EP2304721B1 - Spatial synthesis of multichannel audio signals - Google Patents

Abstract

A method and associated device are provided for spatial synthesis of a sum signal to obtain at least two output signals, the sum signal as well as the spatialization parameters being output from a parametric coding by matrixing of an original multi-channel signal. The method comprises: decorrelation of the sum signal to obtain a decorrelated signal; applying a synthesis matrix, whose coefficients depend on the spatialization parameters, to the decorrelated signal and to the sum signal to obtain said output signals, wherein for at least one range of value of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion of minimizing a quantitative function, relating to the quantity of decorrelated signal in each of the output signals obtained by applying the synthesis matrix.

Description

The present invention relates to the field of coding / decoding multichannel digital audio signals.

More particularly, the present invention relates to the parametric encoding / decoding of multichannel audio signals.

This type of coding / decoding is based on the extraction of spatialization parameters so that at decoding, the spatial perception of the listener can be reconstituted.

Such a coding technique is known under the name of "Binaural Cue Coding" in English (BCC) which aims on the one hand to extract and then code the indices of auditory spatialization and on the other hand to code a monophonic or stereophonic signal from a mastering of the original multi-channel signal.

This parametric approach is a low rate coding. The main advantage of this coding approach is to allow a better compression rate than conventional multi-channel digital audio compression methods while ensuring the backward compatibility of the compressed format obtained with the existing coding formats and broadcasting systems.

Thus, the invention relates more particularly to the spatial decoding of a sound scene 3 D from a reduced number of transmitted channels.

The MPEG Surround standard described in the MPEG document ISO / IEC 23003-1: 2007 and in the document " Breebaart, J. and Hotho, G. and Koppens, J. and Schuijers, E. and Oomen, W. and van de Par, S., "entitled" Background, concept, and architecture for the recent MPEG surround standard on multichannel audio Compression "in Journal of the Audio Engineering Society 55-5 (2007) 331-351 , describes a specific coding / decoding structure of the multi-channel audio signal.

The figure 1 describes such an encoding / decoding system in which the encoder 100 constructs a sum signal ("downmix") S _s by matrixing at 110 the channels of the original multichannel signal S and delivers via an extraction module parameters 120, a reduced set of parameters P which characterize the spatial content of the original multi-channel signal.

At the decoder 150, the multichannel signal is reconstructed (S ') by a synthesis module 160 which takes into account both the sum signal and the transmitted parameters P.

The sum signal has a reduced number of channels. These channels can be encoded by a conventional audio encoder before transmission or storage. Typically, the sum signal has two channels and is compatible with conventional stereo broadcasting. Before transmission or storage, this sum signal can thus be encoded by any conventional stereo encoder. The signal thus coded is then compatible with the devices comprising the corresponding decoder which reconstruct the sum signal while ignoring the spatial data.

• les blocs TTO (pour "Two To One" en anglais) qui permettent d'extraire les paramètres spatiaux entre deux canaux et de construire un signal somme monophonique à partir de ces deux canaux,
• les blocs TTT (pour "Three To Two" en anglais) qui permettent d'extraire les paramètres spatiaux entre trois canaux et de construire un signal somme contenant deux canaux à partir de ces trois canaux.

The MPEG Surround standard has adopted a specific structure for the representation of spatial data: the encoder relies on a coding tree structure constructed from a reduced number of elementary coding blocks, each of which makes it possible to extract spatial parameters on a single space. reduced number of channels. There are two basic block types of coding:

• the TTO blocks (for "Two To One" in English) which make it possible to extract the spatial parameters between two channels and to construct a monophonic sum signal from these two channels,
• the TTT blocks (for "Three To Two" in English) which make it possible to extract the spatial parameters between three channels and to construct a sum signal containing two channels from these three channels.

The figure 2 illustrates a first example of a coding structure or code tree using TTO blocks (TTO ₀ , TTO ₁ , TTO ₂ , TTO ₃ and TTO ₄ ) to obtain a monophonic signal S from a multi-channel signal 5.1 with 6 channels (L, R, C, LFE, Ls and Rs).

The figure 3 illustrates a second example of an encoding structure using both TTO blocks and TTT blocks to obtain a stereo signal Sl and Sr from the 5.1 signal.

The decoding of the monophonic or stereophonic signals thus received is effected by using a decoding tree symmetrical to those represented in FIGS. Figures 2 and 3 .

Thus, for the decoding of a signal encoded according to the tree of the figure 2 , the decoding can be seen as a succession of reconstruction step.

In this case, the first decoding step consists in reconstructing the signals corresponding to the input signals of the block TTO ₀ from the sum signal S and the spatial parameters extracted by the block TTO ₀ , the next step then consists in reconstructing the signals corresponding to the input signals of the block TTO ₁ from the signal reconstructed in the previous step and the spatial parameters extracted by the block TTO ₁ , the decoding then proceeds in a similar manner until the reconstruction of all the channels encoded multi-channel signal. In practice, the decoder constructs a matrix making it possible to go directly from the monophonic sum signal to the 6 reconstructed channels by combining the smaller size matrices of the different blocks TTO and TTT.

The technique used in the MPEG Surround standard for decoding TTO blocks, however, imposes a very disadvantageous limitation for the coding of multichannel signals comprising channels in phase opposition.

This decoding technique is more precisely described in the patent application entitled "signal synthesizing" published under the number WO 03/090206 A1 October 30, 2003 (Applicant: Koninklijke Philips Electronics NV, Inventor: Dirk J. Breebaart).

This technique consists, as shown with reference to the figure 4 , to perform a decorrelation step at 410 by filtering the sum signal to obtain a decorrelated signal d. The sum signal and the decorrelated signal thus obtained are then processed by a synthesis module 420 via a synthesis matrix M, as a function of the spatial parameters R and I to create the two signals I and R respecting the specified spatial parameters. The parameters R and I are respectively the energy ratio between the channels of the multi-channel signal and an inter-channel correlation index of the multi-channel signal channels.

avec ${\lambda }_1=\sqrt{\frac{R}{1+R}},$

${\mathit{\lambda }}_2=\sqrt{\frac{1}{1+R}},$

$\alpha =\frac{1}{2}\arccos \left(I\right)$

et $\beta =\arctan \left(\frac{{\lambda }_2-{\lambda }_1}{{\lambda }_2+{\lambda }_1}\tan \left(\alpha \right)\right)\mathrm{.}$

The matriculation of the signals s and d is done according to the following relations: $$\left[\begin{array}{c}l\\ r\end{array}\right]=\left[\begin{array}{ll}{\mathrm{<figure-callout\; id="1"\; label="\lambda "\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_1\cos \left(\alpha +\mathrm{<figure-callout\; id="1"\; label="\beta \; \lambda "\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">\beta \; </figure-callout>}\right)& {\lambda }_{}{}_1\sin \left(\alpha +\mathrm{<figure-callout\; id="2"\; label="\beta \; \lambda "\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">\beta \; </figure-callout>}\right)& \\ {\lambda }_{}{}_2\cos \left(-\alpha +\mathrm{<figure-callout\; id="2"\; label="\beta \; \lambda "\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">\beta \; </figure-callout>}\right)& {\lambda }_{}{}_2\sin \left(-\alpha +\beta \right)\end{array}\right]\left[\begin{array}{c}s\\ d\end{array}\right]$$

with ${\mathrm{<figure-callout\; id="1"\; label="\lambda "\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_1=\sqrt{\frac{R}{1+R}},$

${\mathit{<figure-callout\; id="2"\; label="\lambda "\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_2=\sqrt{\frac{1}{1+R}},$

$\alpha =\frac{1}{2}\arccos \left(I\right)$

and $\beta =\arctan \left(\frac{{\lambda }_2-{\mathrm{<figure-callout\; id="1"\; label="\lambda "\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_1}{{\mathrm{<figure-callout\; id="2"\; label="\lambda "\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_2+{\mathrm{<figure-callout\; id="1"\; label="\lambda "\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">\lambda </figure-callout>}}_1}\tan \left(\alpha \right)\right)\mathrm{.}$

However, this matrixing has the limitation mentioned above and makes this method unsuitable for coding multichannel audio signals with negative inter-channel correlations.

In particular, such a technique is not suitable for decoding the surround signals which comprise phase oppositions between channels.

Indeed, when the inter-channel correlation I is negative, and in particular when it is close to -1, the proportion of decorrelated signal used to synthesize the signals I and R becomes very important, in some cases clearly exceeding the quantity of the signal. signal sum used. In the most problematic case, it can be seen that for an interchannel level difference of 0 dB, ie for R = 1, when the interchannel correlation I tends to -1, the mixing matrix tends to the following matrix: $$\left[\begin{array}{cc}0& \frac{\sqrt{2}}{2}\\ 0& -\frac{\sqrt{2}}{2}\end{array}\right]\mathrm{.}$$

et $r=-\frac{\sqrt{2}}{2}d$

qui ne font pas intervenir le signal somme dans leur expression, mais utilisent uniquement le signal décorrélé. Ainsi, la forme d'onde du signal reconstruit n'est pas contrôlée puisqu'elle dépend totalement de la décorrélation subie par le signal s.This matrix corresponds to reconstructed signals $l=\frac{\sqrt{2}}{2}d$

and $r=-\frac{\sqrt{2}}{2}d$

which do not involve the sum signal in their expression, but only use the decorrelated signal. Thus, the waveform of the reconstructed signal is not controlled since it totally depends on the decorrelation experienced by the signal s.

The reconstruction problem illustrated in the previous example in an extreme case is also present for other values of R and I, and is even more so. marked that I is close to -1. Thus, the waveform of the reconstructed channels is in these cases not as close as it could be the original signals, which unnecessarily limits the quality of the reconstructed signals.

The effect of this limitation is even more pronounced when the signal has several channels with interchanal correlations close to -1. In this case, more than two channels have near waveforms, but some of them are out of phase.

During the rendering of the original multi-channel signal, the signals of these different channels which have similar waveforms will interact in the rendering zone by creating constructive and destructive interferences which will make it possible to reconstruct the desired sound field.

After decoding, the waveform of the channels will be strongly deformed due to the problem mentioned above.

In addition, since each TTO block decoder involved in the decoding tree uses a different decorrelation filter, the deformation of the waveform will not be the same for the different channels.

The reconstructed channels then no longer have, as in the original signal, near waveforms and the interferences that allowed the reconstruction of the sound field during the restitution, are then no longer as in the original signal. This leads, on the one hand, to a bad spatial reconstruction of the sound stage, and on the other hand to the creation of audible artifacts, the differences in waveforms leading to the creation of perceptible noisy components.

The present invention improves the situation.

• décorrélation du signal somme pour obtenir un signal décorrélé;
• application d'une matrice de synthèse dont les coefficients dépendent des paramètres de spatialisation, au signal décorrélé et au signal somme pour obtenir lesdits signaux de sortie,

caractérisé en ce que pour au moins une plage de valeur d'au moins un paramètre de spatialisation, les coefficients de la matrice de synthèse sont déterminés selon un critère de minimisation d'une fonction quantitative (q), relative à la quantité de signal décorrélé dans chacun des signaux de sortie obtenus par l'étape d'application de la matrice de synthèse, la fonction quantitative étant telle que l'augmentation en valeur absolue des coefficients de la matrice de synthèse appliqués au signal décorrélé fait augmenter la valeur de ladite fonction appliquée à ces mêmes coefficients.For this purpose, the present invention proposes a method of spatial synthesis of a sum signal to obtain at least two output signals, the sum signal as well as spatialization parameters resulting from parametric encoding by mastering a multi signal. - original channel. The method comprises the steps of:

• decorrelation of the sum signal to obtain a decorrelated signal;
• applying a synthesis matrix whose coefficients depend on the spatialization parameters, the decorrelated signal and the sum signal to obtain said output signals,

characterized in that for at least one value range of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion of minimization of a quantitative function (q), relative to the quantity of decorrelated signal in each of the output signals obtained by the step of applying the synthesis matrix, the quantitative function being such that the increase in absolute value of the coefficients of the synthesis matrix applied to the decorrelated signal increases the value of said function applied to these same coefficients.

Thus, taking into account the amount of decorrelated signal in each of the signals and therefore in the signal synthesis step, makes it possible to dispense with the case mentioned above where only the decorrelated signal is involved in the synthesis matrix. The method according to the invention thus makes it possible to handle the cases where a spatialization parameter situated in a predetermined value range causes such a situation.

Minimizing such a quantitative function makes it possible to define coefficients of the synthesis matrix that make it possible to ensure good compliance with the waveform of the input signal in the output signals.

More particularly and simply, such a quantitative function may be a decorrelated signal energy function.

This function meets the characteristics mentioned above.

avec p entier supérieur ou égal à 1.In a more general way, the quantitative function is of the type: $q\left(x,\mathrm{there}\right)={\left({\left|x\right|}^p+{\left|\mathrm{there}\right|}^p\right)}^{\frac{1}{p}}$

with p integer greater than or equal to 1.

In a particular embodiment, the spatialization parameters are a parameter (R) of energy ratio between the channels of the multi-channel signal and a interchannel correlation parameter (I) of the multi-channel signal, a range of values being the range in which the intcrcanal correlation parameter is negative.

Thus, the invention applies more particularly to multi-channel signals having negative interchannel correlations.

It can therefore be implemented only for the values of the negative inter-channel correlation parameter or for any value of this parameter.

In another embodiment, a different quantitative function is chosen by value range of the spatialization parameters.

It is then possible to modulate the relative importance that we want to give to the different synthesis matrices. It is thus possible to give a significant weight to a matrix as defined in the state of the art, for a particular range of parameters and conversely to give a significant weight to the synthesis matrix within the meaning of the invention for a other parameter range. Thus, one can maintain compatibility with existing systems within a certain operating range and improve the quality of the system in a particular range. In addition, the possibility of using several synthesis matrices obtained according to different criteria makes it possible to optimize the overall quality of the system for the entire operating range.

• des moyens de décorrélation (510) du signal somme pour obtenir un signal décorrélé;
• des moyens d'application (520) d'une matrice de synthèse (M Minq) dont les coefficients dépendent des paramètres de spatialisation, au signal décorrélé et au signal somme pour obtenir lesdits signaux de sortie,

caractérisé en ce que pour au moins une plage de valeur d'au moins un paramètre de spatialisation, les coefficients de la matrice de synthèse sont déterminés selon un critère de minimisation d'une fonction quantitative, relative à la quantité de signal décorrélé dans chacun des signaux de sortie obtenus par les moyens d'application de la matrice de synthèse la fonction quantitative étant telle que l'augmentation en valeur absolue des coefficients de la matrice de synthèse appliqués au signal décorrelé fait augmenter la valeur de ladite fonction appliquée à ces mêmes coefficients.The invention also relates to a device for spatially synthesizing a sum signal generating at least two output signals, the sum signal as well as spatialization parameters coming from a parametric coding device implementing a matrixing of an original multichannel signal. The device comprising:

• decorrelation means (510) of the sum signal for obtaining a decorrelated signal;
• application means (520) of a synthesis matrix (M Minq) whose coefficients depend on the spatialization parameters, the decorrelated signal and the sum signal to obtain said output signals,

characterized in that for at least one value range of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion of minimization of a quantitative function, relative to the quantity of signal decorrelated in each of the output signals obtained by the application means of the synthesis matrix, the quantitative function being such that the increase in absolute value of the coefficients of the synthesis matrix applied to the uncorrelated signal increases the value of said function applied to these same coefficients.

It relates to a decoder comprising a synthesis device as described above.

The invention also relates to multimedia equipment comprising a decoder as described above.

Without limitation, such equipment may be for example a mobile phone, an electronic organizer or digital content player, a computer, a set-top box ("set-top box").

Finally, the invention is directed to a computer program comprising code instructions for implementing the steps of the method as described above, when these instructions are executed by a processor.

• la figure 1 illustre un système de codage/décodage paramétrique classique de l'état de l'art tel que décrit précédemment;
• les figures 2 et 3 illustrent des exemples d'arbres de codage tels que décrits précédemment, selon la norme MPEG Surround dans le cas d'un signal multi-canal de type 5.1;
• la figure 4 illustre un système de décodage de l'état de l'art d'un bloc TTO tel que décrit précédemment;
• la figure 5 illustre un dispositif de synthèse selon l'invention pour le décodage d'un bloc TTO;
• la figure 6 illustre un dispositif de synthèse pour le décodage d'un bloc TTO selon un mode particulier de réalisation;
• la figure 7 illustre un décodeur selon l'invention dans le cas de signaux multicanaux de type 5.1; et
• la figure 8 illustre un exemple d'équipement multimédia comportant au moins un dispositif de synthèse selon l'invention.

Other features and advantages of the invention will appear more clearly on reading the following description, given solely by way of nonlimiting example, and with reference to the appended drawings, in which:

• the figure 1 illustrates a conventional parametric coding / decoding system of the state of the art as described above;
• the Figures 2 and 3 illustrate examples of encoding trees as described above, according to the MPEG Surround standard in the case of a 5.1 type multi-channel signal;
• the figure 4 illustrates a decoding system of the state of the art of a TTO block as described above;
• the figure 5 illustrates a synthesis device according to the invention for decoding a TTO block;
• the figure 6 illustrates a synthesis device for decoding a TTO block according to a particular embodiment;
• the figure 7 illustrates a decoder according to the invention in the case of multichannel signals of type 5.1; and
• the figure 8 illustrates an example of multimedia equipment comprising at least one synthesis device according to the invention.

The figure 5 illustrates an embodiment of the invention. It illustrates a synthesis device for the decoding of a TTO block (TTO ^-1 ). This device comprises a decorrelation module 510, able to carry out a step of decorrelation of the received signal s which is a sum signal obtained by coding by a matrix of multichannel signals.

This decorrelation step is for example that described in the MPEG Surround standard mentioned above.

This decorrelated signal d and the sum signal s are taken into account in a synthesis module 520 using a matrix M Minq whose coefficients depend on spatialization parameters R and I received and producing output signals I and r.

en respectant les conditions suivantes :

• l'énergie totale est conservée, c'est-à-dire: $${h_{11}}^2+{h_{12}}^2+{h_{21}}^2+{h_{22}}^2=1$$
  
• le rapport d'énergie entre l et r vaut R, c'est-à-dire: $${h_{11}}^2+{h_{12}}^2=R\left({h_{21}}^2+{h_{22}}^2\right)$$
  
• l'intercorrélation normalisée entre l et r vaut I, c'est-à-dire: $$\frac{h_{11}{h}_{21}+h_{12}{h}_{22}}{\sqrt{\left({h_{11}}^2+{h_{12}}^2\right)\left({h_{21}}^2+{h_{22}}^2\right)}}=I$$
  

More precisely, the signals l and r are generated by the following matrixing: $$\left[\begin{array}{c}l\\ r\end{array}\right]=\left[\begin{array}{cc}{h}_{11}& h_{12}\\ h_{21}& h_{22}\end{array}\right]\left[\begin{array}{c}s\\ d\end{array}\right]$$

respecting the following conditions:

• the total energy is conserved, that is to say: $${h_{11}}^2+{h_{12}}^2+{h_{21}}^2+{h_{22}}^2=1$$
  
• the energy ratio between l and r is R, that is to say: $${h_{11}}^2+{h_{12}}^2=R\left({h_{21}}^2+{h_{22}}^2\right)$$
  
• the standardized cross-correlation between l and r is equal to I, that is to say: $$\frac{h_{11}{h}_{21}+h_{12}{h}_{22}}{\sqrt{\left({h_{11}}^2+{h_{12}}^2\right)\left({h_{21}}^2+{h_{22}}^2\right)}}=I$$
  

Using the first two conditions, we have $${h_{11}}^2+{h_{12}}^2=\frac{R}{R+1}\mathrm{and}{h_{21}}^2+{h_{22}}^2=\frac{1}{R+1}$$

The solutions can be written in the form: $$h_{11}=\sqrt{\frac{R}{R+1}}\cos \left(\mathrm{at}\right),h_{12}=\sqrt{\frac{R}{R+1}}\sin \left(\mathrm{at}\right),h_{21}=\sqrt{\frac{1}{R+1}}\cos \left(b\right),h_{22}=\sqrt{\frac{1}{R+1}}\sin \left(b\right)$$

c'est-à-dire cos(a-b) = I.The third condition is then written: $$\cos \left(\mathrm{at}\right)\cos \left(b\right)+\sin \left(\mathrm{at}\right)\sin \left(b\right)=I$$

that is, cos ( ab ) = I.

avec $\alpha =\pm \frac{\arccos \left(I\right)}{2}\mathrm{.}$

We thus see that the solution matrices of the problem are the set of matrices parameterized by β ∈ [0, 2π) of the form: $$\left[\begin{array}{cc}{h}_{11}& h_{12}\\ h_{21}& h_{22}\end{array}\right]=\left[\begin{array}{cc}\sqrt{\frac{R}{R+1}}& 0\\ 0& \sqrt{\frac{1}{R+1}}\end{array}\right]\left[\begin{array}{cc}\cos \left(\beta +\alpha \right)& \sin \left(\beta +\alpha \right)\\ \cos \left(\beta -\alpha \right)& \sin \left(\beta -\alpha \right)\end{array}\right]$$

with $\alpha =\pm \frac{\mathrm{<figure-callout\; id="2"\; label="arccos\; I"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">arccos\; </figure-callout>}\left(I\right)\left(\right)}{2}\mathrm{.}$

Thus, two values of a are possible. The value of β is a function of R and I and is chosen according to one embodiment of the invention so as to limit the quantity of the decorrelated signal d introduced into the reconstructed signals regardless of the correlation values I, including for values negative.

Thus, the choice of the value β can be formalized by introducing a quantitative function q relating to the quantity of decorrelated signal taken into account in the matrixing for the reconstruction of the signals.

In general, the quantitative function q is such that the increase in absolute value of the coefficients of the synthesis matrix applied to the decorrelated signal increases the value of the function q applied to these same coefficients.

• pour tous réels x, x', y si |x'| ≥ |x| alors q (x', y) ≥ q(x, y)
• et symétriquement pour tous réels x, y, y' si |y'|≥|y| alors q(x, y') ≥ q(x, y).

Thus, this quantitative function q is such that it satisfies the following conditions:

• for all real x, x ', y if | x ' | ≥ | x | then q ( x ' , y ) ≥ q ( x, y )
• and symmetrically for all real x, y, y 'si | y ' | ≥ | y | then q ( x, y ' ) ≥ q ( x, y ).

For I and R fixed, the value of β is then chosen by minimizing the function: $$f\left(\beta \right)=q\left(h_{12},h_{22}\right)=q\left(\sqrt{\frac{R}{R+1}}\sin \left(\beta +\alpha \right),\sqrt{\frac{1}{R+1}}\sin \left(\beta -\alpha \right)\right)$$

Numerous quantitative functions complying with the conditions described above can be chosen and will make it possible to make a satisfactory choice of β.

avec p un entier supérieur ou égal à 1.Thus, the function q can for example be of type: $$q\left(x,\mathrm{there}\right)={\left({\left|x\right|}^p+{\left|\mathrm{there}\right|}^p\right)}^{\frac{1}{p}}$$

with p an integer greater than or equal to 1.

In a particular embodiment, the quantitative function q is a function of energy of the decorrelated signal.

The function q is such that: $$q\left(x,\mathrm{there}\right)=x^2+{\mathrm{there}}^2$$

Thus, the values of β guaranteeing a satisfactory reconstruction according to the embodiment of the invention described here are chosen so as to minimize the total energy of the decorrelated signal d in the reconstructed signals.

c'est-à-dire $${h_{12}}^2+{h_{22}}^2=\frac{1}{2}\left(\frac{R}{R+1}\left(1-\cos \left(2\beta +2\alpha \right)\right)+\frac{1}{R+1}\left(1-\cos \left(2\beta -2\alpha \right)\right)\right)$$

ce qui revient à maximiser : $$g\left(\beta \right)=\frac{R}{R+1}\cos \left(2\beta +2\alpha \right)+\frac{1}{R+1}\cos \left(2\beta -2\alpha \right)$$

We then seek β minimizing: $${h_{12}}^2+{h_{22}}^2=\frac{R}{R+1}{\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}}^2\left(\beta +\alpha \right)+\frac{1}{R+1}{\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}}^2\left(\beta -\alpha \right)$$

that is to say $${h_{12}}^2+{h_{22}}^2=\frac{1}{2}\left(\frac{R}{R+1}\left(1-\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\beta +2\alpha \right)\right)+\frac{1}{R+1}\left(1-\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\beta -2\alpha \right)\right)\right)$$

which amounts to maximizing: $$\mathrm{boy\; Wut}\left(\beta \right)=\frac{R}{R+1}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\beta +2\alpha \right)+\frac{1}{R+1}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\beta -2\alpha \right)$$

$$\mathit{gʹ}\left(\beta \right)=-2\left(\frac{R-1}{R+1}\sin \left(2\alpha \right)\cos \left(2\beta \right)+\frac{R+1}{R+1}\cos \left(2\alpha \right)\sin \left(2\beta \right)\right)$$

The derivative of g is: $$\mathit{boy\; Wut}\left(\beta \right)=-2\left(\frac{R}{R+1}\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}\left(2\beta +2\alpha \right)+\frac{1}{R+1}\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}\left(2\beta -2\alpha \right)\right)$$

$$\mathit{boy\; Wut}\left(\beta \right)=-2\left(\frac{R-1}{R+1}\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}\left(2\alpha \right)\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\beta \right)+\frac{R+1}{R+1}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(2\alpha \right)\mathrm{<figure-callout\; id="2"\; label="sin"\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">sin</figure-callout>}\left(2\beta \right)\right)$$

It vanishes when: $$\tan \left(2\beta \right)=\frac{1-R}{R+1}\tan \left(2\alpha \right)$$

et correspondant bien à une valeur de maximum de g.The value of β retained is therefore chosen from the values verifying $\beta =\frac{1}{2}\arctan \left(\frac{1-R}{R+1}\tan \left(2\mathrm{<figure-callout\; id="2"\; label="\alpha \; mod\; \pi "\; filenames="imgf0003.png,imgf0006.png"\; state="\{\{state\}\}">\alpha \; </figure-callout>}\right)\right)\mathrm{mod}\left(\frac{\pi }{}\right)\left(\frac{}{2}\right)$

and corresponding to a value of maximum of g.

So, the figure 5 represents a synthesis device for decoding a TTO block, here called TTO ^-1 comprising a decorrelation module 510 of the sum signal, a synthesis module 520 able to apply a synthesis matrix to the decorrelated signal and to the sum signal. The coefficients of this synthesis matrix are determined according to a criterion of minimization of a quantitative function q relative to the amount of decorrelated signal as described above.

The figure 5 also illustrates the steps of the spatial synthesis method according to the invention in which from a sum signal, at least two output signals 1 and r are obtained. The sum signal comes from a parametric encoding by mastering a multi-channel signal also providing spatialization parameters.

• décorrélation (Décorr.) du signal somme pour obtenir un signal décorrélé d;
• application (Synth.) d'une matrice de synthèse (M Minq) dont les coefficients dépendent des paramètres de spatialisation (I, R), au signal décorrélé (d) et au signal somme (s) pour obtenir lesdits signaux de sortie.

The method implemented by the synthesis device comprises the steps of:

• decorrelation (decorr.) of the sum signal to obtain a decorrelated signal d;
• application (Synth.) of a synthesis matrix (M Minq) whose coefficients depend on the spatialization parameters (I, R), the decorrelated signal (d) and the sum signal (s) to obtain the said output signals.

This method is such that for at least one value range of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion for minimizing a quantitative function, relating to the quantity of decorrelated signal taken into account in the step of applying the synthesis matrix.

In the embodiment described above with reference to the figure 5 , the spatialization parameters are parameters designating the energy ratio R between the channels of the original multi-channel signal and an inter-channel correlation measurement of this same signal.

Other spatialization parameters derived from parametric coding can also be chosen. These parameters may for example be parameters designating the phase shift between the channels of the multi-channel signal, or time envelope parameters of the audio channels.

The figure 6 illustrates another embodiment of the invention in which, according to a value range of at least one of the spatialization parameters received, here the interchannel correlation parameter I, a different synthesis matrix is chosen.

The example shown in figure 6 shows two types of synthesis matrix.

The first synthesis matrix M is for example that described in the state of the art in the MPEG Surround standard. The corresponding synthesis module is illustrated in 630. This synthesis matrix is applied here to the sum signal and the decorrelated signal d when the parameter I is positive.

When the parameter I is negative, the synthesis matrix M Minq is that described with reference to the figure 5 . The corresponding synthesis module is represented in 620.

Thus, the method implemented by this embodiment makes it possible to efficiently process multi-channel signals that exhibit negative interchannel correlations.

This type of multi-channel signal is for example a surround-type signal. Indeed, this type of signal has channels in phase opposition. This characteristic element of the signals from surround sound is illustrated in the articles by M. Gerzon entitled "Hierarchical System of Surround Sound Transmission for HDTV "or" Ambisonic Decoders for HDTV ".

In an alternative embodiment, several synthesis matrices can be provided for ranges of different values of the spatialization parameters.

Thus, it is possible to modulate the relative importance that we want to give to the different synthesis matrices as a function of the received parameter values.

For example, it is thus possible to give a significant weight to a matrix M as described in the state of the art for a particular range of parameters and conversely to give a significant weight to the synthesis matrix MMinq within the meaning of invention for another parameter range.

Compatibility with existing systems within a certain operating range is then maintained. An improvement in the quality of the synthesis in a particular range of spatialization parameter value is then provided in this embodiment.

In addition, the possibility of using several synthesis matrices obtained according to different criteria makes it possible to optimize the overall quality of the synthesis for the entire operating range.

For example, it is possible to use different synthesis matrices depending on whether the value of at least one spatialization parameter is small or, on the contrary, important.

Thus in this variant of the embodiment, use will be made of two synthetic matrices such that for positive values of the correlation index I, the matrix M as described in the state of the art, and for values negative of the correlation index I, we will use the matrix MMinq.

• pour I > 0, on utilise une matrice Minter = M
• pour 0≥ I>-0.25, on utilisera une interpolation des deux matrices Minter = α M + (1- α) MMinq
• pour -0.25≥ I>-1, on utilisera la matrice Minter = MMinq

It will also be possible to define different operating ranges, for example:

• for I> 0, we use a matrix Minter = M
• for 0≥ I> -0.25, we will use an interpolation of the two matrices Minter = α M + (1- α) MMinq
• for -0.25≥ I> -1, we will use the matrix Minter = MMinq

This type of device TTO ^-1 as represented in figure 5 or in figure 6 is for example integrated in a digital signal decoder. Such a type of decoder is for example illustrated with reference to the figure 7 .

The decoder shown in this figure is typically provided for decoding 5.1 type multi-channel signals. Thus, this decoder comprises a plurality of devices TTO ^-1 (TTO ₀ ^-1 , TTO ₁ ^-1 , TTO ₂ ^-1 , TTO ₃ ^-1 , TTO ₄ ^-1 ) according to the invention for, from a signal S received, obtain a multi-channel signal with 6 channels (L, R, C, LFE, Ls, Rs).

The decoding module 730 comprising this plurality of synthesis devices may, of course, be differently configured depending on the coding tree that has been used for the original multi-channel signal.

The decoder as represented in figure 7 comprises a QMF analysis module (for "quadrature Mirror Filter" in English) capable of performing a transformation of the sum (or downmix) signal S from the encoder into a frequency signal per subband. The signal per frequency band is then supplied to the input of the decoding module 730. At the output of the decoding module, the processed signals enter the QMF synthesis module 720 able to perform an inverse transformation and to bring back the multi-channel signal obtained. in the time domain.

These QMF analysis and QMF synthesis modules may for example be those as described in the MPEG Surround standard.

The decoder as represented in figure 7 receives from the encoder spatialization parameters P which are derived from the parametric encoding of the original multi-channel signal.

Typically, these parameters may be energy ratio parameters between the channels, correlation measurement between the channels or else phase shift between the channels or finally time envelope.

This decoder 700 can be integrated into a multimedia equipment type set-top box or "set-top box", computer or mobile phone, digital content player, personal electronic organizer, etc ...

The figure 8 represents an example of such multimedia equipment which comprises in particular an input module E adapted to receive multi-channel audio signals compressed either by a communication network for example or by means of a multi-channel sound recording.

These multi-channel signals have been compressed by a parametric coding method which, by mastering the original signal, generates a sum signal S and spatialization parameters P. This coding may in an alternative mode be provided in the multimedia equipment.

This equipment comprises one or more synthesis devices according to the invention, represented here physically by a processor PROC cooperating with a memory block BM comprising a memory storage and / or working MEM.

The memory block may advantageously comprise a computer program comprising code instructions for implementing the steps of the method in the sense of the invention, when these instructions are executed by the processor PROC, and in particular a decorrelation step of a signal. sum received to obtain a decorrelated signal and a step of applying a synthesis matrix whose coefficients depend on the spatialization parameters, the decorrelated signal and the sum signal to obtain at least two output signals. The synthesis matrix is such that, for at least one value range of at least one spatialization parameter, its coefficients are determined according to a criterion of minimization of a quantitative function, relative to the quantity of decorrelated signal taken into account in the step of applying the synthesis matrix.

Typically, the description of the figure 5 takes the steps of an algorithm of such a computer program. The computer program can also be stored on a memory medium readable by a reader of the device or downloadable in the memory space of the equipment.

The memory block thus comprises the coefficients of the synthesis matrix as defined above.

This memory block may comprise in another embodiment of the invention as described with reference to FIG. figure 6 , defining coefficients several synthesis matrices which are applied to the sum signal and to the decorrelated signal as a function of the range of values of the spatialization parameters received.

Similarly, the processor of the equipment may also include instructions for implementing the steps of analysis and synthesis of the decoder as described with reference to the figure 7 .

The multimedia equipment as illustrated also comprises an output S for delivering the reconstructed multi-channel signal S 'either by speaker-type reproduction means or by communication means capable of transmitting this multi-channel signal.

Claims (9)

1. Method for spatially synthesizing a sum signal to obtain at least two output digital audio signals, the sum signal together with spatialization parameters being output by a parametric coding by matrixing of an original multi-channel digital audio signal, the method comprising the steps of:

   - decorrelation (Decorr.) of the sum signal (s) to obtain a decorrelated signal (d);

   - application (Synth.) of a synthesis matrix (M Minq) whose coefficients depend on the spatialization parameters (R, I), to the decorrelated signal and to the sum signal so as to obtain said output signals,

   characterized in that for at least one value range of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion for minimizing a quantitative function (q), relating to the quantity of decorrelated signal in each of the output signals obtained by the step of applying the synthesis matrix, the quantitative function being such that the increase in absolute value of the coefficients of the synthesis matrix that are applied to the decorrelated signal increases the value of said function applied to these same coefficients.
2. Method according to Claim 1, characterized in that the quantitative function is an energy function of the decorrelated signal.
3. Method according to Claim 1, characterized in that the quantitative function is of the type:

   $q\left(x,y\right)={\left({\left|x\right|}^p+{\left|y\right|}^p\right)}^{\frac{1}{p}}$

   

   with p an integer greater than or equal to 1.
4. Method according to Claim 1, characterized in that the spatialization parameters are a parameter (R) of energy ratio between the channels of the multi-channel signal and a parameter (I) of interchannel correlation of the multi-channel signal, a value range being the range in which the interchannel correlation parameter is negative.
5. Method according to Claim 1, characterized in that a different quantitative function is chosen per value range of the spatialization parameters.
6. Device for spatially synthesizing a sum signal generating at least two output digital audio signals, the sum signal together with spatialization parameters being output by a parametric coding device implementing a matrixing of an original multi-channel digital audio signal, the device comprising:

   - means (510) for decorrelating the sum signal to obtain a decorrelated signal;

   - means (520) for applying a synthesis matrix (M Minq) whose coefficients depend on the spatialization parameters, to the decorrelated signal and to the sum signal so as to obtain said output signals,

   characterized in that for at least one value range of at least one spatialization parameter, the coefficients of the synthesis matrix are determined according to a criterion for minimizing a quantitative function, relating to the quantity of decorrelated signal in each of the output signals obtained by the means for applying the synthesis matrix, the quantitative function being such that the increase in absolute value of the coefficients of the synthesis matrix that are applied to the decorrelated signal increases the value of said function applied to these same coefficients.
7. Digital audio signal decoder comprising at least one synthesis device according to Claim 6.
8. Multimedia appliance comprising a decoder according to Claim 7.
9. Computer program comprising code instructions for the implementation of the steps of the method according to one of Claims 1 to 5, when these instructions are executed by a processor.

Images