EP2652735B1 - Improved encoding of an improvement stage in a hierarchical encoder - Google Patents

Description

The present invention relates to the field of coding digital signals.

The coding according to the invention is particularly suitable for the transmission and / or storage of digital signals such as audio-frequency signals (speech, music or other).

The present invention relates more particularly to the coding of waveforms such as the coding MIC (for "Coded Pulse Modulation") said PCM (for "Pulse Code Modulation") in English, or the adaptive coding of waveform of the ADPCM encoding type (for "Adaptive Differential Pulse Code Modulation" ) in which the invention relates in particular to embedded code coding for issuing indexes of Scalable bit stream quantization.

The general principle of nested code ADPCM coding / decoding specified by ITU-T Recommendation G.722 or ITU-T G.727 is as described with reference to figures 1 and 2 .

The figure 1 thus represents an encoder with nested codes of the ADPCM type (eg G.722 low band, G.727) operating between B and B + K bits per sample; note that the case of a non-scalable ADPCM encoding (eg G.726, G.722 high band) corresponds to K = 0, where B is a fixed value which can be chosen from among different possible rates.

• un module de prédiction 110 permettant de donner la prédiction du signal $x_P^B\left(n\right)$
  
  à partir des échantillons précédents du signal d'erreur quantifié $e_Q^B\left(\mathit{nʹ}\right)=y_{I^B}^B\left(\mathit{nʹ}\right)v\left(\mathit{nʹ}\right)\mathit{nʹ}=n-1,\dots ,n-N_Z$
  
  , où v(n') est le facteur d'échelle de quantification, et du signal reconstitué r^B (n') n' = n -1,...,n-N_p où n est l'instant courant.
• un module de soustraction 120 qui retranche du signal d'entrée x(n) sa prédiction $x_P^B\left(n\right)$
  
  (n) pour obtenir un signal d'erreur de prédiction noté e(n).
• un module de quantification 130 Q^B+K du signal d'erreur qui reçoit en entrée le signal d'erreur e(n) pour donner des indices de quantification I ^{B+ K} (n) constitués de B+K bits. Le module de quantification Q ^B+K est à codes imbriqués c'est-à-dire qu'il comporte un quantificateur de « coeur » à B bits et des quantificateurs à B + k k =1,...,K bits qui sont imbriqués sur le quantificateur de « coeur ».

It comprises:

• a prediction module 110 making it possible to give the prediction of the signal $x_P^B\left(\mathrm{not}\right)$
  
  from the previous samples of the quantized error signal $e_Q^B\left(\mathit{not}\right)={\mathrm{there}}_{I^B}^B\left(\mathit{not}\right)v\left(\mathit{not}\right)\mathit{not}=\mathrm{not}-1,...,\mathrm{not}-{\mathrm{NOT}}_Z$
  
  , where v ( n ' ) is the quantization scale factor, and the reconstructed signal r ^B ( n' ) n '= n - 1, ..., nN _p where n is the current time.
• a subtraction module 120 which subtracts from the input signal x (n) its prediction $x_P^B\left(\mathrm{not}\right)$
  
  (n) to obtain a prediction error signal denoted e ( n ) .
• a quantization module 130 Q ^{B + K} of the error signal which receives as input the error signal e (n) to give quantization indices I ^{B + K} (n) consisting of B + K bits. The quantification module Q ^{B + K} is with nested codes, that is to say that it comprises a quantizer of "heart" with B bits and quantifiers with B + kk = 1 , ..., K bits which are nested on the quantizer of "heart".

In the case of ITU-T G.722 low band coding, the decision levels and quantizer reconstruction levels Q ^B , Q ^{B +1} , Q ^{B + 2} for B = 4 and K = 0 , 1 or 2 are defined in Tables IV and VI of the review article describing G.722 X. Master "7 kHz audio coding within 64 kbit / s." IEEE Journal on Selected Areas in Communication, Vol.6, no. 2, February 1988 .

The quantization index I ^{B + K} ( n ) of B + K bits at the output of the quantization module Q ^{B + K} is transmitted via the transmission channel 140 to the decoder as described with reference to FIG. figure 2 .

• un module 150 de suppression des K bits de poids faible de l'indice I ^{B+ K} (n) pour donner un indice bas débit I^B (n) sur B bits;
• un module de quantification inverse 121 (Q^B )^-1 pour donner en sortie un signal d'erreur quantifié $e_Q^B\left(\mathit{nʹ}\right)=y_{I^B}^B\left(\mathit{n}\right)v\left(\mathit{n}\right)$
  
  sur B bits;
• un module d'adaptation 170 Q_Adapt des quantificateurs et des quantificateurs inverses pour donner un paramètre de contrôle de niveau v(n) encore appelé facteur d'échelle, pour l'instant suivant;
• un module d'addition 180 de la prédiction $x_P^B\left(n\right)$
  
  au signal d'erreur quantifié pour donner le signal reconstruit à bas débit r^B (n) ;
• un module d'adaptation 190 P_Adapt du module de prédiction à partir du signal d'erreur quantifié sur B bits $e_Q^B\left(\mathit{n}\right)$
  
  et du signal $e_Q^B\left(\mathit{n}\right)$
  
  filtré par 1+P_z (z).

The encoder also includes:

• a module 150 for suppressing the K least significant bits of the index I ^{B + K} (n) to give a low bit rate index I ^B ( n ) on B bits;
• an inverse quantization module 121 ( Q ^B ) ^-1 for outputting a quantized error signal $e_Q^B\left(\mathit{not}\right)={\mathrm{there}}_{I^B}^B\left(\mathit{not}\right)v\left(\mathit{not}\right)$
  
  on B bits;
• an adaptation module 170 Q _Adapt quantizers and inverse quantizers to give a level control parameter v (n) still called scale factor, for the following moment;
• an addition module 180 of the prediction $x_P^B\left(\mathrm{not}\right)$
  
  the quantized error signal to give the reconstructed low rate signal r ^B (n);
• an adaptation module 190 P _Adapt of the prediction module from the quantized error signal on B bits $e_Q^B\left(\mathit{not}\right)$
  
  and signal $e_Q^B\left(\mathit{not}\right)$
  
  filtered by 1+ P _z ( z ).

We can notice that on the figure 1 the dashed portion referenced 155 represents the low rate local decoder which contains the predictors 165 and 175 and the inverse quantizer 121. This local decoder thus makes it possible to adapt the inverse quantizer at 170 from the low bit rate index I ^B ( n ) and to adapt the predictors 165 and 175 from the reconstructed low bit rate data.

This part is found identically on the decoder ADPCM nested codes as described with reference to the figure 2 .

Le symbole " ' " indique une valeur décodée à partir des bits reçus, éventuellement différente de celle utilisée par le codeur du fait d'erreurs de transmission.The ADPCM decoder with embedded codes of the figure 2 receives as input the indices I ^{'B + K} coming from the transmission channel 140, version of I ^{B + K} possibly disturbed by binary errors, and performs inverse quantization by the inverse quantization module 210 ( Q ^B ) ^-1 B bit rate per sample to obtain the signal $e_Q^{\mathit{\text{'}B}}\left(\mathit{not}\right)={\mathrm{there}}_{{\mathit{I\text{'}}}^B}^{\mathit{\text{'}B}}\left(\mathit{not}\right)\mathit{V\text{'}}\left(\mathit{not}\right)\mathrm{.}$

The symbol "'" indicates a decoded value from the received bits, possibly different from that used by the encoder due to transmission errors.

The output signal r ' ^B ( n ) for B bits will be equal to the sum of the signal prediction and the output of the B-bit inverse quantizer. This part 255 of the decoder is identical to the local low speed decoder 155 of the figure 1 .

With the mode flow indicator and the selector 220, the decoder can improve the restored signal.

et de la sortie du quantificateur inverse 230 à B+1 bits $y_{{\mathit{I}}^{B+1}}^{\mathit{ʹB}+1}\left(\mathit{n}\right)\mathit{vʹ}\left(\mathit{n}\right)\mathrm{.}$

In fact, if mode indicates that B + 1 bits have been received, the output will be equal to the sum of the prediction $x_P^B\left(\mathrm{not}\right)$

and from the output of the inverse quantizer 230 to B + 1 bits ${\mathrm{there}}_{{\mathit{I}}^{B+1}}^{\mathit{\text{'}B}+1}\left(\mathit{not}\right)\mathit{V\text{'}}\left(\mathit{not}\right)\mathrm{.}$

et de la sortie du quantificateur inverse 240 à B+2 bits $y_{{\mathit{I}}^{B+2}}^{\mathit{ʹB}+2}\left(\mathit{n}\right)\mathit{vʹ}\left(\mathit{n}\right)\mathrm{.}$

If mode indicates that B + 2 bits have been received then the output will be equal to the sum of the prediction $x_P^B\left(\mathrm{not}\right)$

and the output of the inverse quantizer 240 to B + 2 bits ${\mathrm{there}}_{{\mathit{I}}^{B+2}}^{\mathit{\text{'}B}+2}\left(\mathit{not}\right)\mathit{V\text{'}}\left(\mathit{not}\right)\mathrm{.}$

en définissant le bruit de quantification à B+k bits Q ^{B+ k} (z) par : $$Q^{B+k}\left(z\right)=E_Q^{B+k}\left(z\right)-E\left(z\right)$$

By using the notations of the transform in z, one can write that in this looped structure: $$R^{B+k}\left(z\right)=X\left(Z\right)+Q^{B+k}\left(z\right)$$

defining the quantization noise at B + k bits Q ^{B + k} (z) by: $$Q^{B+k}\left(z\right)=E_Q^{B+k}\left(z\right)-E\left(z\right)$$

ITU-T G.722 nested code ADPCM (hereinafter referred to as G.722) coding broadband signals which are defined with a minimum bandwidth of [50-7000 Hz] and sampled at 16 kHz. The G.722 encoding is an ADPCM coding of each of the two sub-bands of the signal [0-4000 Hz] and [4000-8000 Hz] obtained by decomposition of the signal by quadrature mirror filters. The low band is coded by a 6, 5 and 4 bit nested code ADPCM coding while the high band is coded by a 2 bit ADPCM coder per sample. The total bit rate will be 64, 56 or 48 bit / s depending on the number of bits used for decoding the low band.

This coding was first developed for use in ISDN (Digital Integrated Services Network). It has recently been deployed in high quality voice over IP telephony applications.

For a quantizer with a large number of levels, the quantization noise spectrum will be relatively flat. However, in the frequency zones where the signal has a low energy, the noise may have a comparable level or higher than the signal and is therefore not necessarily masked. It can then become audible in these regions.

Coding noise formatting is therefore necessary. In an encoder such as G.722, coding noise formatting suitable for nested code encoding is furthermore desirable.

In general, the purpose of the formatting of the coding noise is to obtain a quantization noise whose spectral envelope follows the short-term masking threshold; this principle is often simplified so that the noise spectrum follows the signal spectrum approximately, providing a more homogeneous signal-to-noise ratio so that the noise remains inaudible even in the lower energy areas of the signal.

A noise shaping technique for MIC coding (for "Coded Pulse Modulation") with nested codes is described in the Recommendation G.711.1 "Wideband embedded extension for G.711 pulse code modulation" or "G.711.1: A wideband extension to ITU-T G.711". Y. Hiwasaki, S. Sasaki, H. Ohmuro, T. Mori, J. Seong, MS Lee, B. Kvesi, S. Ragot, J.-L. Garcia, C.Marro, LM, J. Xu, V. Malenovsky, J. Lapierre, R. Lefebvre. EUSIPCO, Lausanne, 2008 .

This recommendation thus describes coding with coding noise formatting for heart rate coding. A perceptual filter for shaping the coding noise is calculated based on the decoded past signals from a reverse core quantizer. A local heart rate decoder thus makes it possible to calculate the noise shaping filter. Thus, at the decoder, it is possible to calculate this noise shaping filter from decoded heart rate signals.

A quantizer delivering improvement bits is used at the encoder.

The decoder receiving the core bit stream and the improvement bits, calculates the coding noise shaping filter in the same way as the coder from the decoded heart rate signal and applies this filter to the output signal of the decoder. inverse quantizer of the enhancement bits, the shaped high-speed signal being obtained by adding the filtered signal to the decoded heart signal.

The shaping of the noise thus improves the perceptual quality of the heart rate signal. It offers a limited improvement in quality for improvement bits. Indeed, the formatting of the coding noise is not carried out for the coding of the improvement bits, the input of the quantizer being the same for the quantization of the core as for the improved quantization.

The decoder must then remove a resulting parasitic component by a matched filtering, when the improvement bits are decoded in addition to the core bits.

The additional calculation of a decoder filter increases the complexity of the decoder.

This technique is not used in existing standard scalable decoders of the G.722 or G.727 decoder type. There is therefore a need to improve the quality of the signals regardless of the bit rate while remaining compatible with standard scalable existing decoders.

A solution that does not require a decoder, complementary signal processing is described in the patent application WO 2010/058117 . In this application, the signal received at the decoder can be decoded by a standard decoder capable of decoding the heart rate signal and nested data rates without requiring calculation of noise shaping or correction term.

This document describes that for an improvement stage of a hierarchical coder, quantization is performed by minimizing a quadratic error criterion in a perceptually filtered domain.

For this, a coding noise shaping filter is defined and applied to a given error signal from at least one reconstructed signal of a preceding coding stage. The method also requires the calculation of the reconstructed signal of the current improvement stage in anticipation of a next coding stage.

In addition, improvement terms are calculated and stored for the current improvement stage. This therefore brings significant complexity and significant storage enhancement terms or reconstructed signal samples of previous stages.

This solution is therefore not optimal from a complexity point of view.

There is therefore a need to improve the state of the art methods for encoding and formatting enhancement coding noise, while remaining compatible with existing hierarchical decoders.

The present invention improves the situation.

• obtention de valeurs possibles de quantification pour l'étage d'amélioration courant k à partir des niveaux de reconstruction absolus du seul étage courant k et des indices du codeur imbriqué précédent;
• quantification du signal d'entrée du codeur hiérarchique ayant subi ou non un traitement de pondération perceptuelle, à partir des dites valeurs possibles de quantification pour former un indice de quantification de l'étage k et un signal quantifié correspondant à une des valeurs possibles de quantification.

To this end, it proposes a method for encoding an input digital audio signal (x (n)) in a hierarchical coder comprising a B-bit core coding stage and at least one current improvement coding stage k , the core coding and the coding of the improvement stages preceding the current stage k delivering quantization indices which are concatenated to form the indices of the preceding nested encoder (I ^{B + k-1} ). The method is such that it comprises the following steps:

• obtaining possible quantization values for the current improvement stage k from the absolute reconstruction levels of the single current stage k and the indices of the preceding nested encoder;
• quantization of the input signal of the hierarchical coder which has or has not undergone a perceptual weighting treatment, from the said possible values of quantification to form a quantization index of the stage k and a quantized signal corresponding to one of the possible quantization values.

Thus, the quantization of the improvement stage determines the quantization index bit or bits which are directly concatenated with the indices of the preceding stages. Unlike the state-of-the-art methods, there is no computation of an improvement signal or improvement terms.

In addition, the input signal of the quantization is either directly the input signal of the hierarchical coder, or the same input signal having directly undergone perceptual weighting processing. This is not a signal difference between the input signal and a reconstructed signal of the previous coding stages as in the techniques of the state of the art.

The complexity in terms of computing load is therefore reduced.

In addition, unlike state-of-the-art methods, stored quantization values are not differential values. Thus, it is not useful to memorize the quantization values used for reconstruction in the previous stages to form a quantization dictionary of the improvement stage.

Ainsi l'invention évite la duplication des dictionnaires que l'on peut rencontrer dans les méthodes de l'état de l'art où un dictionnaire différentiel est utilisé au codeur et un dictionnaire absolu au décodeur.On the other hand, unlike state-of-the-art methods, it is not necessary to construct and store a differential dictionary, since the enhancement stage directly uses absolute levels stored by the encoder and hierarchical decoder. existing $\left({\mathrm{there}}_i^{B+k}\left(\mathrm{not}\right)\right)\mathrm{.}$

Thus, the invention avoids the duplication of the dictionaries that can be encountered in the methods of the state of the art where a differential dictionary is used at the encoder and an absolute dictionary at the decoder.

The memory required for the storage of the dictionaries and the quantification operations at the encoder and inverse quantization at the decoder is therefore reduced.

Finally, the fact of directly obtaining the quantization values of the improvement stage without making any difference brings an additional precision between the values obtained at the encoder and those obtained at the decoder when working, for example, in finite precision.

The various particular embodiments mentioned below may be added independently or in combination with each other, to the steps of the method defined above.

In a particular embodiment, the input signal has undergone perceptual weighting processing using a predetermined weighting filter to provide a modified input signal, prior to the quantization step, and the method further includes a step of adapting the weighting filter memories from the quantized signal of the current enhancement coding stage.

This perceptual weighting processing applied directly to the input signal of the hierarchical coder for the enhancement coding of the stage k also reduces the complexity in terms of computational load compared to state-of-the-art techniques which performed this perceptual weighting processing on a difference signal between the input signal and a reconstructed signal of the previous coding stages.

Thus, the encoding method described also allows existing decoders to decode the signal without having to make any additional modifications or processing to be expected while benefiting from the improvement of the signal by formatting the effective coding noise.

In a particular embodiment, the possible quantization values for the improvement stage k further contain a scale factor and a prediction value from the adaptive type core coding.

This makes it possible to adapt the quantization values with respect to the values defined in the core coding.

In an alternative embodiment, the modified input signal to be quantized at the improvement stage k is the perceptually weighted input signal from which a prediction value derived from the adaptive type core coding is subtracted.

This also makes it possible to adapt the quantization values with respect to the values defined in the core coding but by making this input adaptation of the quantizer rather than on each quantization value. This is advantageous in the case where the improvement is carried out on several bits.

In particular, the perceptual weighting treatment is performed by prediction filters forming an ARMA type filter.

The formatting of the improvement coding noise is then of good quality.

• un module d'obtention de valeurs possibles de quantification pour l'étage d'amélioration courant k par la détermination de niveaux de reconstruction absolus du seul étage courant k à partir des indices du codeur imbriqué précédent;
• un module de quantification du signal d'entrée du codeur hiérarchique ayant subi ou non un traitement de pondération perceptuelle, à partir des dites valeurs possibles de quantification pour former un indice de quantification de l'étage k et un signal quantifié correspondant à une des valeurs possibles de quantification.

The present invention also relates to a hierarchical coder of an input digital audio signal, comprising a B-bit core coding stage and at least one current improvement coding stage k, the core coding and the coding of the stages. of improvement preceding the current stage k delivering quantization indices which are concatenated to form the indices of the preceding nested encoder. The encoder is such that it comprises:

• a module for obtaining possible quantization values for the current improvement stage k by determining absolute reconstruction levels of the single current stage k from the indices of the preceding nested encoder;
• a quantization module of the input signal of the hierarchical coder which has or has not undergone perceptual weighting processing, from said possible quantization values to form a quantization index of the stage k and a quantized signal corresponding to one of the values possible quantification.

The hierarchical coder further comprises a perceptual weighting pre-processing module using a predetermined weighting filter to give a modified input signal of the quantization module and a weighting filter memory adaptation module from the quantized signal. of the current improvement coding stage.

The hierarchical coder provides the same advantages as those of the method it implements.

It also relates to a computer program comprising code instructions for implementing the steps of the encoding method according to the invention, when these instructions are executed by a processor.

The invention finally relates to a storage means readable by a processor storing a computer program as described.

• la figure 1 illustre un codeur de type MICDA à codes imbriqués selon l'état de l'art et tel que décrit précédemment;
• la figure 2 illustre un décodeur de type MICDA à codes imbriqués selon l'état de l'art et tel que décrit précédemment;
• la figure 3 illustre un mode de réalisation général du procédé de codage selon l'invention et d'un codeur selon l'invention;
• la figure 4 illustre un premier mode de réalisation particulier du procédé de codage et d'un codeur selon l'invention;
• la figure 5 illustre un deuxième mode de réalisation particulier du procédé de codage et d'un codeur selon l'invention;
• la figure 6 illustre un troisième mode de réalisation particulier du procédé de codage et d'un codeur selon l'invention;
• la figure 7 illustre une alternative de réalisation générale du procédé de codage et d'un codeur selon l'invention;
• la figure 7b illustre une autre alternative de réalisation générale du procédé de codage et d'un codeur selon l'invention;
• la figure 8 illustre un exemple de réalisation du codage coeur d'un codeur selon l'invention;
• la figure 9 illustre un exemple de niveaux de reconstruction de quantification utilisés dans l'état de l'art; et
• la figure 10 illustre un mode de réalisation matérielle d'un codeur 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 coder of the ADPCM type with nested codes according to the state of the art and as previously described;
• the figure 2 illustrates a decoder of the ADPCM type with nested codes according to the state of the art and as described previously;
• the figure 3 illustrates a general embodiment of the coding method according to the invention and an encoder according to the invention;
• the figure 4 illustrates a first particular embodiment of the coding method and an encoder according to the invention;
• the figure 5 illustrates a second particular embodiment of the coding method and an encoder according to the invention;
• the figure 6 illustrates a third particular embodiment of the coding method and an encoder according to the invention;
• the figure 7 illustrates an alternative general embodiment of the coding method and an encoder according to the invention;
• the figure 7b illustrates another alternative general embodiment of the coding method and an encoder according to the invention;
• the figure 8 illustrates an exemplary embodiment of the core coding of an encoder according to the invention;
• the figure 9 illustrates an example of quantization reconstruction levels used in the state of the art; and
• the figure 10 illustrates a hardware embodiment of an encoder according to the invention.

With reference to the figure 3 , an encoder as well as a coding method according to an embodiment of the invention is described.

It will be recalled here that the case of an encoder with nested codes or hierarchical encoder is considered in which a B-bit core coding and at least one rank improvement stage k is provided. The core coding and the improvement stages preceding the coding of the improvement stage k as represented at 306, deliver multiplexed scalar quantization indices in the index I ^{B + k-1} (n) of B + k. -1 bits per sample.

In the exemplary embodiments described below, for the sake of simplification of presentation, the improvement stage (of rank k) is presented as producing one additional bit per sample. In this case, the coding in each improvement stage involves selecting one of two possible values. As it will appear later, the "absolute dictionary" - in terms of absolute levels (in the sense of "non-differential") - corresponding to all the quantization values that can be produced by the rank improvement stage k, is of size 2 ^{B + k} , or sometimes slightly less than 2 ^{B + k} as for example in the G.722 coder which has only 60 possible levels instead of 64 in the quantizer of 6 bits of low band. Hierarchical coding implies a binary tree structure of the "absolute dictionary", which explains why it suffices to have an improvement bit to perform the coding given the B + k-1 bits of the preceding stages.

The figure 9 is an excerpt from Table VI of the aforementioned article X. Master and represents the first 4 levels of the B-bit core quantizer for B = 4 bits and the quantizer levels at B + 1 and B + 2 bits of the coding of the low band of a G.722 encoder as well as the output values of the state-of-the-art enhancement quantizer for B + 2 bits.

As illustrated in this figure, the quantizer nested at B + 1 = 5 bits is obtained by "splitting" the quantizer levels at B = 4 bits. The quantizer nested at B + 2 = 6 bits is obtained by "splitting" the quantizer levels at B + 1 = 5 bits. The duplication of the reconstruction levels is in fact a consequence of the low band hierarchical coding constraint which is implemented in G.722 in the form of a scalar quantization dictionary (at 4, 5 or 6 bits per sample ) structured in a tree.

désignant des niveaux de reconstructions de quantification pour un étage d'amélioration k sont définis par la différence entre

• oles valeurs désignant les niveaux de reconstruction de la quantification d'un quantificateur imbriqué à B+k bits (B désignant le nombre de bits du codage coeur) et
• o les valeurs désignant les niveaux de reconstruction de quantification d'un quantificateur imbriqué à B+k-1 bits, les niveaux de reconstruction du quantificateur imbriqué à B+k bits étant définis par dédoublement des niveaux de reconstruction du quantificateur imbriqué à B+k-1 bits.

Avec l'invention les niveaux de reconstruction différentiels ${\mathit{enh}}_{}^{}$${\mathit{}}_{2I^{B+k-1}+j}^{B+k}$

listés à droite et encadrés en pointillés n'ont pas à être calculés ni stockés. Selon l'invention seuls les niveaux de reconstruction absolus $y_i^{B+k}$

de l'étage k sont calculés et stockés.In the state of art, the values ${\mathit{enh}}_{2I^{B+k-1}+j}^{B+k}$

designating levels of quantization reconstructions for an improvement stage k are defined by the difference between

• oles denoting the quantization reconstruction levels of a quantized quantizer at B + k bits (where B denotes the number of bits of the core coding) and
• o the values designating the quantization reconstruction levels of a nested quantizer at B + k-1 bits, the reconstruction levels of the Nested quantizer at B + k bits being defined by splitting the reconstruction levels of the nested quantizer at B + k-1 bits.

With the invention the differential reconstruction levels ${\mathit{enh}}_{2I^{B+k-1}+j}^{B+k}$

listed on the right and dotted boxes do not have to be calculated or stored. According to the invention only the absolute reconstruction levels ${\mathrm{there}}_i^{B+k}$

of the stage k are calculated and stored.

de l'étage k sont utilisés au codeur de la même façon qu'au décodeur, dans le sens où le signal reconstruit peut être obtenu dans le cas général du codage MICDA à partir de ces niveaux de reconstruction absolus $y_i^{B+k}$

par multiplication par le facteur d'échelle v(n) et ajout du signal de prédiction $x_P^B\left(n\right),$

comme déjà présenté en référence à la description de la Figure 2 qui représente le décodeur MICDA standard à codes imbriqués. Ces niveaux étant déjà définis et stockés dans le décodeur, le codeur ne rajoute donc aucune table de quantification supplémentaire dans le codec (codeur + décodeur).These absolute reconstruction levels ${\mathrm{there}}_i^{B+k}$

of the stage k are used at the encoder in the same way as at the decoder, in the sense that the reconstructed signal can be obtained in the general case of the ADPCM coding from these absolute reconstruction levels ${\mathrm{there}}_i^{B+k}$

by multiplication by the scale factor v (n) and addition of the prediction signal $x_P^B\left(\mathrm{not}\right),$

as already presented with reference to the description of the Figure 2 which represents the standard nested code ADPCM decoder. Since these levels are already defined and stored in the decoder, the encoder does not add any additional quantization tables in the codec (encoder + decoder).

The coding of the improvement stage according to the invention is very easily generalizable for cases where the improvement stage adds several bits per sample. In this case, the size of the dictionary D _k (n) used in the improvement stage, as defined later, is simply 2 ^U where U> 1 is the number of bits per sample of the improvement stage.

The encoder as represented in figure 3 shows a nested coder or hierarchical coder in which a B-bit core coding and at least one rank improvement stage k is provided. The core coding and the improvement stages preceding the coding of the improvement stage k as represented at 306, deliver scalar quantization indices which are concatenated to form the indices of the preceding nested encoder I ^{B + k-1} ( not).

The figure 3 simply illustrates a PCM / ADPCM coding module 302 representing the nested coding preceding the enhancement coding at 306.

The core encoding of the preceding nested encoding can optionally be performed using the masking filter determined at 301 to format "core" coding noise. An example of this type of core coding is described later with reference to the figure 8 .

et le facteur d'échelle v(n) dans le cas où il s'agit bien d'un codage prédictif MICDA similaire à celui décrit en référence à la figure 1.This module 302 thus delivers the indices I ^{B + k-1} (n) of the nested encoder as well as the prediction signal $x_P^B\left(\mathrm{not}\right)$

and the scale factor v (n) in the case where it is indeed a predictive coding ADPCM similar to that described with reference to the figure 1 .

et v(n) =1.In the case of a PCM coding, the module 302 simply delivers the nested quantization indices I ^{B + k-1} (n). Moreover, it can be noticed that the MIC coding is a particular case of the ADPCM coding taking $x_P^B\left(\mathrm{not}\right)=0$

and v ( n ) = 1.

, ainsi que le cas échéant, le signal de prédiction et $y_i^{B+k}$

le facteur d'échelle v(n) permettent de déterminer les valeurs de quantification D_k(n) = $\left\{d_1^{B+k}\left(n\right),d_2^{B+k}\left(n\right)\right\}$

pour l'étage d'amélioration courant k dans le module de construction du dictionnaire des valeurs de quantification 303. Ce dictionnaire D_k(n) est utilisé par le quantificateur qualifié ici de "quantificateur d'amélioration" pour l'étage d'amélioration de rang k.The knowledge of nested quantification indices I ^{B + k-1} (n) and absolute reconstruction levels ${\mathrm{there}}_i^{B+k}$

, as well as where appropriate, the prediction signal and ${\mathrm{there}}_i^{B+k}$

the scaling factor v (n) makes it possible to determine the quantization values D _k (n) = $\left\{d_1^{B+k}\left(\mathrm{not}\right),d_2^{B+k}\left(\mathrm{not}\right)\right\}$

for the current improvement stage k in the construction module of the quantization value dictionary 303. This dictionary D _k (n) is used by the quantizer described here as an "improvement quantizer" for the improvement stage of rank k.

où $y_{2I^{B+k-1}+j}^{B+k},$

avec j= 0 ou 1, représentent deux valeurs possibles de quantification d'un quantificateur imbriqués de B+k bits, prédéfinies et stockées au codeur et au décodeur. On peut voir les valeurs $y_i^{B+k}$

comme issues d'un « dédoublement » du dictionnaire $y_i^{B+k-1}$

de l'étage précédent k-1.Thus, according to the preferred embodiment, the quantization values of the dictionary are defined as follows, in the case of the ADPCM coding: $$d_1^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right)\mathrm{and}{d}_2^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right),$$

or ${\mathrm{there}}_{2I^{B+k-1}+j}^{B+k},$

with j = 0 or 1, represent two possible quantification values of a nested quantizer of B + k bits, predefined and stored at the encoder and the decoder. We can see the values ${\mathrm{there}}_i^{B+k}$

as issues of a "splitting" of the dictionary ${\mathrm{there}}_i^{B+k-1}$

from the previous floor k-1.

Le "dictionnaire absolu" est un dictionnaire structuré en arbre. L'indice I ^B+k-1 conditionne les différentes branches de l'arbre à prendre en compte pour déterminer les valeurs de quantification possibles de l'étage k (D_k(n)).Note that the two elements of the dictionary D _k (n) depend on I ^{B + k- 1} . In fact, this dictionary is a subset of the "absolute dictionary" defined as: $${\displaystyle \underset{I^{B+k-1}}{\cup }{D}_k\left(\mathrm{not}\right)=\underset{I^{B+k-1}}{\cup }\left\{x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right),x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right)\right\}}$$

The "absolute dictionary" is a dictionary structured in tree. The index I ^{B + k -1} conditions the different branches of the tree to be taken into account in order to determine the possible quantization values of the stage k (D _k (n)).

The scaling factor v (n) is determined by the core stage of the ADPCM coding as illustrated in FIG. figure 1 , the improvement stage therefore uses this same scale factor to scale the codewords of the quantization dictionary.

In one embodiment of the invention, the coder of the figure 3 does not include the modules 301 and 310, that is to say that there is no provision for encoding noise shaping processing. Thus, it is the input signal x (n) itself that is quantized by the quantization module 306.

In a particular embodiment, the encoder further comprises a module 301 for calculating a masking filter and for determining the weighting filter W (z) or a predictive version W _PRED (z) described later. The masking or weighting filter is determined here from the input signal x (n) but could very well be determined from a decoded signal, for example from the decoded signal of the preceding nested encoder x ^{B + k-1} ( n ) . The masking filter can be determined or adapted sample by sample or by block of samples.

Indeed, the encoder according to the invention performs a shaping of the coding noise of the improvement stage by using a quantization in the domain weighted by the filter W (z), that is to say by minimizing the quantization noise energy filtered by W (z).

This weighting filter is used at 311 by the filtering module and more generally by the perceptual weighting module 310 of the input signal x (n). This pretreatment is applied directly to the input signal x (n) and not to an error signal as could be the case in state-of-the-art techniques.

This pretreatment module 310 delivers a modified signal x '(n) at the input of the enhancement quantizer 307.

et $d_2^{B+k}\left(n\right)$

du dictionnaire adaptatif D_k(n).The quantization module 307 of the improvement stage k delivers a quantization index I in _h ^{B + k} (n) which will be concatenated with the indices of the preceding nested encoding (I ^{B + k-1} ) to form the nested encoding indices. current (I ^{B + k} ), by a module not shown here. The quantization module 307 of the improvement stage k chooses between the two values $d_1^{B+k}\left(\mathrm{not}\right)$

and $d_2^{B+k}\left(\mathrm{not}\right)$

of the adaptive dictionary D _k (n).

soit à $d_2^{B+k}\left(n\right)),$

en minimisant l'erreur quadratique entre x'(n) et x̃ ^B+k (n). Le dictionnaire adaptatif D_k(n) contient donc directement la valeur quantifiée de sortie de l'étage k.It receives as input the signal x '(n) and gives as output, through the local decoding module 308, the quantized value x ^{B + k} ( n ) (where x ^{B + k} ( n ) is equal to $d_1^{B+k}\left(\mathrm{not}\right)$

either to $d_2^{B+k}\left(\mathrm{not}\right)),$

by minimizing the squared error between x '(n) and x ^{B + k} ( n ) . The adaptive dictionary D _k (n) therefore directly contains the quantized output value of the stage k.

Au décodeur la même valeur est obtenue simplement en utilisant directement le quantificateur inverse de l'étage k et l'indice concaténé : ${\tilde{x}}^{B+k}\left(n\right)=x_P^B\left(n\right)+y_{I^{B+k}}^{B+k}v\left(n\right)\mathrm{.}$

The module 308 gives the quantized value of the input signal by inverse quantization of the index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)\mathrm{.}$

At the decoder the same value is obtained simply by directly using the inverse quantizer of the stage k and the concatenated index: ${\tilde{x}}^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{I^{B+k}}^{B+k}v\left(\mathrm{not}\right)\mathrm{.}$

This quantized signal is used to update the memories of the weighting filter W (z) of the enhancement stage to obtain memories corresponding to an input x ( n ) -x ^{B + k} ( n ) . Typically one subtracts from the memory (or memories in the case of the filter type ARMA) more recent (s) the current value of the decoded signal x ^{B + k} ( n ) .

Thus, the quantization of the signal x (n) is done in the weighted domain, which means that we minimize the squared error between x ( n ) and x ^{B + k} ( n ) after filtering by the filter W ( z ). The quantization noise of the enhancement stage is thus shaped by a 1 / W (z) filter to make this noise less audible. The energy of the weighted quantization noise is thus minimized.

The general embodiment of the block 310 given on the figure 3 shows the general case where W (z) is an infinite impulse response (IIR) filter or a finite impulse response (FIR) filter. The signal x ' ( n ) is obtained by filtering x ( n ) by W ( z ) and then when the quantified value x ^{B + k} (n) is known, the memories of the filter W ( z ) are updated as if the filtering had been done on the signal x ( n ) -x ^{B + k} ( n ) .

The dotted arrow represents the update of the filter memories.

• obtention en 303 de valeurs possibles de quantification $\left(d_i^{B+k}\left(n\right)\right)$
  
  pour l'étage d'amélioration courant k par la détermination de niveaux de reconstruction absolus du seul étage courant k à partir des indices du codeur imbriqué précédent (I ^B+k-1);
• quantification en 306 du signal d'entrée du codeur hiérarchique ayant subi ou non un traitement de pondération perceptuelle (x(n) ou x'(n)), à partir des dites valeurs possibles de quantification $\left(d_i^{B+k}\left(n\right)\right)$
  
  pour former un indice de quantification de l'étage k (I_enb ^B+k(n))

et un signal quantifié (x̃ ^B+k (n)) correspondant à une des valeurs possibles de quantification;Thus, the steps implemented in the encoder as illustrated in FIG. figure 3 are also represented. We find in fact the following steps:

• obtaining in 303 possible values of quantification $\left(d_i^{B+k}\left(\mathrm{not}\right)\right)$
  
  for the current improvement stage k by determining absolute reconstruction levels of the single current stage k from the indices of the preceding nested encoder ( I ^{B + k- 1} );
• quantization at 306 of the input signal of the hierarchical coder having undergone or not a perceptual weighting treatment (x (n) or x '(n)), from the said possible values of quantification $\left(d_i^{B+k}\left(\mathrm{not}\right)\right)$
  
  to form a quantization index of the stage k (I _enb ^{B + k} (n))

and a quantized signal ( x ^{B + k} ( n )) corresponding to one of the possible quantization values;

In the case represented in figure 3 the input signal has undergone perceptual weighting processing by using a predetermined weighting filter at 301 to give a modified input signal x '(n), before the quantization step at 306. The figure 3 also represents the step of adapting the weighting filter memories 311 to the quantized signal ( x ^{B + k} ( n )) of the current enhancement coding stage.

The figures 4 , 5 and 6 now describe particular embodiments of the pretreatment block 310.

The blocks 301, 302, 303, 306, 307 and 308 then remain identical to those described with reference to FIG. figure 3 .

The figure 4 represents a first embodiment of the pretreatment block 310 with a finite impulse response (FIR) filter W ( z ) = A ' ( z ).

N_D étant l'ordre du filtre perceptuel W(z).In this embodiment, the filter memory contains only the input samples passed from the signal x ( n ) -x ^{B + k} ( n ) , noted: $$b^{B+k}\left(\mathit{not}\right),\mathit{not}=\mathrm{not}-1,...,\mathrm{not}-{\mathrm{NOT}}_D\mathrm{.}$$

N _D being the order of the perceptual filter W ( z ).

At 302, the input signal x ( n ) is encoded by the MIC / ADPCM coding module 302, with or without shaping the coding noise of the nested encoder B + k-1,

du facteur d'échelle v(n) de l'étage coeur dans le cas d'un codage de type adaptatif MICDA et des indices de codage I ^B+k-1 (n) comme expliqué en référence à la figure 3. Le dictionnaire adaptatif D_k comporte dans le mode de réalisation particulier ou un seul bit d'amélioration est prévu dans l'étage d'amélioration k, les deux termes suivants: $d_1^{B+k}\left(n\right)=x_P^B\left(n\right)+y_{I^{B+k-1}}^{B+k}v\left(n\right)$

et $d_2^{B+k}\left(n\right)=x_P^B\left(n\right)+y_{2I^{B+k-1}+1}^{B+k}v\left(n\right)\mathrm{.}$

In 303, an adaptive dictionary D _k is constructed according to the prediction values $x_P^B\left(\mathrm{not}\right),$

the scaling factor v ( n ) of the heart stage in the case of an ADPCM type adaptive coding and I ^{B + k-1} ( n ) coding indices as explained with reference to FIG. figure 3 . The adaptive dictionary D _k comprises in the particular embodiment or a single improvement bit is provided in the improvement stage k, the following two terms: $d_1^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right)$

and $d_2^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right)\mathrm{.}$

In this embodiment, there are the calculation steps in 301 of the filtering filter and the weighting filter W ( z ), and its predictive version W _PRED ( z ) based on predictions, that is to say -describe calculations using only past samples.

• Prenons comme exemple le cas d'un filtrage d'un signal x(n) par le filtre non récursif de fonction de transfert tout-zéro (aussi dit FIR pour Finite Impulse Response en anglais (Filtre à réponse impulsionnelle finie)) A(z) d'ordre 4, $A\left(z\right)=1+{\displaystyle \sum _{i=1}^4}{a}_i{z}^{-i},$
  
  donnant pour résultat un signal y(n). Dans le domaine de la transformée en z, l'équation $$Y\left(z\right)=A\left(z\right)X\left(z\right)$$
  
  
  correspond à l'équation aux différences $$y\left(n\right)=a_{0}x\left(n\right)+a_{1}x\left(n-1\right)+a_{2}x\left(n-2\right)+a_{3}x\left(n-3\right)+a_{4}x\left(n-4\right)$$
  

Recall here the definition of a predictive filter:

• Let us take as an example the case of a filtering of a signal x (n) by the non-recursive all-zero transfer function filter (also called FIR for Finite Impulse Response in English (Finite impulse response filter)) A ( z ) of order 4, $\mathrm{AT}\left(z\right)=1+{\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{\mathrm{at}}_i{z}^{-i},$
  
  resulting in a signal y ( n ). In the domain of the z transform, the equation $$Y\left(z\right)=\mathrm{AT}\left(z\right)X\left(z\right)$$
  
  
  corresponds to the difference equation $$\mathrm{there}\left(\mathrm{not}\right)={\mathrm{at}}_{0}x\left(\mathrm{not}\right)+{\mathrm{at}}_{1}x\left(\mathrm{not}-1\right)+{\mathrm{at}}_{2}x\left(\mathrm{not}-2\right)+{\mathrm{at}}_{3}x\left(\mathrm{not}-3\right)+{\mathrm{at}}_{4}x\left(\mathrm{not}-4\right)$$
  

• la première ne dépend que de l'entrée présente x(n) : a ₀ x(n). Le plus souvent et dans les cas qui nous intéressent dans ce document, a₀ = 1 .
• la seconde qui ne dépend que l'entrée passée x(n-i), i > 0: a ₁ x(n-1)+a ₂ x(n-2)+a ₃ x(n-3)+a₄x(n-4) qui sera donc considérée comme la partie prédictive du filtrage par analogie à la prédiction linéaire où elle représente la prédiction de x(n) à partir des échantillons précédents.

This expression of y (n) can be divided into two parts:

• the first depends only on the present entry x (n): a ₀ x (n) . Most often and in the cases that interest us in this document, ₀ = 1.
• the second which depends only on the input x (ni), i> 0: a ₁ x ( n- 1) + a ₂ x ( n- 2 ) + a ₃ x ( n- 3) + a ₄ x ( n- 4) which will therefore be considered as the predictive part of filtering by analogy to linear prediction where it represents the prediction of x (n) from the previous samples.

This second part corresponds for the moment of sampling n to the "response to the null entry", or in English "zero input response" (ZIR) or "ringing" which is in fact a generalized prediction. The z- transform of this component is: Y _PRED ( z ) = ( A ( z ) -1 ) X (z ) = H _{A, PRED} ( z ) X ( z ) with H _{A, PRED} ( z ) = A ( z ) - 1

résultant en un signal y(n), la fonction de transfert donne : $$Y\left(z\right)=\frac{1}{B\left(z\right)}X\left(z\right)$$

avec comme équation aux différences : $$y\left(n\right)=x\left(n\right)-b_{1}y\left(n-1\right)-b_{2}y\left(n-2\right)-b_{3}y\left(n-3\right)-b_{4}y\left(n-4\right)$$

Similarly, for the filtering of a signal x (n) by an all-pole recursive filter of order 4 with $B\left(z\right)=1+{\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{b}_i{z}^{-i},$

resulting in a signal y ( n ), the transfer function gives: $$Y\left(z\right)=\frac{1}{B\left(z\right)}X\left(z\right)$$

with as difference equation: $$\mathrm{there}\left(\mathrm{not}\right)=x\left(\mathrm{not}\right)-{\mathrm{<figure-callout\; id="1"\; label="b"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_1\mathrm{there}\left(\mathrm{not}-1\right)-{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_2\mathrm{there}\left(\mathrm{not}-2\right)-b_3\mathrm{there}\left(\mathrm{not}-3\right)-{\mathrm{<figure-callout\; id="4"\; label="b"\; filenames="imgf0010.png"\; state="\{\{state\}\}">b</figure-callout>}}_4\mathrm{there}\left(\mathrm{not}-4\right)$$

The innovation part is x (n), the predictive part is -b ₁ y ( n- 1) -b ₂ y ( n- 2) -b ₃ y ( n- 3) -b ₄ y ( n- 4) , transform to z Y _PRED ( z ) = - ( B ( z ) - 1) Y ( z ) = (1 -B ( z )) Y ( z ) .

avec comme équation aux différences (dans cet exemple A(z) et B(z) sont d'ordre 4) : $$y\left(n\right)=x\left(n\right)+{\displaystyle \sum _{i=1}^4}{a}_{i}x\left(n-i\right)-{\displaystyle \sum _{i=1}^4}{b}_{i}y\left(n-i\right)$$

It is the same for the case where the filter contains both zeros and poles (ARMA filter for AutoRegressive Moving Average): $$Y\left(z\right)=\frac{\mathrm{AT}\left(z\right)}{B\left(z\right)}X\left(z\right)$$

with as difference equation (in this example A ( z ) and B ( z ) are of order 4): $$\mathrm{there}\left(\mathrm{not}\right)=x\left(\mathrm{not}\right)+{\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{\mathrm{at}}_{i}x\left(\mathrm{not}-i\right)-{\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{b}_i\mathrm{there}\left(\mathrm{not}-i\right)$$

de transformée en z Y_PRED (z)=(A(z)-1)X(z)-(B(z)-1)F(z), ou $Y_{\mathit{PREB}}\left(z\right)=\frac{A\left(z\right)-1}{B\left(z\right)}X\left(z\right)=H_{\mathit{AB},\mathit{PRED}}\left(z\right)X\left(z\right)\mathrm{avec}{H}_{\mathit{AB},\mathit{PRED}}\left(z\right)=\frac{A\left(z\right)-1}{B\left(z\right)}\mathrm{.}$

The innovation part is x (n), the predictive part is ${\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{\mathrm{at}}_{i}x\left(\mathrm{not}-i\right)-{\displaystyle \underset{i=1}{\overset{4}{\Sigma }}}{b}_i\mathrm{there}\left(\mathrm{not}-i\right),$

of transform in z Y _PRED ( z ) = ( A ( z ) - 1) X ( z ) - ( B ( z ) - 1) F ( z ) , or $Y_{\mathit{PREB}}\left(z\right)=\frac{\mathrm{AT}\left(z\right)-1}{B\left(z\right)}X\left(z\right)=H_{\mathit{AB},\mathit{PRED}}\left(z\right)X\left(z\right)\mathrm{with}{H}_{\mathit{AB},\mathit{PRED}}\left(z\right)=\frac{\mathrm{AT}\left(z\right)-1}{B\left(z\right)}\mathrm{.}$

In the following, generally H _PRED (z) denotes a filter whose coefficient for its current input x ( n ) is zero.

ou ARMA $\frac{A\left(z\right)}{B\left(z\right)}$

sont les filtres dit TIR pour Infinite Impulse Response en anglais (Filtre à réponse impulsionnelle infinie).All-pole recursive filters $\frac{1}{B\left(z\right)}$

or ARMA $\frac{\mathrm{AT}\left(z\right)}{B\left(z\right)}$

are the so-called TIR filters for Infinite Impulse Response in English ( Infinite Impulse Response Filter).

In the present case, figure 4 , using the filtering of a filtering into innovation and predictive parts, the term whose energy is to be minimized is then: $$\left(x\left(\mathrm{not}\right)+x_{\mathit{PRED}}\left(\mathrm{not}\right)\right)-\left({\tilde{x}}^{B+k}\left(\mathrm{not}\right)+{\tilde{x}}_{\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)\right)$$

ou X_PRED (n) et ${\tilde{x}}_{\mathit{PRED}}^{B+k}\left(n\right)$

sont obtenus en filtrant x(n) et ${\tilde{x}}^{B+k}\left(n\right)$

par le filtre de prédiction W_PRED (z). Ces deux filtrages peuvent être combinés en un seul, l'entrée du filtre commun W_PRED (z) sera alors b^B+k (n) = x(n) - x̃ ^B+k (n) (par exemple par la mise à jour des mémoire du filtre). En sortie du filtrage, on obtient alors: $$b_{w,\mathit{PRED}}^{B+k}\left(n\right)=x_{\mathit{PRED}}\left(n\right)-{\tilde{x}}_{\mathit{PRED}}^{B+k}\left(n\right)\mathrm{.}$$

The signal to be quantified by the enhancement quantizer of the stage k is therefore $$\mathit{x\; \text{'}}\left(\mathrm{not}\right)=x\left(\mathrm{not}\right)+x_{\mathit{PRED}}\left(\mathrm{not}\right)-{\tilde{x}}_{\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)$$

or X _PRED ( n ) and ${\tilde{x}}_{\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)$

are obtained by filtering x ( n ) and ${\tilde{x}}^{B+k}\left(\mathrm{not}\right)$

by the prediction filter W _PRED ( z ) . These two filterings can be combined into one, the input of the common filter W _PRED ( z ) will then be b ^{B + k} ( n ) = x ( n ) - x ^{B + k} ( n ) (for example by setting day of filter memory). At the end of the filtering, we obtain: $$b_{w,\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)=x_{\mathit{PRED}}\left(\mathrm{not}\right)-{\tilde{x}}_{\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)\mathrm{.}$$

en filtrant par W_PRED (z) en 404, des échantillons passés du signal x(n) - x̃ ^B+k (n) = b ^B+k (n) n=-1, -2, ...-N_D obtenus en 409.The preprocessing module 310 implements the steps of calculating a prediction $b_{w,\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)$

by filtering by W _PRED ( z ) at 404, passed samples of the signal x ( n ) - x ^{B + k} ( n ) = b ^{B + k} ( n ) n = -1, -2, ... - N _D obtained in 409.

est ajoutée au signal d'entrée x(n) en 405 pour obtenir le signal d'entrée modifié x'(n) du quantificateur de l'étage d'amélioration k.This prediction $b_{w,\mathit{PRED}}^{B+k}\left(\mathrm{not}\right)$

is added to the input signal x ( n ) at 405 to obtain the modified input signal x ' ( n ) of the quantizer of the improvement stage k.

de l'étage d'amélioration k et le signal décodé x̃ ^B+k (n) de l'étage k. Le module 307 donne l'indice du mot de code $I_{\mathit{enh}}^{B+k}\left(n\right)$

(1 bit dans l'exemple d'illustration) du dictionnaire adaptatif D_k qui minimise l'erreur quadratique entre x'(n) et les valeurs de quantification $d_1^{B+k}\left(n\right)$

et $d_2^{B+k}\left(n\right)\mathrm{.}$

Cet indice est à concaténer à l'indice du codeur imbriqué précédent I ^B+k-1 pour obtenir au décodeur l'indice du mot code de l'étage k I^B+k . Le module 308 donne la valeur quantifiée du signal d'entrée par quantification inverse de l'indice $I_{\mathit{enh}}^{B+k}\left(n\right),$

${\tilde{x}}^{B+k}\left(n\right)=d_{I_{\mathit{enh}1}^{B+k}}^{B+k}\left(n\right)\mathrm{.}$

The quantization of x ' ( n ) is carried out at 306 by the quantization module of the improvement stage k, to give the quantization index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)$

the improvement stage k and the decoded signal x ^{B + k} ( n ) of the stage k. The module 307 gives the index of the code word $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)$

(1 bit in the illustrative example) of the adaptive dictionary D _k which minimizes the quadratic error between x ' ( n ) and the quantization values $d_1^{B+k}\left(\mathrm{not}\right)$

and $d_2^{B+k}\left(\mathrm{not}\right)\mathrm{.}$

This index is to be concatenated with the index of the nested encoder preceding I ^{B + k- 1} in order to obtain at the decoder the index of the codeword of the stage k I ^{B + k} . The module 308 gives the quantized value of the input signal by inverse quantization of the index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right),$

${\tilde{x}}^{B+k}\left(\mathrm{not}\right)=d_{I_{\mathit{enh}1}^{B+k}}^{B+k}\left(\mathrm{not}\right)\mathrm{.}$

At the decoder the same value is obtained simply by directly using the inverse quantization of the stage k and the concatenated index to obtain: ${\tilde{x}}^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{I^{B+k}}^{B+k}v\left(\mathrm{not}\right)\mathrm{.}$

At 409, a step of calculating the encoding noise b ^{B + k} (n) of the encoder including the stage k is performed by subtracting the input signal x ( n ) from the synthesized signal of the stage k x ^{B + k} ( n ) for the samples present ( n = 0 ) .

The preprocessing operations of the block 310 thus make it possible to format the improvement coding noise of the stage k by performing a perceptual weighting of the input signal x (n). It is the input signal itself that is perceptually weighted and not an error signal as is the case in state-of-the-art methods.

The figure 5 illustrates another embodiment of the preprocessing module using in this embodiment a ARMA type filtering (for Auto Regressive to Adjusted Average) of transfer function: $$W\left(z\right)=\frac{1-P_D\left(z\right)}{1-P_{\mathrm{NOT}}\left(z\right)}$$

• Calcul en 301 du filtre de masquage et détermination du filtre de pondération $$W\left(z\right)=\frac{1-P_D\left(z\right)}{1-P_N\left(z\right)};$$
  
• Codage en 302 du signal d'entrée x(n) par un codeur imbriqué de type MIC/MICDA de B+k-1 bits, éventuellement avec mise en forme du bruit de codage en utilisant le filtre de masquage déterminé en 301 pour mettre en forme le bruit de codage;
• Détermination en 303 du dictionnaire adaptatif D_k en fonction des valeurs de prédiction $x_P^B\left(n\right)$
  
  et de facteur d'échelle v(n) (dans le cas d'un codage MICDA) de l'étage coeur, et des indices de quantification I ^B+k-1(n) $(d_1^{B+k}\left(n\right)=x_P^B\left(n\right)+{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}_{2I^{B+k-1}}^{B+k}v\left(n\right)$
  
  et $d_2^{B+k}\left(n\right)=x_P^B\left(n\right)+{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}_{2I^{B+k-1}+1}^{B+k}v\left(n\right));$
  

The sequence of operations according to the figure 5 is the next :

• 301 calculation of the masking filter and determination of the weighting filter $$W\left(z\right)=\frac{1-P_D\left(z\right)}{1-P_{\mathrm{NOT}}\left(z\right)};$$
  
• Encoding 302 of the input signal x ( n ) by a n- type coded MIC / ADPCM encoder of B + k-1 bits, possibly with encoding noise shaping using the masking filter determined in 301 to implement form the coding noise;
• Determination in 303 of the adaptive dictionary D _k as a function of the prediction values $x_P^B\left(\mathrm{not}\right)$
  
  and scaling factor v ( n ) (in the case of ADPCM coding) of the heart stage, and quantization indices I ^{B + k -1} ( n ) $(d_1^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right)$
  
  and $d_2^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+{\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right));$
  

These steps are equivalent to those described with reference to the figure 3 .

du bruit de quantification filtré ${\mathit{b}}_{\mathit{w}}^{\mathit{B}+\mathit{k}}\left(n\right),$

en ajoutant la prédiction calculée en 510 à partir des échantillons du bruit reconstitué filtré ${\displaystyle \sum _{m=1}^{N_D}}{p}_N\left(m\right)b_w^{B+k}\left(n-m\right)$

et en retranchant la prédiction calculée en 511 à partir du bruit reconstitué ${\displaystyle \sum _{m=1}^{N_D}}{p}_D\left(m\right)b^{B+k}\left(n-m\right)\mathrm{.}$

The pretreatment module 310 comprises a step of calculating a prediction signal in 512 ${\mathit{b}}_{\mathit{w},\mathit{pred}}^{\mathit{B}+\mathit{k}}\left(\mathrm{not}\right)$

filtered quantization noise ${\mathit{b}}_{\mathit{w}}^{\mathit{B}+\mathit{k}}\left(\mathrm{not}\right),$

adding the prediction calculated in 510 from the samples of the filtered reconstituted noise ${\displaystyle \underset{m=1}{\overset{{\mathrm{NOT}}_D}{\Sigma }}}{p}_{\mathrm{NOT}}\left(m\right)b_w^{B+k}\left(\mathrm{not}-m\right)$

and subtracting the calculated prediction in 511 from the reconstructed noise ${\displaystyle \underset{m=1}{\overset{{\mathrm{NOT}}_D}{\Sigma }}}{p}_D\left(m\right)b^{B+k}\left(\mathrm{not}-m\right)\mathrm{.}$

au signal x(n) est effectuée pour donner le signal modifié x'(n).In 505, a step of adding the prediction signal $b_{w,\mathit{pred}}^{B+k}\left(\mathrm{not}\right)$

to the signal x ( n ) is performed to give the modified signal x '( n ).

The quantification step of the modified signal x ' ( n ) is performed by the quantization module 306, in the same way as that explained with reference to the figures 3 and 4 .

et le signal décodé à l'étage k x̃^B+k (n). Thus, the quantization of block 306 outputs the index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)$

and the decoded signal at the stage k x ^{B + k} ( n ) .

In 509, a step of subtracting the reconstructed signal x ^{B + k} (n) from the signal x ( n ) is performed to give the reconstructed noise b ^{B + k} ( n ) .

au signal b^B+k (n) est effectué pour donner le bruit reconstruit filtré $b_w^{B+k}\left(n\right)\mathrm{.}$

In 513, a step of adding the prediction signal $b_{w,\mathit{pred}}^{B+k}\left(\mathrm{not}\right)$

at the signal b ^{B + k} ( n ) is performed to give the filtered reconstructed noise $b_w^{B+k}\left(\mathrm{not}\right)\mathrm{.}$

All the steps carried out at 505, 509, 510, 511, 512 and 513 by the modules of the preprocessing block 310 make it possible to format the coding noise for the improvement coding stage k. This shaping of the noise is then carried out by two prediction filters thus constituting an ARMA filter which brings a better precision of noise shaping.

est calculé. Le bruit reconstruit filtré $b_w^{B+k}\left(n\right)$

est obtenu ici en soustrayant le signal reconstruit x̃ ^B+k (n) du signal x'(n) en 614.The figure 6 illustrates yet another embodiment of the pretreatment block 310 where here the difference lies in the way the filtered reconstructed signal $b_w^{B+k}\left(\mathrm{not}\right)$

is calculated. Reconstructed noise filtered $b_w^{B+k}\left(\mathrm{not}\right)$

is obtained here by subtracting the reconstructed signal x ^{B + k} ( n ) from the signal x ' ( n ) at 614.

pour les échantillons passés.In the figures 5 and 6 described above, we can also talk about updating the memories of the weighting filters from the filtered reconstructed noise signal $b_w^{B+k}\left(\mathrm{not}\right)$

for past samples.

provenant du codage coeur. Ce mode de réalisation est présenté avec l'exemple de bloc de prétraitement 310 présenté à la figure 3, mais peut bien évidement s'intégrer avec des blocs de prétraitement décrits aux figures 4, 5 et 6. L'enchainement des opérations selon la figure 7 est le suivant :

• Calcul en 301 du filtre de masquage et détermination du filtre de pondération W(z) ou de sa version prédictive W_PRED (z).
• Codage en 302 du signal d'entrée x(n), par un codeur imbriqué de type MIC/MICDA de B+k-1 bits, éventuellement avec mise en forme du bruit de codage en utilisant le filtre de masquage déterminé en 301 pour mettre en forme du bruit de codage;
• Détermination en 701 du dictionnaire adaptatif D_k ' en fonction du facteur d'échelle v(n) de l'étage coeur (dans le cas d'un codage MICDA) et des indices de quantification I ^B+k-1(n) du codage imbriqué précédent l'étage k $(d_1^{B+kʹ}\left(n\right)={\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}_{2I^{B+k-1}}^{B+k}v\left(n\right)$
  
  et $d_2^{B+kʹ}\left(n\right)={\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}_{2I^{B+k-1}+1}^{B+k}v\left(n\right));$
  
• Filtrage du signal x(n) par W(z) en 311 pour obtenir le signal d'entrée modifié x'(n) du quantificateur d'amélioration avec pour mémoires du filtre W(z) des valeurs correspondant à un signal d'entrée x(n)-x̃ _B+k (n);
• Quantification de x'(n) en 706 pour donner l'indice $I_{\mathit{enh}}^{B+k}\left(n\right)$
  
  et le signal décodé à l' étage k x̃ ^B+k (n).

The figure 7 illustrates an alternative embodiment for the quantization step 306 of the signal x ' ( n ) by treating differently the predicted signal $x_P^B\left(\mathrm{not}\right)$

from the core coding. This embodiment is presented with the example of pretreatment block 310 presented to the figure 3 , but can of course be integrated with preprocessing blocks described in figures 4 , 5 and 6 . The sequence of operations according to the figure 7 is the next :

• 301 calculation of the masking filter and determination of the weighting filter W ( z ) or its predictive version W _PRED ( z ) .
• Encoding 302 of the input signal x ( n ) , by a n- type MIC / ADPCM encoder of B + k-1 bits, possibly with formatting of the coding noise using the masking filter determined in 301 to set in the form of coding noise;
• Determination at 701 of the adaptive dictionary D _k 'as a function of the scale factor v ( n ) of the core stage (in the case of ADPCM coding) and of the quantification indices I ^{B + k -1} ( n ) of the nested encoding preceding the k stage $(d_1^{B+k\text{'}}\left(\mathrm{not}\right)={\mathrm{there}}_{2I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right)$
  
  and $d_2^{B+k\text{'}}\left(\mathrm{not}\right)={\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right));$
  
• Filtering the signal x ( n ) by W ( z ) at 311 to obtain the modified input signal x ' ( n ) of the enhancement quantizer with the W ( z ) filter memories having values corresponding to an input signal x ( n ) -x _{B + k} ( n );
• Quantification of x ' ( n ) in 706 to give the index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)$
  
  and the decoded signal at the k x ^{B + k} ( n ) stage .

de l'étage coeur est soustrait du signal x'(n) (module 702) pour obtenir le signal modifié $xʹʹ\left(n\right)=xʹ\left(n\right)-x_P^B\left(n\right)\mathrm{.}$

In this embodiment, the predicted signal $x_P^B\left(\mathrm{not}\right)$

of the heart stage is subtracted from the signal x ' ( n ) (module 702) to obtain the modified signal $x\text{'}\text{'}\left(\mathrm{not}\right)=x\text{'}\left(\mathrm{not}\right)-x_P^B\left(\mathrm{not}\right)\mathrm{.}$

(1 bit dans l'exemple d'illustration) du dictionnaire adaptatif D_k' qui minimise l'erreur quadratique entre x"(n) et les mots de code $d_1^{B+kʹ}\left(n\right)$

et $d_2^{B+kʹ}\left(n\right)\mathrm{.}$

Cet indice est à concaténer à l'indice du codage imbriqué précédent I ^B+k-1 pour obtenir au décodeur l'indice du codage imbriqué courant I ^B+k comprenant l'étage k.Module 707 gives the index of the code word $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right)$

(1 bit in the illustrative example) of the adaptive dictionary D _k ' which minimizes the squared error between x " ( n ) and the code words $d_1^{B+k\text{'}}\left(\mathrm{not}\right)$

and $d_2^{B+k\text{'}}\left(\mathrm{not}\right)\mathrm{.}$

This index is to be concatenated with the index of the preceding nested encoding I ^{B + k -1} to obtain at the decoder the index of the current nested encoding I ^{B + k} comprising the stage k.

Le module 703 calcule le signal quantifié de l'étage k en additionnant le signal prédit au signal de sortie du quantificateur ${\tilde{x}}^{B+k}\left(n\right)=x_P^B\left(n\right)+\tilde{x}ʹʹ\left(n\right)\mathrm{.}$

The module 708 gives the quantized value of the signal x "( n ) by inverse quantization of the index $I_{\mathit{enh}}^{B+k}\left(\mathrm{not}\right),\tilde{x}\text{'}\text{'}\left(\mathrm{not}\right)=d_{I_{\mathit{enh}1}^{B+k}}^{B+k}\text{'}\left(\mathrm{not}\right)\mathrm{.}$

The module 703 calculates the quantized signal of the stage k by adding the predicted signal to the output signal of the quantizer ${\tilde{x}}^{B+k}\left(\mathrm{not}\right)=x_P^B\left(\mathrm{not}\right)+\tilde{x}\text{'}\text{'}\left(\mathrm{not}\right)\mathrm{.}$

Finally, a step of updating the memories of the filter W ( z ) is performed at 311, to obtain memories that correspond to an input x ( n ) -x ^{B + k} ( n ) . Typically one subtracts from the memory (or memories in the case of the filter type ARMA) more recent (s) the current value of the decoded signal x ^{B + k} ( n ) .

à tous les mots de code (>2) on ne fait qu'une soustraction avant la quantification et qu'une addition pour retrouver la valeur quantifiée x̃^B+k (n). La complexité est donc réduite.The solution on the figure 7 is equivalent in terms of quality and storage to that of the figure 3 but requires less computation in the case where the enhancement stage uses more than one bit. Instead of adding the predicted value $x_P^B\left(\mathrm{not}\right)$

to all code words (> 2) we only subtract before quantization and add to find the quantized value x ^{B + k} ( n ) . The complexity is reduced.

et $d_1^{B+kʹʹ}\left(n\right)=xʹ\left(n\right)-y_{2I^{B+k-1}+1}^{B+k}v\left(n\right))\mathrm{.}$

Dans ce cas de figure, c'est le signal de prédiction $x_P^B\left(n\right)$

qui est quantifié en minimisant l'erreur quadratique. Puis le signal décodé x̃^B+k (n) pour la mise à jour des mémoires est obtenu de la façon suivante: ${\tilde{x}}^{B+k}\left(n\right)=xʹ\left(n\right)+x_p^B\left(n\right)-d_{I_{\mathit{enh}1}^{B+k}}^{B+k}ʹʹ\left(n\right)\mathrm{.}$

Another alternative embodiment is illustrated in figure 7b . Here, the adaptive dictionary D _k "is constructed by subtracting the levels of reconstruction of the stage k weighted as appropriate by the scale factor v ( n ) , with the modified input signal $(d_1^{B+k\text{'}\text{'}}\left(\mathrm{not}\right)=x\text{'}\left(\mathrm{not}\right)-{\mathrm{there}}_{2I^{B+k-1}}^{B+k}v\left(\mathrm{not}\right)$

and $d_1^{B+k\text{'}\text{'}}\left(\mathrm{not}\right)=x\text{'}\left(\mathrm{not}\right)-{\mathrm{there}}_{2I^{B+k-1}+1}^{B+k}v\left(\mathrm{not}\right))\mathrm{.}$

In this case, it is the prediction signal $x_P^B\left(\mathrm{not}\right)$

which is quantified by minimizing the quadratic error. Then the decoded signal x ^{B + k} ( n ) for updating memories is obtained as follows: ${\tilde{x}}^{B+k}\left(\mathrm{not}\right)=x\text{'}\left(\mathrm{not}\right)+x_p^B\left(\mathrm{not}\right)-d_{I_{\mathit{enh}1}^{B+k}}^{B+k}\text{'}\text{'}\left(\mathrm{not}\right)\mathrm{.}$

Le module 802 calcule l'erreur de codage q_w (n)=x̃(n)-x(n) des instants d'échantillonnage précédents, n-1, n-2, .... Cette erreur est filtrée par un filtre prédicteur H_PRED (z) pour obtenir le signal de prédiction q_w,pred (n).
Le filtre H(z) correspondant à H_PRED (z) peut être égal par exemple soit à $H\left(z\right)=\frac{1}{P_1\left(z\right)}=A\left(z/\gamma \right),\mathrm{soit\; à}H\left(z\right)=\frac{1}{P_2\left(z\right)}=\frac{A\left(z/{\gamma }_2\right)}{A\left(z/{\gamma }_1\right)}\mathrm{.}$

The figure 8 details a possible realization of a noise shaping at the heart coding. The module 801 calculates the coefficients of the noise shaping filter $P_1\left(z\right)=\frac{1}{\mathrm{AT}\left(z/\gamma \right)}\mathrm{or}{P}_2\left(z\right)=\frac{\mathrm{AT}\left(z/{\mathrm{<figure-callout\; id="1"\; label="\gamma "\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\gamma </figure-callout>}}_1\right)}{\mathrm{AT}\left(z/{\mathrm{<figure-callout\; id="2"\; label="\gamma "\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">\gamma </figure-callout>}}_2\right)}\mathrm{.}$

The module 802 calculates the coding error q _w ( n ) = x ( n ) -x ( n ) of the previous sampling instants, n- 1 , n- 2 , .... This error is filtered by a filter predictor H _PRED (z) to obtain the prediction signal q _{w, pred} ( n ) .
The filter H ( z ) corresponding to H _PRED ( z ) may be equal, for example, to $H\left(z\right)=\frac{1}{P_1\left(z\right)}=\mathrm{AT}\left(z/\gamma \right),\mathrm{either\; to}H\left(z\right)=\frac{1}{P_2\left(z\right)}=\frac{\mathrm{AT}\left(z/{\mathrm{<figure-callout\; id="2"\; label="\gamma "\; filenames="imgf0002.png,imgf0004.png"\; state="\{\{state\}\}">\gamma </figure-callout>}}_2\right)}{\mathrm{AT}\left(z/{\mathrm{<figure-callout\; id="1"\; label="\gamma "\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\gamma </figure-callout>}}_1\right)}\mathrm{.}$

At time n, this predicted value will be subtracted from the signal to be encoded in order to obtain the signal to be coded modified x ' ( n ) = x (n) - q _{w, pred} ( n ) .

The difference between the input and the output of the encoder channel PCM / ADPCM - decoder PCM / ADPCM, q (n) = x (n) - x '(n) can be considered as short-term when these white noise Encoders use a quantizer with a large number of levels and assuming the stationary input signal.

Le signal d'entrée de la chaîne de codage standard MIC/MICDA est modifié par la soustraction de la contribution (H(z)-1)(X̃(z)-X(z)). Il en résulte que le bruit de codage de la chaîne complète q_G (n)=x̃(n)-x(n) sera mis en forme par le filtre $\frac{1}{H\left(z\right)}:$

_: $Q_G\left(z\right)=\frac{Q\left(z\right)}{H\left(z\right)}=\frac{1}{A\left(z/\gamma \right)}=Q\left(z\right),$

voici la démonstration en équations : $$\begin{array}{l}\tilde{X}\left(z\right)=\mathit{Xʹ}\left(z\right)+Q\left(z\right)=X\left(z\right)-\left(H\left(z\right)-1\right)\left(\tilde{X}\left(z\right)-X\left(z\right)\right)+Q\left(z\right)=\\ =\tilde{X}\left(z\right)-H\left(z\right)\tilde{X}\left(z\right)+H\left(z\right)X\left(z\right)+Q\left(z\right)\end{array}$$

d' où $$H\left(z\right)\tilde{X}\left(z\right)=H\left(z\right)X\left(z\right)+Q\left(z\right)\mathrm{et\; donc}\tilde{X}\left(z\right)=H\left(z\right)+\frac{Q\left(z\right)}{H\left(z\right)}$$

Take the example where $H\left(z\right)=\frac{1}{P_1\left(z\right)}=\mathrm{AT}\left(z/\gamma \right)\mathrm{.}$

The input signal of the standard MIC / ADPCM coding chain is modified by the subtraction of the contribution ( H ( z ) - 1) ( X ( z ) -X ( z )) . As a result, the coding noise of the complete string q _G ( n ) = x ( n ) -x ( n ) will be formatted by the filter $\frac{1}{H\left(z\right)}:$

_: $Q_{\mathrm{BOY\; WUT}}\left(z\right)=\frac{Q\left(z\right)}{H\left(z\right)}=\frac{1}{\mathrm{AT}\left(z/\gamma \right)}=Q\left(z\right),$

here is the proof in equations: $$\begin{array}{l}\tilde{X}\left(z\right)=\mathit{X\; \text{'}}\left(z\right)+Q\left(z\right)=X\left(z\right)-\left(H\left(z\right)-1\right)\left(\tilde{X}\left(z\right)-X\left(z\right)\right)+Q\left(z\right)=\\ =\tilde{X}\left(z\right)-H\left(z\right)\tilde{X}\left(z\right)+H\left(z\right)X\left(z\right)+Q\left(z\right)\end{array}$$

from where $$H\left(z\right)\tilde{X}\left(z\right)=H\left(z\right)X\left(z\right)+Q\left(z\right)\mathrm{and\; so}\tilde{X}\left(z\right)=H\left(z\right)+\frac{Q\left(z\right)}{H\left(z\right)}$$

In fact, the filter H _PRED ( z ) = H ( z ) - 1 has a zero coefficient in z ⁰ (for the moment n ), it is therefore a predictor acting on q _w ( n ) = x ( n ) -x ( n ) which is known only at the end of the PCM / ADPCM processing when the decoded value x ( n ) is known.

• Calcul en 801 du filtre de masquage et détermination du filtre H(z). A noter que le filtre H(z) peut également être déterminé à partir du signal décodé x̃(n);
• Calcul en 803 de la prédiction q_w,pred (n), ([H(z)-1]Q_w (z)), à partir des valeurs q_w (n) = x̃(n)-x(n) des instants d'échantillonnage précédents, n-1,n-2,... ;
• Soustraction en 804 de la prédiction q_w,pred (n) à x(n) pour obtenir le signal x'(n) modifié;
• Codage/Décodage en 805-806 du signal modifié x'(n) par un codeur/décodeur MIC/MICDA standard. Le décodeur local peut être un décodeur local standard du type MIC/MICDA des normes G.711, G.721, G.726, G.722 ou encore G.727.
• Calcul en 802 du bruit de codage filtré q_w (n) par soustraction du signal d'entrée x(n) du signal de sortie x̃(n)

The sequence of operations of the figure 8 is the following :

• Calculation in 801 of the masking filter and determination of the filter H ( z ). Note that the filter H ( z ) can also be determined from the decoded signal x ( n );
• Computation in 803 of the prediction q _{w, pred} ( n ) , ([ H (z) - 1] Q _w (z)), starting from the values q _w ( n ) = x ( n ) -x ( n ) of previous sampling instants, n- 1 , n -2, ...;
• Subtraction at 804 of the prediction q _{w, pred} ( n ) at x ( n ) to obtain the modified signal x '( n );
• 805-806 encoding / decoding of the modified signal x ' ( n ) by a standard MIC / ADPCM encoder / decoder. The local decoder can be a standard local decoder of the MIC / ADPCM type of the G.711, G.721, G.726, G.722 or G.727 standards.
• 802 calculation of the filtered coding noise q _w (n) by subtracting the input signal x ( n ) from the output signal x ( n )

The surrounding portion 807 can be seen and implemented as a noise shaping pretreatment that modifies the input of the standard encoder / decoder chain.

An exemplary embodiment of an encoder according to the invention is now described with reference to the figure 10 .

Materially, an encoder 900 as described according to the various embodiments above, in the sense of the invention, typically comprises a μP processor cooperating with a memory block BM including a storage and / or working memory, as well as 'a memory MEM buffer mentioned above as a means for storing, for example, a dictionary of quantization reconstruction levels or any other data necessary for the implementation of the coding method as described with reference to FIGS. figures 3 , 4 , 5 , 6 and 7 . This encoder receives as input successive frames of the digital signal x (n) and delivers concatenated quantization indices I ^{B + k} .

The memory block BM may comprise a computer program comprising the code instructions for implementing the steps of the method according to the invention when these instructions are executed by a μP processor of the encoder and in particular the steps of obtaining possible values of quantification. for the current improvement stage k by determining absolute reconstruction levels of the single current stage k from the indices of the preceding nested encoder, quantization of the input signal of the hierarchical coder which has or has not undergone perceptual weighting processing (x (n) or x '(n)), from said possible quantization values to form a quantization index of the stage k and a quantized signal corresponding to one of the possible quantization values.

More generally, a storage means, readable by a computer or a processor, whether or not integrated into the encoder, possibly removable, stores a computer program implementing a coding method according to the invention.

The Figures 3 to 7 can for example illustrate the algorithm of such a computer program.

Claims (8)

1. Method for coding a digital audio input signal (x(n)) in a hierarchical coder comprising a core coding stage with B bits and at least one current improvement coding stage k, the core coding and the coding of the improvement stages preceding the current stage k delivering quantization indices which are concatenated to form the indices of the preceding embedded coder (I^B+k-1), the method being characterized in that it comprises the following steps:

   - obtaining (303) of possible quantization values ( ) for the current improvement stage k on the basis of the absolute reconstruction levels $\left(y_i^{B+k}\right)$

   

   of just the current stage k and of the indices of the preceding embedded coder (I ^B+K-1);

   - quantization (306) of the input signal of the hierarchical coder having undergone or not a perceptual weighting processing (x(n) or x'(n)), on the basis of said possible quantization values $\left(d_i^{B+k}\left(n\right)\right)$

   

   so as to form a quantization index for the stage k (I_enh ^B+k(n)) and a quantized signal (x̃^B+k (n)) corresponding to one of the possible quantization values.
2. Method according to Claim 1, characterized in that the input signal has undergone a perceptual weighting processing using a predetermined weighting filter to give a modified input signal x'(n), before the quantization step (306) and in that it furthermore comprises a step (311) of adapting the memories of the weighting filter on the basis of the quantized signal (x̃^B+k (n)) of the current improvement coding stage.
3. Method according to Claim 1, characterized in that the possible quantization values for improvement stage k furthermore contain a scale factor and a prediction value originating from the core coding of adaptive type.
4. Method according to Claim 2, characterized in that the modified input signal (x"(n)) to be quantized at improvement stage k is the perceptually weighted input signal from which is subtracted a prediction value originating from the core coding of adaptive type.
5. Method according to Claims 1 to 4, characterized in that the perceptual weighting processing is performed by prediction filters forming a filter of ARMA type.
6. Hierarchical coder of a digital audio input signal (x(n)), comprising a core coding stage with B bits and at least one current improvement coding stage k, the core coding and the coding of the improvement stages preceding the current stage k delivering quantization indices which are concatenated to form the indices of the preceding embedded coder (I^B+k-1), the coder being characterized in that it comprises:

   - a module (303) for obtaining possible quantization values ( ) for the current improvement stage k by determining absolute reconstruction levels of just the current stage k on the basis of the indices of the preceding embedded coder (I ^B+k-1);

   - a module (306) for quantizing the input signal of the hierarchical coder having undergone or not a perceptual weighting processing (x(n) or x'(n)), on the basis of said possible quantization values $\left(d_i^{B+k}\left(n\right)\right)$

   

   so as to form a quantization index for the stage k (I_enh ^B+k(n)) and a quantized signal (x̃^B+k (n)) corresponding to one of the possible quantization values.
7. Hierarchical coder according to Claim 6, characterized in that it furthermore comprises a preprocessing module (310) for perceptual weighting using a predetermined weighting filter to give a modified input signal (x'(n)) at the input of the quantization module (306) and a module (311) for adapting the memories of the weighting filter on the basis of the quantized signal (x̃^B+k (n)) of the current improvement coding stage.
8. Computer program comprising code instructions for the implementation of the steps of the coding method according to one of Claims 1 to 5, when these instructions are executed by a processor.

Images