be used by module 3A.

For example, the module 3A can use a spatial analysis method based on the evaluation of correlations and level differences between the channels of the audio signal S.

Alternatively, the module 3A may use a spatial analysis method based on the determination of a Gerzon vector representative of the input audio signal S, as described in US 2007/0269063. The Gerzon vector of a multichannel audio signal reflects in a known manner the spatial location of the multichannel audio signal as perceived by the listener from a given position.

According to another variant, the module 3A can implement a spatial decomposition of the multichannel signal in spherical harmonics, as described in the document WO 2012/025580. This decomposition allows a very precise spatial analysis of the multichannel audio signal and the sound objects composing it. Thus, in particular, several sound objects can be identified for the same frequency subband.

According to the invention, two types of processing are also applied by the generating device 3 to the input audio signal S to generate the signals intended to supply the two sets E1 and E2 of loudspeakers.

These treatments are applied independently for each set of loudspeakers, via two separate signal processing chains comprising for one the module 3B1 and the module 3C1, and for the other the module 3B2 and the module 3C2. These two treatment chains preferentially operate simultaneously.

Thus, more specifically, in the first embodiment described here, the module 3B1 first applies to the input signal S, the delay τΐ and the gain G1 previously mentioned (step E30). This delay τΐ and this gain G1 have been chosen and supplied to the module 3B1 during a preliminary configuration or calibration step (step E_CALIB) described later.

For the sake of simplification, it is assumed here that the same gain G1 and the same delay τΐ are applied by the module 3B1 to all the channels of the input signal S.

This hypothesis is however not limiting and in another embodiment, it is conceivable to apply a distinct gain and delay to the different channels constituting the input signal S.

The signal SI resulting from the application of the delay τΐ and the gain G1 on the signal S is then supplied by the module 3B1 to the module 3C1.

The module 3C1 applies the sound spatialization technique T1 to the signal SI, and thus generates the plurality of signals Sl-1, Sl-2,..., Sl-P (step E40). The T1 technique is for example a holophonic technique of the wavefront synthesis type, also known under the name of Wave Field Synthesis (WFS) technique, as described in document EP 2 206 365 or in the document D E. Corteel cited above.

It takes into account, on the one hand, the POS (OBJ) positions of the sound objects identified and provided by the module 3A, and on the other hand, the position of the speakers HP1-1, HP1-2, HP1-P of the set El, to create virtual sources from the speakers HP1-1, HP1-2, HP1-P from the set El to the positions POS (OBJ).

Of course, this hypothesis is not limiting and other sound spatialization techniques can be envisaged within the scope of the invention, such as dynamic matrix mixing techniques taking the form of, for example, a stereophonic intensity panning. on the speakers, or a VBAP or DBAP type of intensity panning known to those skilled in the art.

In the first embodiment described here, it is assumed that the position of the loudspeakers HP1-1, HP1-2, HP1-P of the set E1 used by the module 3C1 for applying the sound spatialization technique T1 has been provided. to the generation device 3 during the preliminary step E_CALIB, for example by an installer of the set 1 rendering, via a suitable user interface.

Alternatively, the position of the loudspeakers HP1-1, HP1-2, HP1-P of the set E1 can be determined by the generating device 3 itself, automatically or semi-automatically, using for example a position tracking device as described above.

The signals Sl-1, Sl-2,..., Sl-P generated by the module 3C1 are then supplied to the loudspeakers HP1-1, HP1-2, HP1-P of the set El (step E50) for to be returned.

In parallel, and similarly, the module 3B2 applies to the input signal S, the delay τ2 and gain G2 mentioned above (step E60), determined and supplied to the module 3B2 during the configuration step E_CALIB also.

The same gain G2 and the same delay τ2 are applied here to all the channels of the input signal SIG.

The signal S2 resulting from this operation is provided by the module 3B2 to the module 3C2. The latter applies to the signal S2 the sound spatialization technique T2, and thus generates the plurality of signals S2-1, S2-2, ..., S2-N (step E70).

The T2 technique is for example also a holophonic technique of the WFS type as described in the document EP 2 206 365. It takes into account on the one hand the POS (OBJ) positions of the sound objects identified and provided by the module 3A. and on the other hand, the position of the speakers HP2-1, HP2-2, ..., HP2-N of the set E2, to create virtual sources from the loudspeakers HP2-1, HP2- 2, ..., HP2-N from the set E2 to the POS (OBJ) positions. Of course, this hypothesis is not limiting and other sound spatialization techniques can be envisaged within the scope of the invention, such as dynamic matrix mixing techniques mentioned above.

Moreover, it can also be envisaged that the modules 3C1 and 3C2 apply separate sound spatialization techniques T1 and T2 to the signals S1 and S2 for generating the signals intended to supply the two sets E1 and E2 of loudspeakers.

The position of the speakers HP2-1, HP2-2, ..., HP2-N of the set E2 has, in the first embodiment described here, been provided to the device 3 of generation also during the step of preliminary configuration E_CALIB, by the installer of the set 1 rendering via the aforementioned user interface.

The signals S2-1, S2-2, ..., S2-N generated by the module 3C2 are supplied to the loudspeakers HP2-1, HP2-2, ..., HP2-N of the set E2 (step E80), to be restored.

According to the invention, the delays and the gains (T1, G1) and (T2, G2) applied to the input audio signal S by the modules 3B1 and 3B2 for respectively generating the signals Sl-1, Sl-2, ..., S1-P and S2-1, S2-2, ..., S2-N, have been chosen, during the preliminary configuration step E_CALIB, so as to ensure that the combination, at a position of reference Pref (E1, E2) defined for the sets of speakers E1 and E2, signals restored by the first set E1 of the loudspeakers supplied by the signals S1-1, S1-2, ..., S1-P , and signals output by the second set E2 of loudspeakers fed by the signals S2-1, S2-2, ..., S2-N, verify a determined sound effect (in other words, desired, chosen).

This advantageously makes it possible to take into account the vertical directivity of the loudspeakers of the second set E2, and to ensure that the combination of the signals at the selected reference spatial position is harmonious for the listener (no cacophony or echo Not wished).

By reference spatial position, here we mean both a point in the space characterizing the position of a target listener in the room, and a wider area of the space in which is (are) likely to find one or more listeners.

The choice of the reference position Pref (E1, E2) depends of course on several criteria, such as in particular the depth of the room in which the audio signal S is restored. It is chosen preferentially around the middle of the room, slightly shifted towards the stage.

FIG. 3 illustrates an example of the choice of reference position Pref (E1, E2) for room R.

Other criteria can be taken into account for the choice of the reference spatial position Pref (E1, E2), for example:

The position Pref (E1, E2) can be chosen so as to guarantee that the vertical position of the second set E2 of loudspeakers perceived at this position Pref (E1, E2) is approximately equal to an angle (for example 45 °), selected to limit the vertical perception error;

The position Pref (E1, E2) may coincide with the spatial position at which the sound level "naturally" restored (that is to say without applying gains to the signals) by each set of speakers E1 and E2 is similar;

The position Pref (E1, E2) can also depend on the configuration of the room, and in particular the location, if necessary, of the seats intended for the listeners in the room (eg existence of corridors in the room, etc.).

It should be noted that in the first embodiment described here, it is limited to take into account a single reference position in the room to choose the delays and / or gains applied to the input signal S. However, in a variant , one could consider taking into account several reference positions spread over a specific area of the room.

As mentioned above, the delays and the gains applied to the input signal depend on the sound reproduction effect that it is desired to obtain at the reference position Pref (E1, E2) for the combination of the signals restored by the first set El and by the second set E2.

Thus, for example, the desired sound reproduction effect may be the arrival in phase coherence, at the reference spatial position Pref (E1, E2), of the signals restored by the first set El and of the signals restored by the second together E2. This phase coherence is ensured when the signals restored by the first set E1 and the signals restored by the second set E2 simultaneously arrive at the reference position Pref (E1, E2) with a similar sound level.

According to another example, it is possible on the contrary to choose to shift in time the arrival of the signals restored by the second set E2 with respect to the arrival of the signals restored by the first set El so as to benefit from the Haas effect described. previously (or more precisely to accentuate the natural Haas effect which already benefits the system 2 of restitution close to the stage).

By way of illustration, FIG. 6 represents the variations in the propagation time of the signals broadcast by omnidirectional loudspeakers (the propagation time being independent of the vertical directivity of the loudspeakers), as a function of the distance to the scene, for a set of loudspeakers positioned at the stage level (solid line curve) and 6.0 m above the stage (dashed line curve).

This figure illustrates the natural Haas effect, described above, which benefits the system 2 of restitution near the scene (delay offset near the scene).

To maintain this effect over a distance to the larger scene, and allow a good vertical location of the sound objects on the first set of speakers, we can, in the light of Figure 6, try to "raise" the curve in broken lines so to move it away from the curve in solid lines, delaying the second signals with respect to the first signals at the reference position.

The gains Gl and G2 are chosen in order to ensure a sufficient sound level at the back of the room, while optimizing the vertical location of the sound objects restored to the reference position.

As described above, the delays and the gains (T1, G1) and (T2, G2) making it possible to obtain such a desired sound reproduction effect at the reference position Pref (E1, E2), are chosen (ie adjusted) at the during the E_CALIB configuration step. This choice or adjustment can be made manually (for example experimentally, using sound level measurements) or automatically via numerical simulations.

Thus, in the first embodiment described here, the configuration step E_CALIB comprises a first phase during which physical measurements are made at the reference position Pref (E1, E2), on the one hand, of the level sound "naturally" restored by the set El (ie this amounts to applying a gain τΐ zero, a delay Gl = 1, and a gain G2 zero), and secondly, the sound level "naturally" restored by the set E2 (ie this amounts to applying a zero gain T2, a delay G2 = 1, and a zero gain G1).

From these physical measurements, the delays τΐ and τ2, and the gains G1 and G2, to be applied to the input audio signal S are determined and adjusted empirically in a second phase to obtain the desired sound reproduction effect. These delays τΐ and τ2, and these gains Gl and G2 can be the subject of several tests, and be adjusted to each test, by trial and error.

As a variant, this second adjustment phase can be performed by simulation.

The step E_CALIB may further comprise an ultimate phase of more precise adjustment of the delays and gains determined during the second phase, by using known signals, via new physical measurements and / or by numerical simulation.

It will be noted that during the step E_CALIB, the setting of delays and gains can be made either globally or independently on each frequency band of a set of predefined frequency bands. In this case, a gain and a delay are determined per frequency band studied. This advantageously makes it possible to take into account the frequency dependence of the directivity of the loudspeakers of the second set E2.

FIG. 7 illustrates various advantages provided by the set 1 of reproduction, in terms of variations of the sound level as a function of the distance of the listener with respect to the scene. In the example of FIG. 7, the loudspeakers of the set E2 are directional loudspeakers vertically having a vertical opening of 90 °.

More specifically, FIG. 7 shows that, thanks to the use of directional loudspeakers for the set E2, the sound level of the signal restored by the set E2 perceived near the scene is less important (see curve in lines discontinued).

In addition, it ensures an almost uniform sound level to the bottom of the room. In addition, the difference in sound levels between the signals restored by the set E1 (represented by the curve in solid line) and the signals restored by the set E2 is almost constant for distances to the scene greater than 8 m (setting in coherence of the sound levels reached by the invention), so that it is possible by appropriate choice delays τΐ and τ2 to benefit from a Haas effect to the bottom of the room.

In the example illustrated in FIG. 2, the reproduction unit 1 comprises two sets of speakers E1 and E2.

In a variant, the reproduction unit 1 further comprises another set of speakers of the subwoofer or subwoofer type, to add an additional effect during the sound reproduction of the input audio signal.

Similarly, in this variant, a plurality of audio signals for supplying this set of subwoofer type loudspeakers is generated by the generation device 3 by applying to the input audio signal:

At least one delay and / or gain; and

- Low pass type filtering.

The delay and / or the gain thus applied are chosen so that the combination of the signals reproduced by the set of speakers of the subwoofer type and the signals restored by one of the sets E1 and E2 verify a restitution effect. sound determined at a reference spatial position defined for this set and for the set of speakers subwoofers type.

We will now illustrate the implementation of this first embodiment by describing three examples.

Example 1

In example 1:

The set El is a row of loudspeakers HP1-1, HP1-P located on the scene SCENE, at the height h1;

The set E2 is a line of N loudspeakers HP2-1,..., HP2-N situated above the scene SCENE, at the height h2. The number N of speakers and their spacing depend on the width of the scene: N typically varies between 4 and 12 speakers, spaced from 80 cm to 5 m.

The number P of loudspeakers of the set E1 is greater than or equal to the number N of loudspeakers of the set E2.

In accordance with the invention, the speakers HP2-1, ..., HP2-N are on average more vertically oriented than the speakers HP1-1, HP1-P of the set El. More precisely, in this example 1:

The loudspeakers HP1-1, HP1-P of the set El have a vertical opening varying from 80 ° to 120 °, and are oriented towards the front of the stage (ie towards the front of the audience ); The N loudspeakers HP2-1, ..., ΗΡ2-Ν of the set E2 have a vertical opening varying from 60 ° to 90 °, and are oriented towards the bottom of the room (part of the audience the further away from the scene).

The modules 3C1 and 3C2 use techniques T1 and T2 of sound spatialization of the WFS type enabling the creation of virtual sources at the positions of the sound objects of the signal S identified by the module 3A. These positions are provided in Example 1 to Module 3A via a graphical user interface.

The modules 3B1 and 3B2 are configured so as to apply a non-zero delay to the signal S and to attenuate it (gains G1, G2 less than 1). The gains and delays (T1, G1) and (T2, G2) respectively applied by the modules 3B1 and 3B2 are chosen so that the signals restored by the sets E1 and E2 at the reference position Pref (E1, E2) arrive with a similar sound level (within plus or minus 3 decibels), and simultaneously or almost simultaneously (with a tolerance of plus or minus 5 ms). Example 2

In example 2:

The set E2 comprises N loudspeakers HP2-1,..., HP2-N of type "line array" in English (line source) located above the SCENE scene, at the height h2, and distributed on the along the stage. For example for N = 3, we have a loudspeaker on both sides of the stage and one in the center of the stage. Such loudspeakers are in a known manner particularly powerful and have a significant vertical directivity and thus a small opening.

The number P of loudspeakers of the set El and the spacing between them depend on the width of the scene, and the distance separating the loudspeakers of the first set El from the loudspeakers of the second set E2: P varies typically between 4 and 30 speakers, spaced from 50 cm to 3 m.

According to the invention, the loudspeakers HP2-1,..., HP2-N are on average more vertically oriented than the loudspeakers HP1-1, HP1-P of the set El. More precisely, in this case first example:

The loudspeakers HP1-1, HP1-P of the set El have a vertical opening varying from 80 ° to 120 °, and are oriented towards the front of the stage (ie towards the front of the audience );

The N loudspeakers HP2-1, ..., HP2-N of the set E2 have a vertical opening varying from 10 ° to 60 °, and are oriented towards the back of the room (part of the audience the further away from the scene).

The 3C1 module uses a WFS type sound spatialization technique allowing the creation of virtual sources at the positions of the sources identified in the signal S by module 3A. These positions are provided in example 2 to module 3A via a mixer operating on the signal S.

The module 3C2 uses a dynamic matrix mixing sound spatialization T2 technique in the form of a stereophonic intensity panning as a function of the position of the sound objects of the signal S identified by the module 3A.

The modules 3B1 and 3B2 are configured so as to apply a non-zero delay to the signal S and to attenuate it (gains G1, G2 less than 1). The gains and delays (T1, G1) and (T2, G2) respectively applied by the modules 3B1 and 3B2 are chosen so that the signals restored by the sets E1 and E2 at the reference position Pref (E1, E2) arrive with a similar sound level (within plus or minus 3 decibels), and simultaneously or almost simultaneously (with a tolerance of plus or minus 5 ms).

Example 3

In this example 3:

The set E comprises P loudspeakers HP1-1, HP1-P situated on the SCENE stage, at the height h1;

The set E2 comprises N loudspeakers HP2-1,..., HP2-N of the "line array" type, located above the SCENE stage, at the height h2. The number N of loudspeakers and their spacing depend on the width of the scene: N typically varies between 4 and 12 loudspeakers, spaced from 80 cm to 5 m.

The playback assembly 1 further comprises in this example a set designated by Elow of L speakers HPIow-1, HPlow-L low-frequency (subwoofers) arranged on or below the scene SCENE.

The loudspeakers HP1-1, HP1-P of the set El have a vertical dispersion varying from 80 ° to 120 °, and are oriented towards the front of the stage (ie towards the front of the audience );

The N loudspeakers HP2-1,..., HP2-N of the set E2 have a vertical dispersion ranging from 10 ° to 60 °, and are oriented toward the back of the room (part of the audience the further away from the scene).

The module 3C1 uses a sound spatialization technique T1 of the WFS type enabling the creation of virtual sources at the positions of the sound objects identified in the signal S by the module 3A. The module 3C2 uses a technique T2 of sound spatialization of WFS type also allowing the creation of virtual sources at the positions of the sound objects identified in the signal S by the module 3A. The T2 technique is for example a WFS technique as described in the document E. Corteel cited above, so as to compensate for the horizontal directivity characteristics of the line array loudspeakers of the E2 set.

The positions of the sound objects of the signal S are identified directly here by the module 3A by means of a device for tracking the positions of the sound objects present on the SCENE stage (eg instruments or characters, etc.).

The generation device 3 further comprises, according to this example, a 3Blow module adapted to apply a gain and a delay to the input audio signal S, and a 3Clow module capable of applying a low-pass filtering to the signal from the 3Blow module for generate signals to power the Elow speakers.

The modules 3B1, 3B2 and 3Blow are configured so as to apply a non-zero delay to the signal S and attenuate it (gains lower than 1).

The gains and delays (T1, G1) and (T2, G2) respectively applied by the modules 3B1 and 3B2 are chosen so that the signals restored by the set E2 arrive with an offset of 10 or 20 ms with respect to the restored signals. by the set E1 at the reference position Pref (E1, E2), with a similar sound level (within plus or minus 3 decibels), so as to benefit from the Haas effect described in the document by Blauert et al. cited previously.

The delay and gain applied by the 3Blow module are chosen so that the signals restored by the Elow and El sets at a reference spatial position Pref (Elow, El) defined for the Elow and El sets (similar to the position Pref (El, E2)) arrive with a similar sound level (within plus or minus 3 decibels) in their respective frequency bands, and simultaneously or almost simultaneously (with a tolerance of plus or minus 5 ms).

We will now describe, with reference to FIGS. 8 to 10, a second embodiment of the invention in which, for the reproduction of the input audio signal S, a sound reproduction assembly comprising three sets of loudspeakers is considered. El, E2 and E3, the sets E2 and E3 acting as sets of booster speakers.

Thus, FIG. 8 represents, in its environment, a set 1 of sound reproduction, according to the invention, in this second embodiment.

The elements similar to the first embodiment described with reference to FIGS. 2 to 5 are referenced identically in FIGS. 8 to 10.

According to the second embodiment, the set of sound reproduction comprises:

A sound reproduction system 2 'comprising a plurality of loudspeakers; and A device 3 'for generating audio signals intended to supply the loudspeakers of the sound reproduction system 2'.

The sound reproduction system 2 'differs from the sound reproduction system 2 illustrated in FIGS. 2 and 3 in that it also comprises a third set E3 of K loudspeakers, denoted HP3-1, HP3-2, .. ., HP3-K, where K is a positive integer greater than 1.

As illustrated in FIG. 9, the loudspeakers HP3-1, HP3-2,..., HP3-K are positioned at a height h3 greater than the reference height h1, in a vertical plane located closer to the bottom of the room as the vertical plane in which are positioned the speakers of the set E2.

The loudspeakers of the set E3 can be omnidirectional chosen, if the set E3 is intended to compensate for the presence of masking elements such as balconies for example, or on the contrary they can be chosen more directive on average vertically than the El set loudspeakers, and be oriented, similar to the speakers of the set E2 to the bottom of the room.

The number K of loudspeakers in the third set E3 may be equal to or distinct from the numbers P and N of loudspeakers in the sets E1 and E2.

The various loudspeakers of the set E3 are powered by multichannel audio signals, denoted respectively S3-1, S3-2,..., S3-K generated by the device 3 'of generation from the audio signal of entrance S.

To generate these signals, the device 3 'of generation is based on:

On the software module 3A described above, and able to identify the POS positions (OBJ) of the various sound objects OBJ included in the input signal S;

The module 3B3 adapted to apply on the signal S, a delay τ3 and a gain G3, identical here on each channel of the signal S;

The module 3C3 capable of applying to the signal from the module 3B3 a sound spatialization technique T3.

The device 3 'of generation here has the hardware architecture of a computer, similar to that of the generation device 3 described above with reference to FIG. 4. The means of communication of the device 3' of generation are however distinguished from the means communication device of the generation device 3 further comprising an interface (wired or wireless) with the speakers HP3-1, HP3-2, ..., HP3- of the assembly E3.

The ROM of the device 3 'of generation constitutes a recording medium according to the invention, readable by the (micro) processor of the device 3' and on which is recorded a computer program according to the invention, comprising instructions for performing the steps of a generation method now described with reference to FIG.

FIG. 10 illustrates the main steps of a generation method according to the invention, as implemented, in the second embodiment, by the device 3 'for generating on reception of the input audio signal S , to generate and provide the signals S1-1, S1-2, ..., S1-P, S2-1, S2-2, S2-N and S3-1, S3-2, S3-K to the sets E1, E2 and E3 of high- speakers respectively.

The steps similar to the steps implemented by the generation device 3 and illustrated in FIG. 5 are referenced identically and are not described again in detail here.

It is therefore assumed that the generation device 3 'receives the multichannel input audio signal S via its communication means (step E10).

It is furthermore assumed that the device 3 'of generation, via a suitable interface, the positions designated by POS (OBJ) of the sound objects OBJ included in the signal S and which it is desired to restore ( step E20).

According to the invention, different processes are applied by the generating device 3 'to the signal S to generate the signals intended to supply the three sets E1, E2 and E3 of loudspeakers.

These treatments are applied independently for each set of loudspeakers, via three distinct signal processing chains comprising for the first, the module 3B1 and the module 3C1, for the second, the module 3B2 and the module 3C2, and for the third , module 3B3 and module 3C3. These three treatment chains preferentially operate simultaneously.

More specifically, the modules 3B1, 3C1, 3B2 and 3C2 proceed in a manner similar to the first embodiment to generate the signals Sl-1, Sl-2,..., Sl-P intended to feed the set E1 and the signals S2-1, S2-2, ..., S2-N for supplying the set E2 (steps E30 to E80 illustrated in Figure 5).

Moreover, in parallel, the module 3B3 applies to the signal S the delay τ3 and the gain G3 mentioned above (step E90), thereby generating a signal S3. This delay τ3 and this gain G3 have been determined and supplied to the module 3B3 during a preliminary configuration step E_CALIB 'described later.

The module 3C3 then applies to the signal S3 supplied at the output of the module, the sound spatialization technique T3 (step E100), thereby generating the plurality of signals S3-1, S3-2, ..., S3-K.

The technique T3 is for example a holophonic technique of the WFS type as described in document EP 2 206 365. It takes into account firstly the POS positions (OBJ) of the sound objects identified and provided by the module 3A, and on the other hand, the position of the HP3-1, HP3-2, ..., HP3-K speakers in the E3 set, to create virtual sources from the HP3-1, HP3-2 loudspeakers , ..., HP3-K from the E3 set to the POS (OBJ) positions.

Of course, this hypothesis is not limiting and other sound spatialization techniques can be envisaged within the scope of the invention, such as dynamic matrix mixing techniques mentioned above.

The signals S3-1, S3-2, ..., S3-K generated by the module 3C3 are supplied to the speakers HP3-1, HP3-2, ..., HP3-K of the set E3 (step E110), to be restored. In this second embodiment, the delays τΐ, τ2 and τ3, and the gains Gl, G2 and G3 are chosen by considering the sets of speakers El, E2 and E3 two by two. More precisely, they are chosen so as to verify:

On the one hand, that the combination of the signals restored by the set E1 and the signals restored by the set E2 satisfies a determined sound restitution effect at the reference spatial position Pref (E1, E2) defined for the sets E1; and E2; and

On the other hand, that the combination of the signals restored by the set E2 and the signals restored by the set E3 satisfies a determined sound reproduction effect at a reference spatial position Pref (E2, E3) defined for the sets E2 and E3. This spatial position Pref (E2, E3) can be defined according to criteria similar to those adopted to define the spatial position Pref (E1, E2) described above.

As a variant, the combination of the signals restored by the sets E1 and E3 may be considered instead of the sets E2 and E3.

Thus, the preliminary configuration step E_CALIB 'takes place in two stages, during which we consider the sets E1 and E2 respectively, and the sets E2 and E3 on the other hand. A procedure similar to that previously described for the configuration step E_CALIB, based on physical measurements of sound level and / or numerical simulations, is applied to each pair of sets of speakers considered (ie to the pair (E1 , E2) taking into account the reference position Pref (E1, E2), then to the pair (E2, E3) taking into account the reference position Pref (E2, E3)).

The final phase of adjustment can, on the other hand, take into account the signals restored by the three sets in order to be more precise in the estimation of the delays τΐ, τ2 and τ3 and of the gains Gl, G2, and G3.

As for the step E_CALIB, the setting of delays and gains can be made either globally or independently on each frequency band of a set of predefined frequency bands. In this case, a gain and a delay are determined per frequency band studied. This advantageously makes it possible to take into account the frequency dependence of the directivity of the loudspeakers of the second set E2 and, if appropriate, of the third set E3.

It should be noted that in the second embodiment described here, the restitution unit comprises three sets E1, E2 and E3, two sets of which act as a set of return speakers for room R. However, the invention is not limited to two sets of booster speakers, and it is possible, within the context of the invention, to consider adding other sets E4, E5, ... of loudspeakers. recall in the room. The delays and / or the gains applied by the device 3 'of generation to generate the signals intended to supply these sets will be chosen in a manner similar to the delay and the gain applied by the module 3B3, that is to say, considering the sets of speakers of the reproduction system in pairs (eg El with E2, then E2 with E3, then E3 with E4, etc.). We will now illustrate the implementation of this second embodiment by the description of Example 4.

Example 4

In this example 4:

The set E2 comprises N loudspeakers HP2-1,..., HP2-N situated above the scene SCENE, at the height h2. The number N of loudspeakers and their spacing depend on the width of the scene: N typically varies between 4 and 12 loudspeakers, spaced from 80 cm to 5 m;

- The E3 set includes K speakers HP3-1, ..., HP3-K arranged in the room, above the audience, in the middle of the AUDIT part. The number K of loudspeakers and their spacing depend on the width of the scene: K typically varies between 4 and 12 loudspeakers, spaced from 80 cm to 5 m.

In this example 4, the speakers HP2-1, ..., HP2-N are on average more directional vertically than the speakers HP1-1, HP1-P of the set El. More precisely, in this example 4:

The loudspeakers HP1-1, HP1-P of the set El have a vertical opening varying from 80 ° to 120 °, and are oriented towards the front of the stage (ie towards the front of the audience ); The N loudspeakers HP2-1, ..., HP2-N of the set E2 have a vertical opening varying from 60 ° to 90 °, and are oriented towards the back of the room (part of the audience the further away from the scene).

Moreover, in this example 4, the K loudspeakers HP3-1,..., HP3-K of the set E3 are also on average more directional vertically than the loudspeakers HP1-1, HP1-P of the El set. They have a vertical opening varying from 60 ° to 90 °, and are also directed towards the back of the room (part of the audience furthest away from the stage).

The modules 3C1, 3C2 and 3C3 use techniques T1, T2 and T3 for sound spatialization of the WFS type enabling the creation of virtual sources at the positions of the objects identified in the signal S by the module 3A. These positions are provided in Example 4 to Module 3A via a graphical user interface.

The modules 3B1, 3B2 and 3B3 are configured so as to apply a non-zero delay to the signal S and attenuate it (gains G1, G2 and G3 less than 1).

The gains and delays (T1, G1) and (T2, G2) respectively applied by the modules 3B1 and 3B2 are chosen so that the signals restored by the sets E1 and E2 at the reference position Pref (E1, E2) arrive with a similar sound level (within plus or minus 3 decibels), and simultaneously or almost simultaneously (with a tolerance of plus or minus 5 ms).

The delay and the gain (T3, G3) applied by the module 3B3 are chosen so that the signals restored by the sets E2 and E3 at the reference position Pref (E2, E3) arrive with a similar sound level (within plus or minus 3 decibels) in their respective frequency bands, and simultaneously or almost simultaneously (with a tolerance of plus or minus 5 ms). The position Pref (E2, E3) is chosen farther from the SCENE scene than the position Pref (E1, E2), towards the back of the room.

In the two embodiments described here, the gain / delay application and the application of the sound spatialization technique on the input audio signal are implemented in separate steps by separate modules.

In another embodiment, these two processes can be performed simultaneously by the same module (i.e. by the modules 3C1, 3C2 and optionally 3C3 respectively), during the application of the sound spatialization technique.

Claims

A method of generating, from an input audio signal (S), audio signals to be supplied to a sound reproduction system (2,2 '), said sound reproduction system comprising a first set ( El) speakers (HP1_1, ..., HP1_P) positioned at a reference height (hl), and a second set (E2) of speakers (HP2_1, ..., HP2_N) positioned at a height ( h2) greater than the reference height (hl), the loudspeakers of the second set being more vertically oriented than the loudspeakers of the first set, this method comprising:

A step (E20) for identifying a position (POS (OBJ)) of at least one sound object of the input audio signal (S);

A step of generating (E30, E40) a plurality of first audio signals (S1_1, ..., S1_P) intended to feed the first set (El) of loudspeakers, by applying to the input audio signal ( S):

o at least one delay (τΐ) and / or a gain (Gl); and

a sound spatialization technique (Tl) taking into account the identified position of said at least one sound object of the input audio signal and a position of the speakers of the first set;

A step of generating (E60, E70) a plurality of second audio signals (S2_1, ..., S2_N) for supplying the second set (E2) of loudspeakers, by applying to the input audio signal ( S):

at least one delay (τ2) and / or a gain (G2); and

a sound spatialization technique (T2) taking into account the identified position of said at least one sound object of the input audio signal and a position of the speakers of the second set;

at least one said sound spatialization technique applied during the generation of the first and second audio signals being a holophonic technique; and

the delays and / or gains applied during the generation of the first and second audio signals being chosen so that the combination of the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers; speakers verifies a determined sound reproduction effect at at least one reference spatial position (Pref (E1, E2)) defined for the first and second sets of speakers.

2. The method of claim 1 wherein the delays applied during the generation of the first and second audio signals are chosen so that the first signals output by the first set of speakers and the second signals output by the second set of speakers arrive simultaneously at the reference spatial position.

3. The method of claim 1 wherein the delays applied during the generation of the first and second audio signals are chosen so that the second signals output by the second set of speakers arrive with a predetermined offset to the spatial position of reference with respect to the first signals restored by the first set.

4. Method according to any one of claims 1 to 3 wherein at least one said sound spatialization technique applied during the generation of the first and second audio signals is:

A holophonic sound spatialization technique of the WFS type; or

- a dynamic matrix mixing technique.

5. Method according to any one of claims 1 to 4 wherein, during at least one said generation step, said at least one delay and / or gain is applied to the input signal before the spatialization technique. sound.

The method of any one of claims 1 to 5 wherein the input audio signal is a multichannel audio signal and, in at least one said generation step, the same delay and / or gain is applied on each channel of the input audio signal.

The method of any one of claims 1 to 5 wherein the input audio signal is a multichannel audio signal and, in at least one said generation step, a delay and / or distinct gain is applied. on at least two separate channels of the input audio signal.

8. Method according to any one of claims 1 to 7 wherein the sound reproduction system further comprises at least a third set (E3) of speakers (HP3-1, HP3-2, ..., HP3- K) positioned at a height (h3) greater than the reference height (h1), said method comprising a step (E90, E100) of generating a plurality of third audio signals (S3-1, S3-2, .. ., S3-K) for supplying the third set of loudspeakers by applying to the input audio signal (S):

At least one delay (τ3) and / or a gain (G3); and

A sound spatialization technique (T3) taking into account the identified position of said at least one sound object of the input audio signal and a position of the loudspeakers of the third set;

said at least one delay and / or gain being selected such that a combination of the third signals output by the third set of speakers and signals output by another (E2) of said speaker sets of the playback system check a refund effect sound determined at a reference spatial position defined for this other set and for the third set of speakers.

A computer program comprising instructions for performing the steps of the generation method according to any one of claims 1 to 8 when said program is executed by a computer.

A computer-readable recording medium on which a computer program is recorded including instructions for performing the steps of the generation method according to any one of claims 1 to 8.

11. Device (3,3 ') for generating audio signals intended for a sound reproduction system (2,2') from an input audio signal, this sound reproduction system comprising a first set (E1) speakers positioned at a reference height (h1), and a second set (E2) of speakers positioned at a higher height (h2) at the reference height, the speakers of the second set being more vertically oriented than the speakers of the first set,

the generating device (3,3 ') comprising:

Means for identifying (3A) a position of at least one sound object of the input audio signal;

Means for generating (3B1) a plurality of first audio signals intended to supply the first set of loudspeakers, able to apply to the input audio signal:

o at least one delay and / or gain; and

Generation means (3B2) of a plurality of second audio signals intended to supply the second set of loudspeakers, able to apply to the input audio signal:

o at least one delay and / or gain; and

at least one said sound spatialization technique applied by the means for generating the first and second audio signals being a holophonic technique; and

the gains and / or delays applied by the means for generating the first and second audio signals being chosen so that the combination of the first signals output by the first set of loudspeakers and the second signals output by the second set of loudspeakers verifies a determined sound reproduction effect at at least one reference spatial position defined for the first and second sets of loudspeakers.

12. Set (Ι,) for rendering an input audio signal comprising: - a sound reproduction system (2,2 ') comprising a first set (El) of loudspeakers positioned at a reference height, and a second set (E2) of speakers positioned at a height greater than the reference height, the speakers of the second set being more vertically oriented than the speakers of the first set; and

- A device (3,3 ') for generating audio signals according to claim 11 for supplying the speakers of the first set and the speakers of the second set of the sound reproduction system.