EP0608174A1 - System for predictive encoding/decoding of a digital speech signal by an adaptive transform with embedded codes - Google Patents

Abstract

The invention relates to a system for the predictive encoding of a digital speech signal with embedded codes. The encoded digital signal (Sn) is formed by an encoded speech signal and, if appropriate, by auxiliary data. A perceptual weighting filter (11) is formed by a filter for short-term prediction of the speech signal to be encoded so as to produce a frequency distribution of the quantisation noise. A circuit (12) makes it possible to subtract from the perceptual signal @, the contribution from the past excitation signal @ so as to deliver an updated perceptual signal Pn. A long-term prediction circuit (13) is formed, as a closed loop, from a dictionary updated with the past excitation @ modeled for the lowest throughput and enables an optimal wave form and an associated estimated gain @ to be delivered. An orthonormal transform module (MT) includes an adaptive transform module (14) and a module (16) for progressive modelling by orthogonal vectors, this making it possible to deliver indices representing the encoded speech signal. A circuit (19) allows the insertion of auxiliary data through bit stealing from the encoded speech signal. Application to any transmission system and to the storage of speech signals. <IMAGE>

Description

The present invention relates to a predictive coding-decoding system for a digital speech signal by adaptive transform with nested codes.

In the predictive transform coders currently used, this type of coder being represented in FIG. 1, it is sought to construct a synthetic signal Sn as close as possible to the digital speech signal to be coded Sn, resemblance in the sense of a perceptual criterion.

The digital signal to be coded Sn, originating from an analog source speech signal, is subjected to a short-term prediction process, LPC analysis, the prediction coefficients being obtained by prediction of the speech signal on windows comprising M samples. The digital speech signal to be coded Sn is filtered by means of a perceptual weighting filter W (z) deduced from the aforementioned prediction coefficients, to obtain the perceptual signal pn

A long-term prediction process then makes it possible to take into account the periodicity of the residue for the voiced sounds, on all the sub-windows of N samples, N <M, in the form of a contribution P _n , which is subtracted from the signal perceptual pn so as to obtain the signal p'n in the form of a vector P'e _RN.

A transformation followed by a quantification are then carried out on the aforementioned vector P ′ in order to carry out a digital transmission. The reverse operations allow, after transmission, the modeling of the synthetic signal S _n .

In order to obtain good perceptual behavior, according to the usual criteria established by experience, it is necessary to establish a transformation process by orthonormal transform F and quantization of the vector P ', in the presence of gain values G verifying well determined properties, G = F ^T .P 'where F ^T denotes the matrix transposed from the matrix F.

A first solution, proposed by G. Davidson and A. Gersho, in the publication "Multiple-Stage Vector Excitation Coding of Speech Wave forms", ICASSP 88, Vol. 1, pp 163-166, consists in using a transformation matrix not singular V = HC where H is a lower triangular matrix and C a non-singular dictionary, constructed by learning, ensuring the invertibility of the transformation matrix V for any sub-window.

In order to be able to exploit certain decorrelation and ordering properties of the components of the vector of coefficients of the transform G during the quantification step, several solutions using orthonormal transforms have been proposed.

où est le nombre de vecteurs contenu dans le corpus d'apprentissage, permet de maximiser l'expression

où K est un entier, K Z N. On démontre que l'erreur quadratique moyenne de la transformée de Karhunen-Loeve est inférieure à celle de toute autre transformation pour un ordre de modélisation K donné, cette transformée étant, dans ce sens, optimale. Ce type de transformée a été introduit dans un codeur prédictif par transformée orthogonale par N.Moreau et P.Dymarski, confer publication "Successive Orthogonalisations in the Multistage CELP Coder", ICASSP 92 Vol.1, pp 1-61 - 1-64.The Karhunen-Loeve transform, obtained from the eigenvectors of the autocorrelation matrix

where is the number of vectors contained in the learning corpus, allows to maximize the expression

where K is an integer, KZ N. We show that the mean square error of the Karhunen-Loeve transform is lower than that of any other transformation for a given modeling order K, this transform being, in this sense, optimal. This type of transform was introduced into a predictive coder by orthogonal transform by N. Moreau and P. Dymarski, confer publication "Successive Orthogonalisations in the Multistage CELP Coder", ICASSP 92 Vol.1, pp 1-61 - 1-64.

However, in order to reduce the computation complexity of the gain vector G, it is possible to use sub-optimal transforms, such as the Fast Fourier transform (FFT), the discrete cosine transform (TCD) the discrete transform of Hadamard (DHT) or Walsh Hadamard (DWHT) for example.

matrice dans laquelle h(n) est la réponse impulsionnelle du filtre de prédiction à court terme 1/A(z) de la fenêtre courante.Another method for the construction of an orthonormal transform consists in decomposing into singular values the lower triangular Toeplitz matrix H defined by:

matrix in which h (n) is the impulse response of the short-term prediction filter 1 / A (z) of the current window.

The matrix H can then be decomposed into a sum of matrices of rank 1:

Since the matrix U is unitary, it can be used as an orthonormal transform. Such a construction was proposed by BSAtal in the publication "A Model of LPC Excitation in Terms of Eigenvectors of the Autocorrelation Matrix of the Impulse Response of the LPC Filter", ICASSP 89, Vol.1, pp 45-48 and by E .Ofer in the publication "A Unified Framework for LPC Excitation Representation in Residual Speech Coders" ICASSP 89, Vol.1 pp 41-44.

Coders with nested codes currently known make it possible to transmit parvol data of binary elements normally allocated to speech on the transmission channel, and this, in a manner transparent to the coder, which codes the speech signal at the maximum bit rate.

Among this type of encoder, a 64 kbit / s encoder with scaled quantizer with nested codes was standardized in 1986 by standard G 722 established by the CCITT. This coder operating in the field of wideband speech (audio signal with a bandwidth of 50 Hz to 7 kHz, sampled at 16 kHz), is based on a coding in two sub-bands each containing a Pulse Modulation coder and Adaptive Differential Coding (MICDA coding). This coding technique allows broadband speech signals and data, if necessary, to be transmitted on a 64 kbit / s channel, at three different bit rates 64-56-48 kbit / s and 0-8-16 kbit / s for data.

In addition, in the context of the implementation of coded excited coders (or CELP coders) M. Johnson and T. Tanigushi have described a multistage CELP coder with nested codes. See the publication by the aforementioned authors entitled "Pitch Orthogonal Code-Excited LPC", Globecom 90, Vol.1, pp 542-546.

Finally, R.Drogo De lacovo and D.Sereno described a modified CELP type coder allowing to obtain nested codes or modeling the excitation signal of the LPC analysis filter by a sum of different contributions and using only the first of them for updating the memory of the synthesis filter, confer the publication of these authors "Embedded CELP Coding For Variable Bit-Rate Between 6.4 and 9.6 kbit / s" ICASSP 91 Vol.1, pp 681-684 .

The aforementioned prior art transform predictive coders do not make it possible to transmit data and therefore cannot fulfill the function of nested code coders. In addition, the nested code coders of the prior art do not use the orthonormal transform technique, which does not make it possible to tend towards or to achieve optimal transform coding.

The object of the present invention is to remedy the aforementioned drawback by implementing a predictive coding-decoding system for a digital speech signal by adaptive transform with nested codes.

Another object of the present invention is the implementation of a predictive coding-decoding system for a digital speech and data signal allowing transmission at reduced and flexible rates.

The system for predictive coding of a digital signal into a digital code nested code signal, in which the coded digital signal consists of a coded speech signal and, where appropriate, of an auxiliary data signal inserted into the coded speech signal after coding of the latter, object of the present invention, comprises a perceptual weighting filter controlled by a short-term prediction loop making it possible to generate a perceptual signal and a long-term prediction circuit delivering an estimated perceptual signal, this circuit long-term prediction signal forming a long-term prediction loop making it possible to deliver, from the perceptual signal and the estimated past excitation signal, a modeled perceptual excitation signal and adaptive transform and quantization circuits making it possible to of the perceptual excitation signal to generate the coded speech signal.

It is remarkable in that the perceptual weighting filter consists of a short-term prediction filter of the speech signal to be coded, so as to achieve a frequency distribution of the quantization noise and in that it comprises a circuit for subtracting the contribution of the past excitation signal from the perceptual signal to deliver an updated perceptual signal, the long-term prediction circuit being formed, in closed loop, from a dictionary updated by the past excitation modeled corresponding to the lowest bit rate allowing the delivery of an optimal waveform and a gain associated with it, constituting the estimated perceptual signal. The transform circuit is formed by an orthonormal transform module comprising an adaptive orthogonal transformation module and a progressive modeling module by orthogonal vectors. The progressive modeling module and the long-term prediction circuit make it possible to deliver indexes representative of the coded speech signal. An auxiliary data insertion circuit is coupled to the transmission channel.

The system for adaptive transform predictive decoding of a nested coded digital signal in which the coded digital signal consists of a coded digital signal and, where appropriate, of an auxiliary data signal inserted into the coded speech signal after coding of the latter, is remarkable in that it comprises a circuit for extracting the data signal allowing, on the one hand, the extraction of the data for an auxiliary use and, on the other hand, the transmission of 'indexes representative of the coded speech signal. It further comprises a circuit for modeling the speech signal at the minimum bit rate and a circuit for modeling the speech signal at at least one bit rate greater than the minimum bit rate.

The predictive coding-decoding system of a digital speech signal by adaptive transform with nested codes object of the present invention finds application, in general, to the transmission of speech and data at flexible rates, and, more particularly , audiovisual conference protocols, videophone, loudspeaker telephony, storage and transport of digital audio signals over long distance links, transmission with mobiles and channel concentrating systems.

• - la figure 2 représente un schéma de principe du système de codage prédictif d'un signal de parole par transformée adaptative à codes imbriqués objet de la présente invention,
• - la figure 3 représente un détail de réalisation d'un module de prédiction à long terme en boucle fermée utilisé dans le système de codage représenté en figure 2,
• - les figures 4a et 4b représentent un schéma partiel d'un codeur prédictif par transformée et un schéma équivalent au schéma partiel de la figure 4a,
• - la figure 5a représente un organigramme d'un processus de transformée orthonormée construit par apprentissage,
• - la figure 5b représente deux diagrammes comparatifs de valeurs de gain normalisées obtenus par décomposition en valeurs singulières respectivement par apprentissage,
• - les figures 6a et 6b représentent schématiquement le processus de transformation de Householder appliqué au signal perceptuel,
• - la figure 7 représente un module de transformation adaptative mettant en oeuvre une transformation de Householder,
• - la figure 8a représente, pour la décomposition en valeurs singulières respectivement la construction pour apprentissage, un critère normalisé de gain en fonction du nombre de composantes du vecteur de gains,
• - la figure 8b représente un schéma de principe de quantification vectorielle multiétage dans lequel le vecteur de gains G est obtenu par combinaison linéaire de vecteurs issus de dictionnaires stochastiques,
• - la figure 9 est une représentation géométrique de la prospection du vecteur de gain G dans un sous-espace de vecteurs issus de dictionnaires stochastiques,
• - les figures 10a e 10b représentent le schéma de principe d'un processus de quantification vectorielle de gain par modélisations progressives orthogonales, correspondant à une projection optimale de ce vecteur de gain représentée en figure 9, dans le cas d'un seul respectivement de plusieurs dictionnaires stochastiques,
• - la figure 11 représente un mode de réalisation de la modélisation de l'excitation du filtre de synthèse correspondant au débit le plus faible,
• - la figure 12 représente un schéma de principe d'un système de décodage prédictif d'un signal de parole par transformée adaptative à codes imbriqués objet de la présente invention,
• - la figure 13a représente un schéma de principe d'un module de modélisation du signal de parole au débit minimum,
• - la figure 13b représente un mode de réalisation d'un module de transformation orthonormée inverse,
• - la figure 14a représente un schéma d'un module de modélisation du signal de parole aux débits autres que le débit minimum,
• - la figure 14b représente un schéma équivalent au module de modélisation représenté en figure 14a,
• - la figure 15 représente la mise en oeuvre d'un filtre adaptatif de post-filtrage destiné à améliorerla qualité perceptuelle du signal de parole de synthèse Sn.

A more detailed description of the coding-decoding system which is the subject of the invention will be given below in relation to the drawings in which, in addition to FIG. 1 relating to the prior art concerning a predictive transform coder,

• FIG. 2 represents a block diagram of the predictive coding system of a speech signal by adaptive transform with nested codes which is the subject of the present invention,
• FIG. 3 represents a detail of an embodiment of a long-term closed-loop prediction module used in the coding system represented in FIG. 2,
• FIGS. 4a and 4b represent a partial diagram of a predictive coder by transform and a diagram equivalent to the partial diagram of FIG. 4a,
• FIG. 5a represents a flow diagram of an orthonormal transform process constructed by learning,
• FIG. 5b represents two comparative diagrams of normalized gain values obtained by decomposition into singular values respectively by learning,
• FIGS. 6a and 6b schematically represent the Householder transformation process applied to the perceptual signal,
• FIG. 7 represents an adaptive transformation module implementing a Householder transformation,
• FIG. 8a represents, for the decomposition into singular values respectively the construction for learning, a normalized gain criterion as a function of the number of components of the gain vector,
• FIG. 8b represents a principle diagram of multi-stage vector quantization in which the gain vector G is obtained by linear combination of vectors from stochastic dictionaries,
• FIG. 9 is a geometric representation of the prospecting of the gain vector G in a subspace of vectors from stochastic dictionaries,
• - Figures 10a and 10b represent the schematic diagram of a vector quantization process of gain by orthogonal progressive modeling, corresponding to an optimal projection of this gain vector represented in Figure 9, in the case of a single respectively of several stochastic dictionaries,
• FIG. 11 represents an embodiment of the modeling of the excitation of the synthesis filter corresponding to the lowest bit rate,
• FIG. 12 represents a block diagram of a system for predictive decoding of a speech signal by adaptive transform with nested codes which is the subject of the present invention,
• FIG. 13a represents a block diagram of a module for modeling the speech signal at the minimum rate,
• FIG. 13b represents an embodiment of a reverse orthonormal transformation module,
• FIG. 14a represents a diagram of a module for modeling the speech signal at bit rates other than the minimum bit rate,
• FIG. 14b represents a diagram equivalent to the modeling module represented in FIG. 14a,
• FIG. 15 represents the implementation of an adaptive post-filtering filter intended to improve the perceptual quality of the synthetic speech signal Sn.

A more detailed description of a system for predictive coding of a digital speech of adaptive partransformed speech into a digital signal with nested codes will now be given in connection with FIG. 2 and the following figures.

In general, it is considered that the digital signal coded by the implementation of the coding system which is the subject of the present invention consists of a coded speech signal and, where appropriate, by an auxiliary data signal inserted into the coded speech signal. , after coding of this digital speech signal.

Of course, the coding system which is the subject of the present invention may comprise, from a transducer delivering the analog speech signal, an analog-digital converter and an input storage circuit or input buffer making it possible to deliver the digital signal to code Sn.

The coding system which is the subject of the present invention also comprises a perceptual weighting filter 11 controlled by a short-term prediction loop making it possible to generate a perceptual signal, denoted P.

It also includes a long-term prediction circuit, denoted 13, delivering an estimated perceptual signal, which is denoted P ′.

The long-term prediction circuit 13 forms a long-term prediction loop making it possible to deliver, from the perceptual signal and from the estimated past excitation signal, denoted ^p _n ^o , a modeled perceptual excitation signal.

The coding system which is the subject of the invention as shown in FIG. 2 further comprises an adaptive transform and quantization circuit making it possible, from the perceptual excitation signal P _n, to generate the coded speech signal as it will be described below in the description.

,le dispositif de codage selon l'invention comprend ainsi que représenté sur la même figure 2 un circuit 120 de soustraction de la contribution du signal d'excitation passée P̂⁰ _n du signal perceptuel pour délivrer un signal perceptuel réactualisé, ce signal perceptuel réactualisé étant noté P_n.According to a first particularly advantageous aspect of the coding system which is the subject of the present invention, the perceptual weighting filter 11 consists of a filter for short-term prediction of the speech signal to be coded, so as to achieve a frequency distribution of the quantization noise. The perceptual weighting filter 11 delivering the perceptual signal

, the coding device according to the invention thus comprises, as shown in the same figure 2, a circuit 120 for subtracting the contribution of the past excitation signal P̂ ⁰ _n from the perceptual signal to deliver a refreshed perceptual signal, this refreshed perceptual signal being noted P _n .

According to another particularly advantageous characteristic of the coding device which is the subject of the present invention, the long-term prediction circuit 13 is formed in a closed loop from a dictionary updated by the past excitation modeled corresponding to the lowest bit rate, this dictionary to deliver an optimal waveform and an estimated gain associated with it. In Figure 2, the modeled past excitation corresponding to the lowest flow rate is noted r̂ ¹ _n . It is further indicated that the optimal waveform and the estimated gain associated with it constitute the estimated perceptual signal Pn delivered by the long-term prediction circuit 13.

According to another characteristic of the coding system which is the subject of the present invention, as shown in FIG. 2, the transform module circuit, denoted MT, is formed by an orthonormal transform module 14, comprising an adaptive orthogonal transformation module proper and a progressive modeling module using orthogonal vectors, noted 16.

In accordance with a particularly advantageous aspect of the coding system which is the subject of the present invention, the progressive modeling module 16 and the long-term prediction circuit 13 make it possible to deliver indices representative of the coded speech signal, these indices being denoted i (0 ), j (0) respectively i (1), j (1) with 1 e [1, L] in Figure 2.

Finally, the coding system according to the invention further comprises a circuit 19 for inserting auxiliary data coupled to the transmission channel, noted 18.

The operation of the coding device which is the subject of the present invention can be illustrated in the following manner.

As indicated above, it is sought to reconstruct a synthetic signal S _n as closely as possible perceptually resembling the digital signal to be coded Sn.

The synthetic signal S _n is of course the signal reconstituted on reception, that is to say at the decoding level after transmission as will be described later in the description.

A short-term prediction analysis formed by the analysis circuit 10 of the LPC type for "Linear Predictive Coding" and by the perceptual weighting filter 11 is carried out for the digital signal to be coded by a conventional prediction technique on windows comprising for example M samples. The analysis circuit 10 then delivers the coefficients a _i , where the aforementioned coefficients are the linear prediction coefficients.

.The speech signal to be coded Sn is then filtered by the perceptual weighting filter 11 of transfer function W (z), which makes it possible to deliver the perceptual signal proper, noted

.

The coefficients of the perceptual weighting filter are obtained from a short-term prediction analysis on the first correlation coefficients of the sequence of the coefficients a of the analysis filter A (z) of circuit 10 for the current window. This operation makes it possible to achieve a good frequency distribution of the quantization noise. In fact, the perceptual signal delivered tolerates greater coding noise in high-energy areas where the noise is less audible, since it is frequently masked by the signal. It is indicated that the perceptual filtering operation is broken down into two stages, the digital signal to code Sn being filtered a first time by the filter constituted by the analysis circuit 10, in order to obtain the residue to be modeled, then a second times by the perceptual weighting filter 11 to deliver the perceptual signal _n .

In the operating process of the coding device which is the subject of the present invention, the second operation consists in removing the contribution of the past excitation, or estimated past excitation signal, noted P̂ _n ⁰ from the aforementioned perceptual signal.

Indeed, we show that:

In this relation, h _n is the impulse response of the double filtering performed by the circuit 10 and the perceptual weighting filter 11 in the current window and r̂ ¹ _n is the past excitation modeled corresponding to the lowest flow rate, as well as will be described later in the description.

The operating mode of the long-term prediction circuit 13 in closed loop is then as follows. This circuit makes it possible to take into account the periodicity of the residue for the voiced sounds, this long-term prediction being carried out all the sub-windows of N samples, as will be described in connection with FIG. 3.

par rapport aux deux paramètres g_o et q.The closed-loop long-term prediction circuit 13 comprises a first stage constituted by an adaptive dictionary 130, which is updated all the aforementioned sub-windows by the modeled excitation denoted ¹ _n , delivered by the module 17, which will be described later in the description. The adaptive dictionary 130 makes it possible to minimize the error, noted

with respect to the two parameters g _o and q.

Such an operation corresponds, in the frequency domain, to filtering by the transfer function filter:

This operation is equivalent to the search for the optimal waveform, noted f ^{j (0)} and its associated gain g _o in a judiciously constructed dictionary. See the article published by R. Rose, and T. Barnwell, entitled "Design and Performance of an Analysis by Synthesis Class of Predictive Speech Coders", IEEE Trans. on Acoustic Speech Signal Proceessing, September 1990.

issue du dictionnaire adaptatif, est filtrée par un filtre 131 et correspond à l'excitation modélisée au débit le plus faible r retardé de q échantillons par le filtre précité. La forme d'onde optimale fo est délivrée par le dictionnaire adaptatif filtré 133.The waveform of index i. noted

from the adaptive dictionary, is filtered by a filter 131 and corresponds to the excitation modeled at the lowest rate r delayed by q samples by the aforementioned filter. The optimal waveform fo is delivered by the filtered adaptive dictionary 133.

A module 132 for calculating and quantifying the prediction gain makes it possible, from the perceptual signal P _n and all the waveforms f ^{j (0)} _n, to perform a calculation for quantifying the prediction gain, and to deliver an index i (0) representative of the number of the quantization range, as well as its associated quantized gain g (0).

A multiplier circuit 134 delivers from the filtered adaptive dictionary 133, that is to say from the filtering result of the waveform of index j C; ,, or fn, and of the associated quantized gain g (0) , the long-term prediction excitation modeled and filtered perceptually noted P _n ¹ .

A subtractor circuit 135 then makes it possible to perform a minimization relating to en = P _n - _n ¹ |, this expression representing the error signal. A module 136 makes it possible to calculate the Euclidean norm 1 in 12.

A module 137 makes it possible to search for the optimal waveform corresponding to the minimum value of the above-mentioned Euclidean standard and to deliver the index j (0). The parameters transmitted by the coding system object of the invention for the modeling of the long-term prediction signal are then the index j (0) of the optimal waveform f (0) as well as the number i (0 ) of the quantization range of its associated gain g (0) quantized.

A more detailed description of the adaptive orthogonal transformation module MT of FIG. 2 will be given in conjunction with FIGS. 4a and 4b.

In the context of the implementation of the predictive coding system by orthonormal transform object of the present invention, the method used for the construction of this transform corresponds to that proposed by BSAtal and E.Ofer, as mentioned previously in the description .

According to the embodiment of the coding system according to the present invention, this consists in decomposing, not the short-term prediction filtering matrix, but the perceptual weighting matrix W formed by a lower triangular Toeplitz matrix defined by the relation (4):

In this relation, w (n) denotes the impulse response of the perceptual weighting filter W (z) of the current window previously mentioned.

In FIG. 4a, there is shown the partial diagram of a predictive coder by transform and in FIG. 4b, the corresponding equivalent diagram in which the matrix or filter of perceptual weighting W, designated by 140, has been highlighted, a inverse perceptual weighting filter 121 having however been inserted between the long-term prediction module 13 and the subtracting circuit 120. It is indicated that the filter 140 achieves a linear combination of the basic vectors obtained from a decomposition into singular values of the representative matrix of the perceptual weighting filter W.

As shown in FIG. 4b, the signal S ', corresponding to the speech signal to be coded S _{n from} which it has been subtracted the contribution of the past excitation delivered by the module 12, as well as that of the long-term prediction P̂ ¹ _n filtered by a reverse perceptual weighting module with transfer function (W (z)) ^-1 , is filtered by the perceptual weighting filter with transfer function W (z), so as to obtain the vector P '.

et peut être exprimée sous forme d'une combinaison linéaire de vecteurs de base en utilisant la décomposition en valeurs singulières de la matrice W.This filtering operation is written:

and can be expressed as a linear combination of basic vectors using the decomposition into singular values of the matrix W.

As far as the embodiment of the perceptual weighting filter 140 is concerned, it is indicated that this comprises for each matrix W representative of the perceptual weighting filter a first matrix module U = (U ₁ , ..., U _N ) and a second matrix module V = (V ₁ , ..., V _N ).

relation dans laquelle :

• - U^T désigne le module matrice transposée du module U,
• - D est un module matrice diagonale dont le coefficients constituent les valeurs singulières précitées,
• - U et V_j désignent respectivement le ième vecteur singulier gauche et le jème vecteur singulier droit, les vecteurs singuliers droit {Vj} formant une base orthonormée.

The first and second matrix modules verify the relationship:

relationship in which:

• - U ^T designates the matrix module transposed from the module U,
• - D is a diagonal matrix module whose coefficients constitute the aforementioned singular values,
• - U and V _j respectively denote the ith left singular vector and the jth right singular vector, the right singular vectors {Vj} forming an orthonormal base.

Such a decomposition makes it possible to replace the filtering operation by convolution product by a filtering operation by a linear combination.

We indicate that the decomposition into singular values of the perceptual filtering matrix W makes it possible to obtain the two unit matrices U and V verifying the above-mentioned relation where

avec la propriété d'ordonnancement telle que d_i ≧ d_i+1 > 0. Les éléments d_i sont appelés valeurs singulières, et les vecteurs U et V_j, ième vecteur singulier gauche, respectivement jème vecteur singulier droit.

with the scheduling property such that d _i ≧ d _{i + 1} > 0. The elements d _i are called singular values, and the vectors U and V _j , ith left singular vector, respectively jth right singular vector.

The matrix W is then decomposed into a sum of matrices of rank 1, and verifies the relation:

permet d'obtenir le vecteur P' vérifiant la relation :

avec g(k) = ⁹ (k)d_k.The matrix V being unitary, the right singular vectors {V _i } form an orthonormal base and the signal S ', expressed in the form:

allows to obtain the vector P 'verifying the relation:

with g (k) = ⁹ (k) d _k .

By the process of decomposition into singular values, we indicate that a change on a component of the excitation Associated with a small singular value produces a small change at the output of the filter 140 and vice versa for the filtering operation reverse perceptual performed by module 121.

fⁱ _orth U_i pour i = 1 à N.In order to use these properties, the unitary matrix U can be used as an orthonormal transform, checking the relation:

f ⁱ _orth U _i for i = 1 to N.

The weighted perceptual signal P 'then breaks down as follows:

After vector quantization of the gains G, the weighted perceptual signal modeled P is calculated in the following manner:

We indicate that the left singular vectors associated with the largest singular values play a preponderant role in the modeling of the weighted perceptual signal P '. Thus, in order to model the latter, it is possible to keep only the components associated with the K largest singular values, K <N, that is to say the K first components of the gain vector G verifying the relation:

The short-term analysis filtering circuit 10 being updated on windows of M samples, the decomposition into singular values of the perceptual weighting matrix W is carried out at the same frequency.

Decomposition processes into singular values of any matrix allowing rapid processing have been developed, but the calculations remain relatively complex.

In accordance with an object of the present invention, it is, in order to simplify the abovementioned processing operations, proposed to construct a fixed orthonormal transform under suboptimal however possessing good perceptual properties, whatever the current window.

In a first embodiment, as shown in FIG. 5, the orthonormal transform process is constructed by learning. In such a case, the orthonormal transform module can be formed by a stochastic transform sub-module constructed by drawing a Gaussian random variable for initialization, this sub-module comprising in FIG. 5 the process steps 1000, 1001, 1002 and 1003 and being noted SMTS. Step 1002 can consist in applying the K-average algorithm to the aforementioned vector corpus.

The SMTS sub-module is successively followed by a module 1004 for building centers, a module 1005 for building classes and, in order to obtain a vector G whose components are relatively ordered, by a module 1006 for reordering of the transform according to the cardinal of each class.

The aforementioned module 1006 is followed by a Gram-Schmidt calculation module, noted 1007a, so as to obtain an orthonormal transform. The aforementioned module 1007a is associated with a module 1007b for calculating the error under the conventional conditions for implementing the Gram-Schmidt processing process.

The module 1007a is itself followed by a module 1008 for testing the number of iterations, this in order to allow an orthonormal transform carried out offline by learning to be obtained. Finally, the memory 1009 of read-only memory type makes it possible to store the orthonormal transform in the form of a transform vector. It is indicated that the relative ordering of the components of the gain vector G is accentuated by the process of orthogonalisation. When the learning construction process has converged, an orthonormal transform is obtained whose waveforms are gradually correlated with the vector learning corpus delivered by the initial transform step 1001.

FIG. 5b represents the ordering of the components of the gain vector G, that is to say of the normalized mean value G for a transform obtained on the one hand by decomposition into singular values of the perceptual weighting matrix W, and on the other hand, by learning. The transform F obtained by this last method for those of the orthonormal waveforms whose frequency spectra are bandpass and relatively ordered as a function of k, which makes it possible to attribute to this transform pseudo-frequency properties. An evaluation of the quality of transformation in terms of energy concentration made it possible to show that, by way of indication, on a corpus of 38,000 perceptual vectors P ′, the transformation gain is 10.35 decibels for the optimal Karhunen- transform. Loeve, and 10.29 decibels for a transform constructed by learning, the latter therefore tending towards the optimal transform in terms of energy concentration.

As previously mentioned in the description, the orthonormal transform F can be obtained according to two different methods.

By observing that generally the waveform most correlated with the perceptual signal P is that resulting from the adaptive dictionary, it is possible to envisage carrying out an adaptive orthonormal transform F 'whose f ",, th is equal to the form d optimal wave from the normalized adaptive dictionary f! (0), the first component of the gain vector G then being equal to the normalized long-term prediction gain g (0), which it is not necessary to recalculate, since this was quantified during this prediction.

The new dimension of the gain vector G then becomes equal to N-1, which makes it possible to increase the number of binary elements per sample during the vector quantization of the latter and therefore the quality of its modeling.

A first solution for calculating the transform F 'can then consist in making a long-term prediction analysis, shifting the transform obtained by learning a notch, placing the long-term predictor in the first position, then applying the Gram-Schmidt algorithm, in order to obtain a new transform F '.

avec

A second, more advantageous solution consists in using a transformation making it possible to rotate the orthonormal base, so that the first waveform coincides with the long-term predictor, that is to say:

with

avec B = f^j(0) - |f^j(0)| - f¹ _orth In order to keep the orthogonality property, the transformation used must keep the dot product. A particularly suitable transformation is the Household transformation, verifying the relationship:

with B = f ^{j (0)} - | f ^{j (0)} | - f ¹ _orth

A geometric representation of the aforementioned transform is given in FIGS. 6a and 6b.

For a more detailed definition of this type of transformation, one can usefully refer to the publication by Alan O.Steinhardt entitled "Householder Transforms in Signal Processing", IEEE ASSP Magazine, July 1988, pp 4-12.

avec P' = TP = (P-B[wB^TP]).By using this transformation, it is possible to reduce the complexity of the calculations and the projection of the perceptual signal P in this new base is written:

with P '= TP = (PB [wB ^T P]).

In this relation, w denotes a scalar equal to w = 2 / B ^T B.

It is indicated that in this embodiment of the orthonormal transform, the transformation is applied only to the perceptual signal P, and the modeled perceptual signal P can then be calculated by the inverse transformation.

A particularly advantageous embodiment of the orthonormal transform module proper 14 in the case where a Householder transformation is used will now be described in conjunction with FIG. 7.

As shown in FIG. 7 above, the adaptive transformation module 14 can include a Householder transformation module 140 receiving the estimated perceptual signal constituted by the optimal waveform and by the estimated gain and the perceptual signal. P to generate a transformed perceptual signal P ". It is indicated that the Householder transformation module 140 comprises a calculation module 1401 of the parameters B and wB as defined previously by the relation 13. It also comprises a module 1402 comprising a multiplier and a subtractor making it possible to carry out the transformation proper according to relation 14. It is indicated that the transformed perceptual signal P "is delivered in the form of vector of perceptual signal transformed of component P" _k , with ke [0, N-1].

The adaptive transformation module 14 as shown in FIG. 7 also includes a plurality N of registers for memorizing orthonormal waveforms, the current register being denoted r, with re [1, N]. It is indicated that the aforementioned N storage registers form the read-only memory previously described in the description, each register comprising N storage cells, each component of rank k of each vector, component denoted f ¹ _{orth (k)} being stored in a cell of corresponding rank of the current register r considered.

In addition, as will be seen in FIG. 7, the module 14 comprises a plurality of N multiplier circuits associated with each rank register reforming the plurality of the previously mentioned storage registers. In addition, each multiplier register of rank k receives on the one hand the component of rank k of the stored vector and on the other hand the component P " _k of the transformed perceptual signal vector of corresponding rank k. The multiplier circuit Mrk delivers the product P "kf: 'tt (k) of the components of transformed perceptual signal.

Finally, a plurality of N-1 summing circuits is associated with each register of rank r, each summing circuit of rank k, denoted Srk, receiving the product of prior rank k-1, and the product of corresponding rank k delivered by the circuit. multiplier Mrk of the same rank k. The summing circuit of highest rank, SrN-1, then delivers a component g (r) of the estimated gain expressed in the form of gain vector G.

We indicate that the predictive coding system using the adaptive orthonormal transform constructed by learning is likely to give better results, while the Householder transformation makes it possible to obtain a reduced complexity.

As will be observed in FIG. 2, the progressive modeling module by orthogonal vectors in fact comprises a module 15 for normalizing the gain vector to generate a normalized gain vector, denoted G _k , by comparison of the normalized value of the gain vector G with respect to a threshold value. This normalization module 15 also makes it possible to generate a signal of length of the normalized gain vector linked to the modeling order k to the decoder system as a function of this modeling order.

The progressive modeling module by orthogonal vectors further comprises, in cascade with the module 15 for normalization of the gain vector, a stage 16 of progressive modeling by orthogonal vectors. This modeling stage 16 receives the normalized vector Gk and delivers the indexes representative of the coded speech signal, these indexes being denoted I (1), J (1), these indexes being representative of the selected vectors and their associated gain. The transmission of the auxiliary data formed by the indexes is carried out by overwriting the parts of the frame allocated to the indices and track numbers to form the auxiliary data signal.

The operation of the standardization module 15 is as follows.

est constante pour une sous fenêtre donnée. Dans ces conditions, maximiser cette énergie est équivalent à minimiser l'expression :

où G_{k =} (0,g₂,9₃,...,9_K,0,...0).The energy of the perceptual signal given by

is constant for a given sub window. Under these conditions, maximizing this energy is equivalent to minimizing the expression:

where G _{k =} (0, g ₂ , 9 ₃ , ..., 9 _K , 0, ... 0).

It is indicated that during such an operation, an additional way of increasing the number of binary elements per sample during the vector quantization of the vector G is to use the following normalized criterion, consisting in choosing K such that:

The gain vector thus obtained G _K is then quantified and its length k is transmitted by the coding system object of the invention in order to be taken into account by the corresponding decoding system, as will be described later in the description.

The average normalized criterion as a function of the modeling order K is given in FIG. 8a for an orthonormal transform obtained on the one hand by decomposition into singular values of the perceptual weighting matrix W and on the other hand by learning.

A particularly advantageous embodiment of the progressive modeling module by orthogonal vectors 16 will now be given in connection with FIG. 8b. The aforementioned module makes it possible in fact to perform a multistage vector quantization.

The gain vector G is obtained by linear combination of vectors, noted

These vectors being derived from stochastic dictionaries, noted 161, 162, 16 L, constructed either by drawing a Gaussian random variable, or by learning. The estimated gain vector G verifies the relation:

In this relation, 8 ₁ is the gain associated with the optimal vector Ψ ^{j (1)} _K from the stochastic dictionary of rank 1, noted 16 1.

However, the vectors selected iteratively are generally not linearly independent and therefore do not form a basis. In such a case, the subspace generated by the L optimal vectors Ψ ^{j (L)} _K is of dimension less than L.

In FIG. 9, the projection of the vector G is represented on the subspace generated by the optimal vectors of rank I, respectively 1-1, this projection being optimal when the aforementioned vectors are orthogonal.

It is therefore particularly advantageous to orthogonalize the stochastic dictionary of rank 1 with respect to the optimal vector of the stage of previous rank Ψ ^{j (I-1)} _K.

Thus, whatever the optimal vector of rank 1 from the new dictionary or stage of corresponding rank I, it will be orthogonal to the optimal vector Ψ ^{j (I-1)} _K of prior rank, and we obtain:

correspond à l'énergie de l'onde sélectionnée à l'étape I,

représente l'intercorrélation des vecteurs optimaux de rang j et de rang j(I) et

représente la matrice d'orthogonalisation.In this relation, we indicate that:

corresponds to the energy of the wave selected in step I,

represents the intercorrelation of the optimal vectors of rank j and of rank j (I) and

represents the orthogonalization matrix.

The previous operation makes it possible to remove from the dictionary the contribution of the previously selected wave and thus imposes linear independence for any optimal vector of rank i comprised between 1 + 1 and L with respect to the optimal vectors of lower rank.

Diagrams of the principle of vector quantization by orthogonal progressive modeling are given in FIGS. 10a and 10b depending on whether there are one or more stochastic dictionaries.

In order to reduce the complexity of the vector quantization process, it is indicated that the recursive modified Gram-Schmidt algorithm can be used as proposed by N. Moreau, P.Dymarski, A.Vigier, in the publication entitled: "Optimal and Suboptimal Algorithms for Selecting the Excitation in Linear Predictive Products ", Proc. ICASSP 90, pp 485-488.

Given the orthogonalization properties, we show that:

Given this expression, the modified recursive Gram-Schmidt algorithm as proposed above can be used.

It is then no longer necessary to explicitly recalculate the dictionaries at each stage of the orthogonalization.

The above calculation process can be explained in matrix form from the matrix

We indicate that Q is an orthonormal matrix, and R is an upper triangular matrix whose elements of the main diagonal are all positive, which ensures the uniqueness of the decomposition.

ce qui implique que Rθ = θ̃.The gain vector G verifies the matrix relation:

which implies that Rθ = θ̃.

The upper triangular matrix R thus makes it possible to recursively calculate the gains 0 (k) relative to the original base.

The contribution of the optimal vectors to the orthonormal base, noted: {Ψ ^{j (I)} _{orth (L)} } in the modeling of the gain vector G _K tends to decrease, and the gains {θ̃ ₁ } are ordered in decreasing order. The residue can be modeled gradually in the following way where θ̂ ^cod _k denotes the gains associated with the optimal orthogonal vector Ψ ^{j (k)} _{orth (k)} quantized, taking into account the relationships:

We then obtain the gain vectors Ĝ ¹ , Ĝ ² , Ĝ ³ orthogonal whose contribution in the modeling of the gain vector G is decreasing, which allows the gradual modeling of the residue r _n effectively. The parameters transmitted by the coding system object of the invention for the modeling of the gain vector G are then the indices j (I) of the selected vectors as well as the numbers i (l) of the ranges of quantification of their associated gains , θ⃝ ₁ . The data transmission is then done by overwriting the parts of the frame allocated to the indices and track numbers j (I), i (I), for 1 ∈ [L1, L2-1] and [L2, L] as required. of communication.

Plusieurs solutions peuvent être utilisées pour coder les rapports précités.The previously mentioned processing process uses the recursive modified Gram-Schmidt algorithm in order to code the gain vector G. The parameters transmitted by the coding system according to the invention being the aforementioned indices, j (0) to j (L ) of the different dictionaries as well as the quantified gains g (0) and {θ⃝ _k }, it is necessary to code the various aforementioned gains g (0) and {θ⃝ _k }. A study has shown that the gains relative to the orthogonal base {Ψ ^{j (I)} _{orth (L)} } being decorrelated, these have good properties for their quantification. Furthermore, since the contribution of the optimal vectors in the modeling of the gain vectors G tends to decrease, the gains {θ⃝ ¹ } are ordered in a relatively decreasing order, and it is possible to use this property by coding not the aforementioned gains but their ratio given by

Several solutions can be used to code the aforementioned reports.

As will be seen in FIG. 2, the coding device which is the subject of the present invention comprises a module for modeling the excitation of the synthesis filter corresponding to the lowest bit rate, this module being noted 17 in the aforementioned figure.

The principle diagram for calculating the excitation signal of the synthesis filter corresponding to the lowest bit rate is given in FIG. 11. An inverse transformation is applied to the gain vectors modeled G ¹ , this inverse adaptive transformation can for example correspond to a reverse transformation of the Householder type, which will be described later in the description, in conjunction with the decoding device which is the subject of the present invention. The signal obtained after inverse adaptive transformation is added to the long-term prediction signal B ' ¹ _n by means of a summator 171, the estimated perceptual signal or long-term prediction signal being delivered by the long-term prediction circuit 13 in closed loop. The resulting signal delivered by the adder 171 is filtered by a filter 172, which corresponds from the point of view of the transfer function to the filter 131 of FIG. 3. The filter 172 delivers the residual signal modeled r̂ ¹ _n .

A system for adaptive transform predictive decoding with nested codes of a coded digital signal constituted by a coded speech signal, and if necessary, by an auxiliary data signal inserted into the coded speech signal after coding of the latter will now be described. in conjunction with Figure 12.

According to the aforementioned figure, the decoding system comprises a circuit 20 for extracting the data signal allowing on the one hand the extraction of the data for an auxiliary use, by an output of the auxiliary data and, on the other hand , the transmission of indexes representative of the coded speech signal. It is of course understood that the aforementioned indexes are the indices i (l) and j (I), for 1 between 0 and L ₁ -1 previously described in the description and for I between I ₁ and L under the conditions which will be described below. As shown in addition in FIG. 12, the decoding system according to the invention comprises a circuit 21 for modeling the speech signal at the minimum bit rate, as well as a circuit 22 or 23 for modeling the speech signal at at least one flow greater than the minimum flow above.

In a preferred embodiment, as shown in FIG. 12, the decoding system according to the invention comprises, in addition to the data extraction system 20, a first module 21 for modeling the speech signal at the minimum bit rate directly receiving the coded signal and delivering a first estimated speech signal, denoted S ¹ _n and a second module 22 for modeling the speech signal at an intermediate rate connected to the data extraction system 20 via a switching circuit 27 conditional on the criterion of the actual bit rate allocated to the speech signal and delivering a second estimated speech signal, denoted AS _n 2 _.

The decoding system shown in FIG. 12 also includes a third module 23 for modeling the speech signal at a maximum rate, this module being connected to the data extraction system 20 via a circuit 28 for conditional switching on criterion of the actual bit rate allocated to speech and delivering a third estimated speech signal Sn-In addition, a summing circuit 24 receives the first, second and third estimated speech signal, and delivers at its output a resulting estimated speech signal , noted S _n . At the output of the summing circuit 24 are connected in cascade an adaptive filtering circuit 25 receiving the resulting estimated speech signal S _n and delivering a reconstituted estimated speech signal, denoted S ' _n . A digital-to-analog converter 26 may be provided to receive the reconstructed speech signal and to output an audio-frequency reconstituted speech signal.

According to a particularly advantageous characteristic of the decoding device which is the subject of the present invention, each of the modules for modeling the speech signal at a minimum, intermediate and maximum bit rate, that is to say the modules 21, 22 and 23 of the figure 12, includes an inverse adaptive transformation sub-module, followed by an inverse perceptual weighting filter.

The block diagram of the speech signal modeling module at minimum bit rate is given in FIG. 13a.

In general, the decoding system object of the present invention takes into account the constraints imposed by the transmission of data at the level of the coding system and in particular at the level of the adaptive dictionary, as well as the contribution of the past excitation.

The circuit for modeling the speech signal at minimum bit rate 21 is identical to that described for circuit 17 of the coding system according to the invention from an inverse adaptive transformation module similar to module 170 described in relation to FIG. 11. We simply note that in FIG. 13a, we have explained how the perceptual signal P was obtained from the indices {i (0), j (0)}, the modeling order K and the indices i (l), j (l) for 1 = 1 to L1-1.

As regards the inverse adaptive transformation, an advantageous embodiment of this is shown in FIG. 13b. It is indicated that the embodiment represented in FIG. 13b corresponds to a reverse Householder type transform using elements identical to the Householder transform represented in FIG. 7. It is simply indicated that for a perceptual signal delivered by the long-term prediction circuit 13, this signal being denoted ^p1 entering a similar module 140, the signals entering the module 1402, respectively at the level of the multipliers associated with each register, are inverted. The resulting signal delivered by the summator corresponding to the summator 171 of FIG. 11 is filtered by a filter of inverse transfer function of the transfer function of the perceptual weighting matrix and corresponding to the filter 172 of the same FIG. 11.

The modules for modeling the speech signal at the intermediate rate or at the maximum rate, module 22 or 23, are shown in FIGS. 14a and 14b.

Of course, it is possible for reasons of complexity to group the different models of the speech signal corresponding to the other bit rates in a single block as shown in FIG. 14a and 14b. According to the actual bit rate allocated to speech, the gain vectors modeled G ² , G ³ are added, as represented in FIG. 14b, by a summator 220, subjected to the process of inverse adaptive transformation in a module 221 identical to the module 210 of FIG. 13a, then filtered by the inverse weighting filter W - '(z) previously mentioned, this filter being designated by 222, the filtering starting from zero initial conditions, which makes it possible to perform an operation equivalent to multiplication by the inverse matrix W- ¹ , in order to obtain a progressive modeling of the synthesis signal S _n . Note in FIG. 14b the presence of switching devices, which are none other than the switching devices 24 and 28 shown in FIG. 12, which are controlled as a function of the actual bit rate of the data transmitted.

Finally, with regard to the adaptive filter 25, a particularly advantageous embodiment is given in FIG. 15. This adaptive filter makes it possible to improve the perceptual quality of the synthesis signal S _n obtained following the summation by the summator 24. A such a filter includes for example a post-filtering module long-term noted 250, followed by a short-term post-filtering module and an energy control module 252, which is controlled by a scale factor calculation module 253. Thus, the adaptive filter 25 delivers the filtered signal ŜS'n, this signal corresponding to the signal in which the quantization noise introduced by the encoder on the synthesized speech signal has been filtered in the places of the spectrum where this is possible. It is indicated that the diagram represented in figure 15 corresponds to the publications of JHChen etA.Gersho, "Real Time Vector APC Speech Coding at 4800 Bps with Adaptative Postfiltering", ICASSP 87, Vol.3, pp 2185-2188.

We have thus described a predictive coding system by orthonormal transform with nested codes making it possible to provide new solutions in the field of coders with nested codes. It is indicated that, in general, the coding system which is the subject of the invention allows wideband coding at speech / data rates of 32/0 kbit / s, 24/8 kbit / s and 16/16 kbit / s.

Claims (10)

1. System for predictive coding of a digital signal into a digital signal with nested codes, coded by adaptive transform with nested codes, in which the digital coded signal is constituted by a coded speech signal and, where appropriate, by a signal of auxiliary data, inserted into the coded speech signal after coding of the latter, system of the type comprising a perceptual weighting filter (11) controlled by a short-term prediction loop making it possible to generate a perceptual signal and a prediction circuit to long term delivering an estimated perceptual signal P., this long term prediction circuit forming a long term prediction loop making it possible to deliver, from the perceptual signal and the estimated past excitation signal, a modeled perceptual excitation signal , and means of adaptive transform and quantization allowing from the perceptual excitation signal to generate the signal d coded speech, characterized in that the perceptual weighting filter consists of a short-term prediction filter of the speech signal to be coded, so as to achieve a frequency distribution of the quantization noise, and in that it comprises means (12) subtracting the contribution of the excitation signal P _n ^o passed from said perceptual signal to deliver an updated perceptual signal P _n , and in that the long-term prediction circuit is formed, in closed loop, from d '' a dictionary updated by the past excitation modeled corresponding to the lowest bit rate making it possible to deliver an optimal waveform and an estimated gain associated with it, constitutive of the estimated perceptual signal, and in that the transform means are formed by an orthonormal transform module comprising an adaptive orthogonal transformation module and a progressive modeling module by orthogonal vectors, these progressive modeling means and the long-term prediction circuit making it possible to deliver indexes representative of the coded speech signal, said system further comprising means (19) for inserting auxiliary data coupled to the transmission channel. 2. Coding system according to claim 1, characterized in that said adaptive orthogonal transformation module comprises: - a filter performing a linear combination of the basic vectors, obtained from a decomposition into singular values of the matrix representative of the perceptual weighting filter. 3. Coding system according to claim 2, characterized in that said filter comprises for any matrix W representative of the perceptual weighting filter: - a first matrix module U = (U ₁ , ..., U _N ) and - a second matrix module V = (V ₁ , ..., V _N ), said first and second matrix modules verifying the relation:

in which U ^T denotes the matrix module transposed from the module U, and where D is a diagonal matrix module whose coefficients constitute the said singular values, U and V _j denoting respectively the i ^th singular vector _me left and jth singular right vector, the said right singular vectors {V _j} forming an orthonormal basis, which allows transform the filtering operation by convolution product by a filtering operation by a linear combination. 4. Coding system according to claim 1, characterized in that said orthonormal transform module is formed by: - a stochastic transform submodule constructed by drawing a Gaussian random variable, for initialization, - a global averaging module on a plurality of vectors from a transform predictive coder, - a reordering module, - a Gram-Schmidt processing module, a reiteration of the processing by the previous modules making it possible to obtain an orthonormal transform, carried out offline, formed by learning, - a read-only memory, allowing the orthonormal transform to be stored, in the form of transform vectors. 5. Coding system according to claim 4, characterized in that said transform is formed by orthonormal waveforms whose frequency spectra are bandpass and relatively ordered, the first waveform of orthonormal waveforms relatively ordered being equal to the normalized optimal waveform from said adaptive dictionary and the first component of estimated gain is equal to the normalized long-term prediction gain. 6. Coding system according to claim 2 and 5, characterized in that said adaptive transformation module comprises: - a Householder transformation module receiving said perceptual signal estimated p; constituted by said optimal waveform and by said estimated gain, and said perceptual signal to generate a transformed perceptual signal P "in the form of a transformed perceptual signal vector with component P" _k , and a plurality of N registers for memorizing said orthonormal waveforms, said plurality of registers forming said read-only memory, each register of rank r comprising N memory cells, a component of rank k of each vector being stored in a cell of rank corresponding, - A plurality of N multiplier circuits associated with each register forming said plurality of storage registers, each rank k multiplier circuit receiving, on the one hand, the rank k component of the stored vector, and, on the other hand, the component P " _k of the transformed perceptual signal vector of rank k, and delivering the product P" _k .f ^k _orth (k) of the components of transformed perceptual signal vector, a plurality of N-1 summing circuits associated with each register of rank r, each summing circuit of rank k receiving the product of prior rank k-1 delivered by the multiplier circuit of prior rank and the product of corresponding rank k delivered by the multiplier circuit of previous rank and the product of corresponding rank k delivered by the multiplier circuit of same rank k, the summing circuit of highest rank, N-1, delivering a component g (r) of the estimated gain, expressed in the form of gain vector G. 7. System according to claim 1, characterized in that said progressive modeling module by orthogonal vector comprises: a gain vector normalization module for generating a normalized gain vector Gk, by comparison of the normalized value of the gain vector G with respect to a threshold value, said normalization module making it possible to generate a signal of length of the normalized gain vector Gk, towards the decoder system as a function of the modeling order, a progressive modeling stage by orthogonal vectors proper receiving said normalized vector Gk and delivering said indexes representative of the coded speech signal, said indexes being representative of the selected vectors and their associated gains, the transmission of auxiliary data formed by the indexes being performed by overwriting the parts of the frame allocated to the indices and track numbers to form the auxiliary data signal. 8. Adaptive transform predictive decoding system of a nested code coded digital signal in which the coded digital signal consists of a coded speech signal and, where appropriate, of an auxiliary data signal inserted into the coded speech signal after coding the latter, characterized in that it comprises: means for extracting said data signal allowing, on the one hand, the extraction of said data for an auxiliary use, and on the other hand, the transmission of indexes representative of the coded speech signal, - means for modeling the speech signal at minimum bit rate, means for modeling the speech signal at at least one bit rate greater than the minimum bit rate, 9. Decoding system according to claim 8, characterized in that this decoder comprises, in addition to the data extraction system, a first module for modeling the speech signal at the minimum rate, directly receiving the coded signal and delivering a first speech signal, estimated S., a second module for modeling the speech signal at an intermediate rate connected to said data extraction system by means of conditional switching means on the criterion of the value of said indexes, and delivering a second estimated speech signal, S. , a third module for modeling the speech signal at a maximum rate, connected to said data extraction system by means of conditional switching means on the criterion of the value of said indexes and delivering a third estimated speech signal, S3 a summing circuit receiving on its summing inputs the first, the second respectively the third estimated speech signal and delivering at its output a resulting estimated speech signal, and connected in cascade at the output of said summing circuit, - an adaptive filtering circuit receiving said resulting estimated speech signal and delivering a reconstituted estimated speech signal, and a digital to analog converter receiving said reconstructed estimated speech signal and delivering an audio-frequency reconstituted speech signal. 10. Decoding system according to claim 9, characterized in that each of the modules for modeling the speech signal at a minimum, intermediate or maximum rate comprises a sub-module of adaptive inverse transformation followed by a filter of inverse perceptual weighting.

Images