EP0906613B1 - Method and device for coding an audio signal by "forward" and "backward" lpc analysis - Google Patents

Description

The invention relates to a method and a device for coding an audio signal, such as a speech signal, by " forward " and " backward " LPC analysis.

Currently, coding techniques for audio signals, including speech signals, have aim to allow the transmission of these signals in digital form, under conditions of reduction of transmission rate, in particular to ensure adapted management of the transmission networks for these signals, given the significant increase in transactions between users.

Among the coding techniques used, that designated by LPC analysis, for "Linear Predictive Coding" in Anglo-Saxon language, consists in performing a linear prediction of the audio frequency signal to be coded, the coding being carried out temporally by means of prediction filtering. linear applied to successive blocks of this signal.

In the aforementioned techniques, that known under the name of CELP coding, for "Code Excited Linear Prediction ", is the most widespread and one of the most effective. Other techniques, such as the technique designated by MP-LPC, for " Multi Pulse Linear Predictive Coding ", or the technique VSELP, for " Vector Sum Excited Linear Prediction " in Anglo-Saxon language, are relatively close to CELP coding.

The above coding techniques are said to be analysis by synthesis ". In particular, they made it possible, for audio signals belonging to the band telephone frequency, reduce transmission rate of these signals from 64 kbps (MIC coding) to 16 kbps using the CELP coding technique, and even up to 8 kb / s in the case of coders implementing the most advanced recent results of this coding technique, without degradation noticeable from the quality of the speech restored after transmission and decoding.

A particularly important field of application of these coding techniques is, in particular, that of mobile telephony. In this area of application, the necessary limitation of the frequency band granted to every mobile operator and the increase very rapid number of user subscribers make it necessary the corresponding decrease in coding rate, while user requirements for speech quality do not cease to grow. Other areas of application of these coding techniques relate, for example, to storage digital data representative of these signals on storage media, high quality telephony for video or audio conference applications, multimedia, or digital satellite transmissions.

The linear prediction filters used in the aforementioned techniques are obtained using an analysis module called "LPC analysis" operating on successive blocks of the digital signal. These filters are capable, according to the order of analysis, that is to say according to the number of coefficients of the filter, of modeling more or less faithfully the contours of the frequency spectrum of the signal to be coded. In the case of a speech signal, these contours are called formants.
However, for good quality coding, required by most current applications, the filter thus defined is not sufficient to perfectly model the signal. It is then essential to proceed to the coding of the linear prediction residue. Such an operating mode relating to the linear prediction residue is notably implemented by the LD-CELP coding technique, for Low Delay CELP in Anglo-Saxon language, previously mentioned in the description. The residual signal is in this case modeled by a waveform extracted from a stochastic dictionary and multiplied by a gain value. The MP-LPC coding technique, for example, models this residue using variable position pulses assigned respective gain values, while the VSELP coding technique performs this modeling by a linear combination of pulse vectors extracts from appropriate directories.
A didactic reminder of the operating mode of the LPC analysis and in particular of the " back " LPC analysis and of the "front" LPC analysis or the "backward" LPC analysis and the "forward" LPC analysis respectively in Anglo-Saxon language, will be very first given below.

où p désigne le nombre de coefficients a_i du filtre et l'ordre de la prédiction linéaire mise en oeuvre, z désignant la variable de la transformée en z de l'espace des fréquences.The general envelope of the frequency spectrum is modeled by means of a short-term synthesis filter, constituting the LPC filter, the coefficients of which are evaluated by means of a linear prediction of the speech signal to be coded. This LPC filter, autoregressive filter, has a shape transfer function, relation (1):

where p denotes the number of coefficients a _i of the filter and the order of the linear prediction implemented, z denoting the variable of the transform in z of the frequency space.

où s(n) désigne l'échantillon de rang n dans la trame de N échantillons. A method for evaluating the coefficients a _i consists in applying a criterion for minimizing the energy of the prediction error signal of the speech signal over the analysis length of the latter.
The analysis length for a digital speech signal formed by successive samples is in practice a number N of these samples, constituting a coding frame. The energy of the prediction error signal then checks the relation (2):

where s (n) denotes the sample of rank n in the frame of N samples.

In a block coding process, the frame of coding can advantageously be divided into several subframes or adjacent LPC blocks. Analysis length N then exceeds the length of each block to allow the taking into account a certain number of past samples and, where applicable, future, by means and at the cost of delays appropriate coding.

The analysis is called LPC "before" when the LPC analysis process is carried out on the block of the current frame of the speech signal to be coded, the coding at the level of the coder intervening "in real time", that is to say say during the block of the current frame to the sole processing delay introduced by calculating the coefficients of the filter. This analysis involves the transmission of the calculated values of the coefficients of the filters to the decoder.

The "rear" LPC analysis implemented in the LD-CELP coder at 16 kb / s is the subject of standard ITU-T G728. This analysis technique consists in performing the LPC analysis, not on the block or the block of the current frame of the speech signal to be coded, but on the synthesis signal. It will then be understood that this LPC analysis is in fact carried out on the synthesis signal of the block preceding the current block, since this signal is available simultaneously at the level of the coder and the decoder. This simultaneous operation with the coder and the decoder thus makes it possible to avoid the transmission from the coder to the decoder of the value, obtained at the coder, of the coefficients of the LPC filter. For this reason, the “ backward ” LPC analysis makes it possible to free up the transmission rate, the rate thus freed up being able to be used for example in order to enrich the excitation dictionaries in the case of CELP coding. The “ backward ” LPC analysis also allows an increase in the order of analysis, the number of coefficients of the LPC filter being able to reach 50 in the case of an LD-CELP coder against 10 coefficients for most coders implementing performs a " before " LPC analysis.

• bonne qualité du signal de synthèse, très proche du signal de parole à coder, ce qui implique un débit de codage suffisamment élevé, supérieur à 13 kb/s compte tenu de la qualité actuelle des codeurs CELP ;
• trame et bloc de longueur réduite en raison du retard d'un bloc entre signal analysé et signal à coder. La longueur de trame et de bloc doit donc être faible par rapport au temps de stationnarité moyen du signal de parole à coder ;
• fidélité de la transmission et respect de l'intégrité des données transmises entre codeur et décodeur, par l'introduction de peu d'erreurs de transmission. Dès que les signaux de synthèse diffèrent de manière significative du signal de parole à coder, codeur et décodeur ne calculent plus le même filtre et des divergences importantes peuvent survenir, en l'absence de toute chance de retour à une sensible identité des filtres calculés au codeur ou au décodeur.

Thus, proper functioning of the " back " LPC analysis requires the following conditions:

• good quality of the synthesis signal, very close to the speech signal to be coded, which implies a sufficiently high coding bit rate, greater than 13 kb / s taking into account the current quality of CELP coders;
• frame and block of reduced length due to the delay of a block between the analyzed signal and the signal to be coded. The frame and block length must therefore be small compared to the average stationarity time of the speech signal to be coded;
• fidelity of the transmission and respect for the integrity of the data transmitted between encoder and decoder, by the introduction of few transmission errors. As soon as the synthesis signals differ significantly from the speech signal to be coded, the coder and decoder no longer calculate the same filter and significant divergences may occur, in the absence of any chance of returning to a sensitive identity of the filters calculated at encoder or decoder.

Due to the respective advantages and disadvantages of the aforementioned types of LPC analysis "rear" and "front", a technique consisting in selectively combining the LPC analysis "rear" and "front" was proposed in the article entitled: "Dual Rate Low Delay CELP Coding (8 kbits / s / 16 kbits / s) using a Mixed Backward / Forward Adaptive LPC Prediction " published by S.PROUST, C.LAMBLIN and D.MASSALOUX, Proc.IEEE Workshop Speech Cod. Telecomm., Sept. 1995, pp 37-38. The conditions previously mentioned, relating to the proper functioning of the " rear " LPC analysis, reveal that this only type of analysis has obvious limits when operating at transmission rates significantly lower than 16 kb / s. In addition to the reduction in the quality of the synthesis signal, which degrades the performance of the LPC filter, it is most often necessary, in order to reduce the transmission rate, to operate on a longer LPC frame length, of the order of 10 to 30 ms. It can then be seen that under these conditions, the degradation occurs before all during frequency spectrum transitions and more generally in areas with little stationary, while for generally very stationary signals such as those relating to music, the LPC analysis "rear " retains a very significant advantage over LPC analysis " before. ".

• analyse LPC "avant" pour le codage des transitions et des zones non-stationnaires ;
• analyse LPC "arrière", d'ordre plus élevé, pour le codage des zones stationnaires.

The purpose of combining the two types of LPC analysis mentioned above is to overcome these drawbacks by taking advantage of the advantages inherent in each of them:

• " forward " LPC analysis for coding transitions and non-stationary areas;
• higher order "back" LPC analysis for coding stationary areas.

In addition, the introduction of LPC frames coded by " forward " LPC analysis among LPC frames coded by "back" LPC analysis allows the coder and the decoder to converge again to the same synthesis signal in the event of transmission error. and therefore offers a robustness to these errors much superior to a coding by pure "back" LPC analysis.

• un premier critère fondé sur les gains de prédiction des filtres :

• un deuxième critère fondé sur un paramètre de distance entre filtres LPC "avant" calculés successivement. Pour chacun de ces deux critères, des valeurs de seuil, fixes, sont établies.

Overall, the aforementioned "front" - "rear" mixed LPC analysis consists of carrying out two LPC analyzes, a "front" LPC analysis on the speech or audio frequency signal to be coded and a "rear" LPC analysis on the synthesis signal.
Two filters are calculated for each LPC block, these filters being designated by " front " LPC filter and " rear " LPC filter respectively. A procedure for choosing the filter applied for the LPC block considered as a function of the stationarity of the signal is then implemented. This procedure uses two separate criteria:

• a first criterion based on the filter prediction gains:

• a second criterion based on a distance parameter between LPC filters " before " calculated successively. For each of these two criteria, fixed threshold values are established.

First criterion:

The choice of the "rear" LPC filter is retained if the difference between the prediction gain of the "rear" and "front" LPC filters is greater than a first threshold value.

Second criterion:

For a current analysis in "back" LPC analysis mode , prohibition of switching from "back" LPC analysis mode to "front" LPC analysis mode if the distance calculated on the parameter vectors representing two " front " LPC filters "consecutive is less than a second threshold value, a too small distance characterizing a substantially stationary zone for which it is advisable to avoid any change in LPC analysis mode. The calculated distance is a Euclidean distance between the spectral lines of the speech or audio frequency signal to be coded.

For a more detailed description of the analysis mode LPC mixed above, we can usefully refer to the article published by S.PROUST, C.LAMBLIN and D.MASSALOUX previously cited.

• pour certains signaux, les valeurs des gains de prédiction des filtres LPC "avant" et "arrière" peuvent osciller de part et d'autre de la première valeur de seuil. Ce phénomène occasionne des changements de filtre LPC "arrière" - LPC "avant", ou réciproquement, brusques et fréquents. Les discontinuités de filtrage alors introduites constituent une source de dégradation importante du signal de synthèse et ne sont pas, la plupart du temps, liées à des réelles modifications spectrales du signal de parole ou audiofréquence à coder ;
• la valeur optimale du premier seuil, qu'il convient de fixer, varie très fortement en fonction de la stationnarité du signal à coder, ce d'autant plus que le débit de codage est faible. Pour un retard de codage correspondant à une trame LPC de 10 à 30 ms, ou lorsque le débit de transmission diminue, il apparaít une divergence très nette du mode de codage de signaux de musique et de parole. Pour les signaux de musique, l'analyse LPC "arrière" est utilisée quasiment en permanence alors que pour les signaux de parole, l'analyse LPC "avant" est utilisée majoritairement.

Alors que dans le cas de signaux de musique, la stationnarité étant très élevée, l'analyse en mode LPC "arrière" est retenue même pour une longueur de trame LPC importante, dans le cas des signaux de parole, au contraire, les zones très stationnaires sont de durée très limitée et le passage en mode d'analyse LPC "arrière" en conséquence bref, ce qui provoque des transitions de filtres indésirables qui dégradent la qualité du codage. Le codeur n'est alors plus en mesure de corriger les phénomènes engendrés par la discontinuité introduite par le basculement des filtres ;

• le filtre LPC qui donne la meilleure qualité subjective et modélise donc le mieux le spectre du signal à coder n'est pas toujours celui qui possède le meilleure gain de prédiction. Certains basculements d'un mode d'analyse LPC à l'autre, liés à une décision instantanée, sont donc inutiles.

Extensive investigations carried out on the operating method of the above-mentioned mixed analysis have made it possible to highlight the following significant drawbacks:

• for certain signals, the values of the prediction gains of the “front” and “rear” LPC filters can oscillate on either side of the first threshold value. This phenomenon causes abrupt and frequent LPC filter changes "rear" - LPC "front", or vice versa. The filtering discontinuities then introduced constitute a source of significant degradation of the synthesis signal and are not, most of the time, linked to real spectral modifications of the speech or audio frequency signal to be coded;
• the optimal value of the first threshold, which should be fixed, varies very greatly as a function of the stationarity of the signal to be coded, all the more so when the coding rate is low. For a coding delay corresponding to an LPC frame of 10 to 30 ms, or when the transmission rate decreases, there appears a very clear divergence in the coding mode of music and speech signals. For music signals, the "back" LPC analysis is used almost continuously while for speech signals, the " front " LPC analysis is used mostly.

Whereas in the case of music signals, the stationarity being very high, the analysis in LPC mode "back" is retained even for a long LPC frame length, in the case of speech signals, on the contrary, the very are of very limited duration and the transition to LPC analysis mode "back" consequently brief, which causes unwanted filter transitions which degrade the quality of coding. The encoder is then no longer able to correct the phenomena caused by the discontinuity introduced by the tilting of the filters;

• the LPC filter which gives the best subjective quality and therefore best models the spectrum of the signal to be coded is not always the one with the best prediction gain. Certain switches from one LPC analysis mode to another, linked to an instant decision, are therefore unnecessary.

The object of the present invention is to remedy the aforementioned drawbacks by implementing a method and a device for coding a digital audio frequency signal by specific " front " and "rear" LPC analysis.

Another object of the present invention is also the implementation of a process of dynamic adaptation of the function of choice between the "front" LPC analysis and the "rear" LPC analysis according to the degree of stationarity of the signal. to code.

Another object of the present invention is also the implementation of a process of dynamic adaptation of the aforementioned choice function on the basis of a discrimination between strongly stationary signals, such as music or background noise, and other signals. , such as speech, to allow the most appropriate coding processing by "back" and "front" LPC analysis respectively.

Another object of the present invention is also, the choice of the aforementioned most appropriate coding having been carried out, for a signal to be coded of a given type or characteristics, to avoid any untimely switchover in the non-LPC analysis mode. retained, and, thus, to avoid the appearance of transitions of LPC filters " front " - " rear " or vice versa likely to degrade the quality of the reproduced synthesis signal.

Another object of the present invention is finally the implemented a process of dynamic adaptation of the above-mentioned choice function for which the mode change LPC analysis is a true reflection of a change stationarity of the signal to be coded and risks, therefore, to be much less linked to a simple crossing effect point of the first and second threshold values.

• déterminer le degré de stationnarité du signal numérique audiofréquence selon un paramètre de stationnarité, dont la valeur est comprise entre une valeur de stationnarité maximale et une valeur de stationnarité minimale ;
• établir une valeur de choix d'analyse, en appliquant une fonction de décision adaptative au paramètre de stationnarité ;
• appliquer la valeur de choix d'analyse au filtrage LPC pour effectuer le codage du signal numérique audiofréquence par filtrage LPC "avant" pour les zones non-stationnaires sur le signal numérique audiofréquence et par filtrage LPC "arrière" pour les zones stationnaires sur le signal de synthèse.

Ce mode opératoire permet de privilégier le maintien dans l'un des modes de filtrage LPC "avant" et "arrière" respectivement, en liaison avec le degré de stationnarité du signal numérique audiofréquence et de limiter le nombre de basculements de l'un à l'autre des modes de filtrage et réciproquement.The method and the device for coding a digital audio signal, objects of the present invention, implement a double analysis on criteria of choice of LPC analysis analysis "front" and "rear" respectively to generate a transmitted coded signal consisting of LPC filtering parameters accompanied by analysis decision information and a coding residue signal, not transmitted. The digital audio signal is subdivided into frames, a succession of blocks of a determined number of samples and the coding of this digital audio signal is carried out on this signal using " forward " LPC filtering for non-stationary areas and on a synthesis signal respectively, this synthesis signal being obtained from the residual coding signal, from a “backward” LPC filtering for the stationary zones.
They are remarkable in that they consist of and allow, respectively:

• determining the degree of stationarity of the digital audio frequency signal according to a stationarity parameter, the value of which lies between a maximum stationarity value and a minimum stationarity value;
• establishing an analysis choice value, by applying an adaptive decision function to the stationarity parameter;
• apply the analysis choice value to the LPC filtering to carry out the coding of the digital audio signal by "front" LPC filtering for non-stationary zones on the digital audio signal and by "rear" LPC filtering for stationary zones on the signal of synthesis.

This operating mode makes it possible to favor the maintenance in one of the LPC filtering modes "front" and "rear" respectively, in connection with the degree of stationarity of the digital audio signal and to limit the number of toggles from one to the other. 'other filtering methods and vice versa.

The method and the device, objects of the present invention find application not only in the field of mobile telephony but also to the industry creation and reproduction of phonograms, at the high quality satellite and telephone transmission for video or audio conference applications, multimedia.

• la figure 1 représente, sous forme d'un organigramme général, un schéma illustratif des étapes permettant la mise en oeuvre du procédé de codage, objet de la présente invention ;
• la figure 2a représente un organigramme général des étapes de calcul du paramètre de stationnarité pour chaque bloc LPC courant ;
• la figure 2b représente un mode de réalisation particulier avantageux des étapes essentielles du calcul du paramètre de stationnarité selon la figure 2a ;
• la figure 2c représente un détail de réalisation de la figure 2b, plus particulièrement un détail du processus d'affinage de la valeur du paramètre de stationnarité intermédiaire pour l'obtention du paramètre de stationnarité ;
• les figures 2d et 2e représentent un premier, respectivement un deuxième exemple de réalisation non limitatif de mise en oeuvre d'une fonction d'affinage permettant de calculer une valeur d'affinage du paramètre de stationnarité intermédiaire en fonction des valeurs relatives du gain de filtrage LPC "avant" et "arrière" ;
• la figure 2f représente, à titre d'exemple illustratif, un organigramme des étapes permettant la mise en oeuvre de la fonction de décision et de la valeur du choix d'analyse LPC "avant" ou "arrière" ;
• la figure 3 représente, sous forme de blocs fonctionnels, le schéma général d'un codeur permettant d'effectuer le codage d'un signal audiofréquence conformément à l'objet de la présente invention ;
• la figure 4 représente, sous forme de blocs fonctionnels, le schéma général d'un décodeur permettant d'effectuer le décodage d'un signal audiofréquence codé grâce à la mise en oeuvre d'un codeur tel que représenté en figure 3.

They will be better understood on reading the description and on observing the drawings below, in which:

• FIG. 1 represents, in the form of a general flowchart, an illustrative diagram of the steps allowing the implementation of the coding method, object of the present invention;
• FIG. 2a represents a general flow diagram of the steps for calculating the stationarity parameter for each current LPC block;
• FIG. 2b represents a particular advantageous embodiment of the essential steps of the calculation of the stationarity parameter according to FIG. 2a;
• FIG. 2c represents a detail of embodiment of FIG. 2b, more particularly a detail of the process of refining the value of the intermediate stationarity parameter for obtaining the stationarity parameter;
• FIGS. 2d and 2e represent a first, respectively a second nonlimiting example of implementation of a refining function making it possible to calculate a refining value of the intermediate stationarity parameter as a function of the relative values of the filtering gain LPC "front" and "rear" ;
• FIG. 2f represents, by way of illustrative example, a flow diagram of the steps allowing the implementation of the decision function and of the value of the LPC analysis choice " front " or "back" ;
• FIG. 3 represents, in the form of functional blocks, the general diagram of an encoder making it possible to carry out the coding of an audio frequency signal in accordance with the object of the present invention;
• FIG. 4 represents, in the form of functional blocks, the general diagram of a decoder making it possible to carry out the decoding of an audio-coded signal signal thanks to the implementation of an encoder as represented in FIG. 3.

A more detailed description of the coding method of a digital audio signal by double analysis, on the criterion of choice of LPC analysis " front " or "rear" respectively into a coded transmitted signal, object of the present invention, will now be given in connection with figure 1.

In general, it is indicated that the transmitted coded signal, noted s_c _n (t), consists partly of LPC filtering parameters accompanied by LPC analysis decision information. In addition, a non-transmitted coding residue signal res _n (t) is available by the implementation of the coding method.

The digital audio frequency signal is subdivided into LPC frames, a succession of LPC blocks, each block, for the convenience of the description, being denoted B _n and provided with a determined number N of samples.

In accordance with one aspect of the coding method which is the subject of the present invention, it consists in carrying out the aforementioned coding on the digital audio frequency signal as defined above from " forward " LPC filtering for the non-stationary areas, respectively on a synthesis signal obtained from the residual coding signal from a "rear" LPC filtering for stationary areas.

According to a particularly remarkable aspect of the method which is the subject of the present invention, it consists, in order to establish the criterion of choice of LPC filtering "front" or "rear", on each current block of the succession of current blocks constituting a frame current, as shown in FIG. 1, each current block, denoted B _n , being available in a starting step 10, to be determined in a step 11 the degree of stationarity of the digital audio signal according to a stationarity parameter, denoted STAT (n ). This stationarity parameter has a numerical value between a maximum stationarity value, denoted STAT _M , and a minimum stationarity value, denoted STAT _m .

By convention and without in any way detracting from the degree of generality of the coding method which is the subject of the present invention, it is indicated that the stationarity parameter has the maximum value STAT _M for a very strongly stationary signal, while this stationarity parameter has the minimum value STAT _m for a very strongly non-stationary signal.

Following the aforementioned step 11, the coding method which is the subject of the present invention consists in establishing, in a step 12, from the stationarity parameter STAT (n), an analysis choice value LPC, this choice value corresponding analysis of course, either to the choice of LPC analysis "before" , or on the contrary to the choice of LPC analysis "back". The choice of analysis value is denoted d _n (n) and is obtained from a specific decision function, denoted D _n .

The above-mentioned step 12 is then followed by a test step 13 allowing the application of the analysis choice value d _n (n), symbolized by C, to the LPC filtering to effect the coding of the digital audio-frequency signal by filtering "Front" LPC for non-stationary areas on the digital audio signal, respectively by "rear" LPC filtering for stationary areas on the synthesis signal.

The implementation of the decision function D _n and of the abovementioned analysis choice values d _n (n), in accordance with a particularly advantageous aspect of the coding method which is the subject of the present invention, makes it possible to favor maintenance in the one of the LPC filtering modes "front" respectively "rear", in connection with the degree of stationarity of the audio frequency signal, and to limit the number of switches from one to the other of the filtering modes, and vice versa.

In general, it is indicated that the decision function implemented in step 12, this decision function being denoted D _n , is an adaptive function updated for each current block B _n , from the stationarity parameter.

The updating of the adaptive function makes it possible to privilege the maintenance in one of the LPC filtering modes "front", respectively "rear" , according to the degree of stationarity of the digital audio-frequency signal and thus limit the number of toggles of l 'to one another filtering modes, and vice versa.

More specifically, it is indicated that the analysis choice value d _n (n) established from the above-mentioned decision function D _n corresponds to a filtering mode priority value LPC " front " or "rear " as well as to another priority value representing in fact a value of absence of priority to return to the LPC filtering mode " rear " or " front ".

By LPC filtering mode priority value, it is indicated that the analysis choice value d _n (n) can for example correspond to a logical value, the true value of this logical value, value 1 for example, corresponding to a LPC filtering choice "backward" while the value complemented by this true value, the value zero, corresponds to a LPC filtering choice "forward". It is thus understood that the test function in step 13 can be summed up as a test value on the logical value of the above-mentioned analysis choice value to ensure in step 14 the "back" LPC filtering for the zones. of the signal to be coded or the LPC filtering "before" in step 15 for the non-stationary zones, the aforementioned steps 14 and 15 then being followed by steps 14a and 15a on returning to the next block denoted B _n-1 , for n = n + 1.

Although the analysis choice value d _n (n) is represented by a logical value, it is understood that this logical value can be associated with a priority and probability value of filtering mode established by the decision function D _n specifically. It is understood in particular that this probability value can correspond, for each current block B _n , to the true logic value for a range of probability values between zero and 1 for "backward" LPC filtering whereas the complemented logic value, value logic zero, for example, may correspond to the complement of the above range of probability values between zero and 1 of the first aforementioned range. This probability is linked to the number of successive filtering decisions in the same filtering mode.

The operating mode of the decision function D _n making it possible in fact to associate with the logic variable d _n (n) the priority of filtering mode, is adaptive over time, for each current block B _n .

In general, it is indicated that the adaptation of the decision function D _n aims to progressively favor the "back" LPC filtering mode or on the contrary the "front" LPC filtering mode which works best, counts given the overall stationarity of the signal to be coded, in order to avoid as much as possible any unnecessary switching from one of the filtering modes to the other.

• plus le signal à coder est stationnaire et plus la fonction de décision D_n privilégie l'analyse LPC "arrière" en limitant le plus possible le basculement en mode d'analyse LPC "avant",
• au contraire, moins le signal à coder est stationnaire et plus la fonction de décision D_n privilégie l'analyse LPC "avant" en limitant le plus possible tout basculement en mode d'analyse LPC "arrière".

More specifically, we indicate that:

• the more the signal to be coded is stationary and the more the decision function D _n favors the " back " LPC analysis by limiting as much as possible the switch to " forward " LPC analysis mode,
• on the contrary, the less the signal to be coded is stationary and the more the decision function D _n favors the " forward " LPC analysis by limiting as much as possible any switch to "backward" LPC analysis mode .

A more detailed description of implementation a specific decision function allowing to realize the adaptation of this decision function, depending on the value of the stationary parameter STAT (n), will be given later in the description.

A preferred method of calculating the stationarity parameter STAT (n) relative to each current LPC block B _n will now be given and described in connection with FIG. 2a.

According to the aforementioned figure, it is indicated that step 11 consisting in determining the degree of stationarity of each current block B _n of the digital audio frequency signal consists, from an arbitrary starting value of the stationarity parameter, as represented in l step 110 of FIG. 2a, this arbitrary value being denoted STAT (O), to be calculated in a step 111 for this current block B _n an intermediate stationarity parameter value, denoted STAT * (n), function of a number determined from successive analysis choice values, these LPC analysis choice values, denoted d _n-1 (n-1), ..., to d _np (np), being obtained for different successive blocks prior to the block current B _n of the succession of blocks LPC, and of the value of the stationarity parameter of the block preceding the current block, this stationarity value being denoted STAT (n-1). In step 111 represented in FIG. 2a, it is indicated that the function of the determined number of earlier analysis choice values is given in relation to these earlier values, denoted d _n-1 (n-1) to d _np (np ). With regard to the arbitrary starting value of the stationary parameter STAT (0), it is indicated that this can, by way of nonlimiting example, be taken equal to the average value between the maximum value and the minimum value of the stationarity parameter previously mentioned in the description, STAT _M and STAT _m .

The aforementioned step 111 is then followed by a step 112, which consists in refining the value of the intermediate stationarity parameter as a function of the value of the prediction gains of the filters or LPC analysis mode "front" and "rear" of the frame preceding the current frame. In step 112 of FIG. 2a, it is indicated that the above-mentioned function is denoted g (STAT * (n), Gpf, Gpb) where Gpf denotes the prediction gain of the LPC filter " before " and Gpb denotes the prediction gain the "back" LPC filter for the frame preceding the current frame. In step 112, that is to say at the end of the step consisting in refining the value of the intermediate stationarity parameter, the value of stationary parameter STAT (n) of the current LPC block B _n is assigned the value , relation (3): STAT (n) = g (STAT * (n), Gpf, Gpb) corresponding to the refined value of the intermediate stationarity parameter.

A more detailed description of the calculation step 111 of the intermediate stationary parameter STAT * (n) and of step 112 consisting in refining this parameter value will now be given in connection with FIG. 2b.

In accordance with the aforementioned figure, step 111 consists, from an initialization step 1110 in which the value of the stationarity parameter STAT (n-1) and the value of analysis choice d _n-1 (n -1) relating to the block LPC B _n-1 prior to the current block B _n is available, to be carried out, in a step 1111, a step consisting in discriminating the analysis mode LPC " front " or LPC "rear" of block B _n-1 preceding the current block B _n . This discrimination step 1111 can, as shown in FIG. 2b, consist of a test step on the value of choice of analysis d _n-1 (n-1) with respect to the symbolic value "fwd" or to the value logic zero corresponding to the value complemented by the true logic value.

On negative response to the aforementioned test 1111, that is to say for any block B _n-1 preceding the current block LPC B _n analyzed in “backward” LPC analysis mode , the step of calculating the parameter value of intermediate stationarity consists, in a step 1113, in determining the number of anterior frames analyzed consecutively in LPC analysis mode " rear ", number noted N_BWD, then, in a step 1114, in comparing on criterion of comparison of superiority the number of frames prior to a first arbitrary value, denoted Na, representative of a number of successive frames analyzed in "backward" LPC mode .

On a positive response to the comparison of superiority of test 1114, the calculation step then consists in assigning, in a step 1114b, to the value of the intermediate stationarity parameter STAT * (n), the value of the stationarity parameter of the preceding block the current block, STAT (n-1), increased by a determined value as a function of the first arbitrary value representative of a number of successive frames analyzed, that is to say in fact of the number of previous frames N_BWD analyzed consecutively in LPC analysis mode " back ". In step 1114b, the determined value as a function of the first arbitrary value is denoted f _a (N_BWD). During the aforementioned step, it is understood that the value of the intermediate stationarity parameter STAT * (n) for the current LPC block B _n is thus increased relative to the corresponding value of the same stationarity parameter for the previous block B _{n -1} .

On a negative response to the comparison of superiority to the comparison test 1114, the value of the intermediate stationarity parameter STAT * (n) is assigned, in a step 1114a, the value of the stationarity parameter STAT (n-1) of the preceding block the current block B _n .

On the contrary, for any previous block B _n-1 analyzed in LPC analysis mode "before", that is to say on a positive response to test 1111, the step of calculating the intermediate stationarity parameter 111 consists, thus as shown in FIG. 2b, to be determined in a step 1112, on test criteria, the occurrence of a transition from the "back" LPC analysis mode to the " front " LPC analysis mode between the anterior block and the anterior block to the current block B _n-1 , of rank n-2, that is to say the existence of a value of choice of analysis LPC d _n-2 (n-2) = symbolic value "bwd", either logical value zero as mentioned previously. The positive response to test 1112 indicates the existence of such a transition from the "backward" analysis mode for the LPC block B _n-2 preceding the block preceding the current block B _n-1 , while a negative response to the test 1112 cited above indicates the absence of such a transition.

On a positive response to the aforementioned occurrence test 1112, the calculation step 111 then consists in comparing, on the basis of an inferiority comparison criterion, the number of anterior frames N_BWD mentioned above with a second arbitrary value N _b representative of a number of successive frames analyzed in "backward" LPC mode preceding block B _n-1 preceding the current block.

On a positive response to the comparison carried out in test 1118, this test is followed by a step 1118a consisting in assigning to the value of the intermediate stationarity parameter STAT * (n) the value of the stationarity parameter of the block preceding the current block, STAT (n-1) reduced by a determined value, function of the second arbitrary value N _b , this determined value being noted f _b (N_BWD). It is thus understood that during the allocation step 1118a, the value of the intermediate stationarity parameter is thus reduced accordingly.

On the contrary, on a negative response to the comparison of inferiority carried out in test 1118, step 111 consists then to assign, in a step 1118b, to the value of intermediate stationary parameter STAT * (n) the value of stationarity parameter of the block preceding the block current, i.e. STAT (n-1).

In FIG. 2b, it will be noted that the allocation steps 1118a and 1118b are then followed by a step for resetting to zero the number of successive blocks processed in "backward" LPC analysis mode , this step of zeroing carrying the reference 1118c and making it possible to update the entire process for calculating the value of the intermediate stationarity parameter.

On a negative response to comparison test 1112, no LPC analysis transition " before " having appeared, the value of the intermediate stationary parameter STAT * (n) is assigned the value of the stationary parameter STAT (n-1) of the block previous B _n-1 in a step 1119.

At the end of step 111, there is the value of the intermediate stationarity parameter STAT * (n) for the current block B _n .

As regards step 112 consisting in refining the value of the aforementioned intermediate stationarity parameter, it is indicated, with reference to FIG. 2b, that this can advantageously consist, in a step 1120, in discriminating the prediction gains from the LPC filtering "backward" LPC filtering and "advan t", the gain values being denoted Gpb and Gpf respectively. It will be understood that the aforementioned discrimination step simply consists in storing and reading the gain values calculated for the aforementioned "front" and "back" LPC filtering. In addition to the aforementioned gain values, step 1120 can consist in calculating the relative value of the prediction gains, denoted DGfb, such as the difference or the ratio between the aforementioned "front" and "rear" prediction gains.

As shown in addition in FIG. 2b, step 112 of FIG. 2a comprises, after the above-mentioned step 1120, a step 1121 consisting in modifying the value of the intermediate stationarity parameter STAT * (n) a refinement value ΔS, this refinement value in accordance with a particularly remarkable characteristic of the method which is the subject of the present invention being a function of the relative value of the LPC filtering prediction gains "front" and "rear".

• ΔS = f_r(Gpf, Gpb) où Gpf et Gpb désignent ainsi que précédemment les gains de prédiction du filtrage LPC "avant" respectivement "arrière".

In general, it is indicated that the function representative of the refinement value ΔS is noted:

• ΔS = f _r (Gpf, Gpb) where Gpf and Gpb thus designate as previously the prediction gains of the LPC filtering "front" respectively "rear".

In general, it is indicated that the function f _r (GPf, Gpb) making it possible to establish the refinement value ΔS is an increasing respectively decreasing function of this relative value, according to the direction in which this relative value is considered. When the relative value designates the value of the "back" LPC filtering gain compared to the "front" LPC filtering gain , this choice can be arbitrarily retained without in any way detracting from the generality of the process, object of the invention, relative value DGfb above, the function f _r is then increasing. Otherwise, it decreases.

In other words, the modification, by increase or by decrease, of the value of the intermediate stationarity parameter of the ripening value ΔS is proportional to this relative value of the gains. Generally, this modification is written STAT (n) = STAT * (n) + kΔS. In practice we will take k = 1. In a more specific way, we indicate that the refinement value ΔS increases in algebraic value when the difference between the prediction gains of the LPC filtering "front" and "rear" increases, the function f _r (GPf, Gpb) then being an increasing function, while this refinement value ΔS decreases in algebraic value when this same abovementioned deviation decreases, the abovementioned deviation being defined between the "backward" LPC filtering prediction gain and the LPC filtering prediction gain " before ". In fact, this function is increasing or decreasing depending on the definition of this difference.

Consequently, at the end of step 1121 as shown in FIG. 2b, the value of the stationarity parameter intermediate STAT * (n) can then, for k = 1, be corrected by the algebraic value of the refinement value ΔS above to calculate the value of the stationarity parameter STAT (n).

Following step 1121, we thus have the value of the stationarity parameter STAT (n) in step 1122.

A more detailed description of step 1121 of FIG. 2b will now be given in conjunction with FIG. 2c in a preferred embodiment in which a plurality of test criteria are applied both to the refining value and to the values of "forward" and "backward" LPC prediction gain in order to optimize the stationarity parameter calculation process.

As shown in FIG. 2c above, step 1121 can consist of a first step 1121a making it possible to calculate the refinement value ΔS from the function f _r (Gpf, Gpb) previously cited. Different examples of usable functions will be given later in the description.

First, the refinement value ΔS is subject to a comparison test of superiority to the value 0, in a step 1121b, this comparison test in fact allowing to determine the increase in this refinement value ΔS.

On a positive response to the aforementioned test 1121b, the refinement value ΔS being positive and corresponding to an increase in the relative value of the LPC filtering prediction gains " front " and " rear ", the step of increasing the value of intermediate stationarity parameter of the refinement value ΔS is further subject to a condition of superiority of the “ backward ” filtering gain value LPC, compared with a first determined positive value, in a step of comparing the superiority of the value of the "back" LPC filtering gain Gpb with respect to this first determined positive value, denoted S _i .

On a negative response to the aforementioned test 1121c, to the stationary parameter value STAT (n) is assigned the value of the intermediate stationarity parameter STAT * (n) in one step 1121g.

On a positive response to the aforementioned test 1121c, the increase in the value of the intermediate stationarity parameter of the ripening value ΔS is furthermore subject to a condition of inferiority of the value of the intermediate stationarity parameter STAT * (n) by compared to a second determined positive value STAT _i of course representing a stationarity value. This inferiority condition test is carried out in step 1121e.

On a negative response to the aforementioned test 1121e, to the value of the intermediate stationary parameter STAT (n) the value of the stationarity parameter is assigned intermediate STAT * (n) in the aforementioned step 1121g.

On positive response to the inferiority condition test 1121e to the value of the stationarity parameter intermediate STAT (n) is assigned the value of the parameter STAT * (n) intermediate stationarity increased by positive value ΔS of the ripening value in step 1121i.

On the contrary, on a negative response to the aforementioned test 1121b, the refinement value ΔS being negative, the step of decreasing the intermediate stationarity parameter by the refinement value ΔS, this value being negative, is also subjected to a test. of inferiority condition of the gain value of the "back" LPC filtering Gpb with respect to a third determined positive value denoted S _d in a comparison step 1121d. This third determined positive value is of course representative of an LPC filtering gain value.

On a negative response to test 1121d mentioned above in stationary parameter value STAT (n) is assigned the value of the intermediate stationarity parameter STAT * (n) in step 1121g.

On the contrary, on a positive response to the 1121d test mentioned above, the step of decreasing the value of the parameter of intermediate stationarity of the ripening value ΔS is further subject to a value superiority condition of the intermediate stationarity parameter STAT * (n) by compared to a fourth determined positive value, noted STATd in a comparison test noted 1121f. Of course, the fourth determined positive value is representative of a selected stationarity parameter value.

On a negative response to the aforementioned test 1121f, to the stationary parameter value STAT (n) is assigned the value of the intermediate stationarity parameter STAT * (n) in step 1121g.

On a positive response to the aforementioned test 1121f, to the stationary parameter value STAT (n) is assigned the value of the STAT * intermediate stationary parameter increased by the algebraic value by the refinement value ΔS, negative, the value of the intermediate stationarity parameter thus being reduced to establish the value of stationarity parameter STAT (n) in step 1121h.

At the end of steps 1121g, 1121h and 1121i, we have thus in step 1122 of FIG. 2b of the parameter of STAT (n) stationarity.

With regard to the function f _r (Gpf, Gpb), it is indicated that this can consist of. a non-linear function of the relative value of the "forward" and "backward" LPC filtering gains in which the relative value of the "forward" and "backward" LPC filtering prediction gains may itself consist of either the ratio or unlike the " forward " and " backward " LPC filtering prediction gains. Other types of functions, such as linear functions, can be used.

A first example of a non-linear function f _r (Gpf, Gpb) is shown in Figure 2d.

In the exemplary embodiment of FIG. 2d, pairs of values of the LPC filtering prediction gain "rear" Gpb plotted on the ordinate and the LPC filtering gain gain " before " Gpf make it possible to assign positive refinement values ΔS , ΔS> 0 or negative ΔS <0 for a value of the ratio ρ = Gpb / Gpf corresponding to a higher slope respectively lower than that of the straight line ΔS = 0.

In FIG. 2e, we have represented the case where the relative value of the LPC filtering prediction gains "front" and "rear" no longer corresponds to the ratio of the gains ρ but to the difference of the aforementioned gains.

In this case, the function of the relative value of the prediction gains from the LPC filtering " front " and "rear" f _r (Gpf, Gpb) can also be a non-linear function allowing the refinement value ΔS to be assigned to values of this difference corresponding to pairs of value Gpb, Gpf corresponding to straight lines whose abscissa at the origin is respectively respectively greater, in algebraic value, than the abscissa at the origin of the line ΔS = 0.

In the case of Figure 2e, the lines delimiting zones according to the sign of the ripening value ΔS are parallel to each other.

In accordance with another particular aspect of the method which is the subject of the invention, it is further indicated that the stationarity index of the current block B _n should not be adapted during silence frames, when for example the audio frequency signal is constituted by a speech signal including silences. In such a case, step 1111 of step 111 shown in FIG. 2b can be preceded by a step 1111a consisting, for each successive current block, in determining the average energy of the digital audio frequency signal and in comparing in this same step ,. on an inferiority comparison criterion, this average energy at a determined threshold value representative of a frame of silence. On figure 2b, this threshold value is noted ENER_SIL. On positive response to the aforementioned test, the value of the stationarity parameter of the current block STAT (n) is assigned the value of the stationarity parameter of the previous block STAT (n-1) in the allocation step 1111b represented in the figure. 2b. The steps 1111a and 1111b are, in the above-mentioned figure, shown in dotted lines, since they are reserved for example for coding a speech signal.

A more detailed description of the implementation of the decision function D _n allowing the obtaining of the decision values d _n (n) will now be given in connection with FIG. 2f. This description is given in a preferred embodiment in which this decision function, which can be compared to that described in the article previously mentioned by the description, published by S.PROUST, C.LAMBLIN and D.MASSALOUX, is however adapted temporally, in accordance with the object of the present invention in order to obtain the successive analysis choice values d _n (n).

From a step 120, for the current block B _n , a distance, denoted d _LPC , is first calculated between the filter LPC of the current block and that of the previous block B _n-1 . This distance calculation is carried out for example using the LSP frequency parameters as mentioned previously in the description relative to the method described in the aforementioned article.

• S_PRED(n) et S_TRANS, S_STAT et G₁ les valeurs des seuils intervenant dans le critère fondé sur les gains de prédiction des filtres LPC "arrière" et "avant" ;
• S_LSP_L et S_LSP_H les valeurs des seuils intervenant dans le critère fondé sur les distances entre vecteurs de fréquence LSP représentant deux filtres. LPC "avant" relatifs à deux blocs consécutifs B_n-1 et B_n ;
• Gpf le gain de prédiction du filtre LPC "avant" ;
• Gpb le gain de prédiction du filtre "arrière" ; et
• Gpi le gain de prédiction du filtre "avant" interpolé selon la méthode exposée dans l'article publié, mentionné précédemment dans la description.

We notice :

• S_PRED (n) and S_TRANS, S_STAT and G ₁ the values of the thresholds intervening in the criterion based on the gains in prediction of the LPC filters "rear" and "front" ;
• S_LSP_L and S_LSP_H the values of the thresholds intervening in the criterion based on the distances between vectors of frequency LSP representing two filters. LPC " before " relating to two consecutive blocks B _n-1 and B _n ;
• Gpf the prediction gain of the LPC filter "before";
• Gpb the prediction gain of the "back" filter ; and
• Gpi the prediction gain of the " before " filter interpolated according to the method exposed in the published article, mentioned previously in the description.

• si les filtres LPC consécutifs sont très stationnaires, c'est-à-dire pour d_LPC < S_LSP_L, alors, aucun basculement du filtrage LPC "arrière" au filtrage LPC "avant" n'est réalisé si l'on est en mode de filtrage LPC "arrière", à condition que le gain de prédiction du filtre LPC "arrière" soit supérieur au gain de prédiction du filtre LPC "avant" diminué d'une valeur S_STAT. On indique que la valeur S_STAT est choisie de manière à favoriser le choix d'un filtre LPC "arrière" en présence d'une grande stationnarité du spectre mesurée à l'aide de la distance d_LPC ;
• si les filtres LPC consécutifs présentent une transition importante, c'est-à-dire pour d_LPC > S_LSP_H et si Gpf > Gpb-S_TRANS, alors, le mode de filtrage choisi est le filtrage LPC "avant", c'est-à-dire d_n(n) = 0, valeur symbolique "fwd", sinon, d_n(n) est pris égal à 1, valeur symbolique "bwd". On indique que la valeur de S_TRANS est choisie de manière à favoriser fortement le choix du filtre LPC "avant" en présence d'une transition de spectre mesurée à l'aide de la distance d_LPC ;
• sinon, dans tous les autres cas, si Gpb > Gpf-S_PRED et Gpi > Gpf-S_PRED, alors, le filtre LPC retenu est le filtre LPC "arrière" interpolé, à condition que le gain de ce dernier et celui du filtre LPC "arrière" pur excèdent la valeur de seuil G₁ précédemment mentionnée. Si la condition sur les valeurs de gain de prédiction précitées n'est pas remplie, alors, on choisit le filtrage LPC "avant".

The criterion for establishing the decision function, in relation to FIG. 2f, is established as follows:

• if the consecutive LPC filters are very stationary, that is to say for d _LPC <S_LSP_L, then, no switching from LPC filtering " back " to LPC filtering " before " is carried out if one is in " rear " LPC filtering, provided that the prediction gain of the "rear" LPC filter is greater than the prediction gain of the "front" LPC filter reduced by a value S_STAT. It is indicated that the value S_STAT is chosen so as to favor the choice of a "rear" LPC filter in the presence of a large stationarity of the spectrum measured using the distance d _LPC ;
• if the consecutive LPC filters have a significant transition, that is to say for d _LPC > S_LSP_H and if Gpf> Gpb-S_TRANS, then, the filtering mode chosen is the LPC filtering "before", that is to say -d say d _n (n) = 0, symbolic value "fwd" , otherwise, d _n (n) is taken equal to 1, symbolic value " bwd ". It is indicated that the value of S_TRANS is chosen so as to strongly favor the choice of the LPC filter " before " in the presence of a spectrum transition measured using the distance d _LPC ;
• otherwise, in all other cases, if Gpb> Gpf-S_PRED and Gpi> Gpf-S_PRED, then the LPC filter used is the interpolated "back" LPC filter, provided that the gain of the latter and that of the LPC filter " rear " pure exceed the threshold value G ₁ previously mentioned. If the condition on the aforementioned prediction gain values is not fulfilled, then the “before” LPC filtering is chosen .

In order to increase the number of " forward " LPC filters transmitted and thus increase the robustness of the coding system for transmission errors, the " forward " LPC filtering mode can be advantageously chosen as soon as the energy of the signal to be encoded E _n , that is to say the energy of the corresponding block B _n , becomes lower than the value of the energy of a frame of silence ENER_SIL, this value of energy corresponding to the minimum audible level.

The set of conditions allowing the establishment of the decision function D _n and the obtaining of the corresponding analysis choice values d _n (n), is illustrated in FIG. 2f with temporal adaptation of the decision function D _n .

The value of the stationary parameter STAT (n) can for example be located on a scale of 0, corresponding to the value STAT _m very little stationary, to 100, corresponding to the value STAT _M very stationary.

Depending on the value of the stationary parameter STAT (n), the decision function D _n is modified by adapting the value of the thresholds.

The more the stationarity of the signal increases, the more the LPC filtering mode "rear" is favored: the thresholds S_PRED, S_LSP_L and S_LSP_H are increased.

• S_PRED(n) = f_{S_PRED}(STAT(n)) avec f_{S_PRED} fonction croissante de la valeur de STAT(n) ;
• S_LSP_L(n) = f_{S_LPC_L}(STAT(n)) avec f_{S_LPC_L} fonction croissante ;
• S_LSP_L(n) = f_{S_LPC_H}(STAT(n)) avec f_{S_LPC_H} fonction croissante.

By way of nonlimiting example, the modification functions are indicated for each current LPC boc B _n of the aforementioned threshold values:

• S_PRED (n) = f _{S_PRED} (STAT (n)) with f _{S_PRED} increasing function of the value of STAT (n);
• S_LSP_L (n) = f _{S_LPC_L} (STAT (n)) with f _{S_LPC_L} increasing function;
• S_LSP_L (n) = f _{S_LPC_H} (STAT (n)) with f _{S_LPC_H} increasing function.

In the adaptation of the aforementioned threshold values, it is indicated that the increasing functions mentioned are for example staircase functions as regards the functions f _{S_LPC_L} and f _{S_LPC_H} . The function f _{S_PRED} is an affine function of the stationarity parameter variable, of the form: S_PRED (n) = α.STAT (n) + β where α and β are two real values between 0 and 1 and where the value of S_PRED (n) is bounded in the interval [S_PRED _m , S_PRED _M ], S_PRED _m and S_PRED _M represent two values determined experimentally.

In order to further limit the risk of filter tilting, it is then possible to choose, when the stationarity parameter STAT (n) is less than a given threshold value S _FWD , to impose the LPC filtering mode " before " .

On the other hand, the threshold values S_TRANS, S_STAT and G ₁ keep a fixed value, these values being able for example to be equal to -1 dB, 5 dB and 0 dB respectively.

The establishment of the decision function D _n and the obtaining of the analysis choice values d _n (n) are illustrated in the following manner in FIG. 2f: following the above-mentioned step 120, carrying out a step of test 121 relating to the energy of the current LPC block B _n , by a comparison of inferiority to the value of energy of silence ENER_SIL or of the value of the stationarity parameter STAT (n), compared by a comparison of inferiority to the value S _FWD previously cited in the description. On a positive response to the aforementioned test 121, the value of choice of analysis d _n (n) is taken equal to 0, that is to say symbolic value "fwd" in step 122.

On negative response to the aforementioned test 121, a new test is carried out relative to the value of choice of analysis d _n-1 (n-1) with the logical value 1, that is to say with the symbolic value " bwd ".

On a positive response to the aforementioned test 123, a new test is performed on the aforementioned filtering distance LPC d _LPC , in a step 124, relative to the threshold value S_LSP_H (n) by comparison of superiority to this threshold value.

On a positive response to the aforementioned test 124, a new test 126a is carried out, consisting in comparing the prediction gain of the LPC filtering " before ", Gpf, to the prediction gain of the LPC filtering "back", Gpb, reduced by the threshold value S_TRANS.

On a positive response to the above test 126a, the analysis value d _n (n) is assigned the logical value 0, symbolic value " fwd ", and on a negative response to the above test 126a, is assigned to the same value choice of analysis the logical value 1, symbolic value " bwd ". The corresponding steps are noted 128 and 129.

On a negative response to the test 124 previously mentioned, a new test 125 is carried out. Test 125 consists in making a comparison of the filtering distance LPC, d _LPC , by comparison of inferiority to the threshold value S_LSP_L (n).

On a positive response to test 125, a new test 126b is carried out by comparing the superiority of the LPC filtering prediction gain "back" to the LPC filtering prediction gain "before" reduced by the value S_STAT previously mentioned.

On a positive response to test 126b, the value of analysis choice d _n (n) is assigned in step 129 the logical value 1, that is to say the symbolic value " bwd ".

On a negative response to test 126b, the value of choice of analysis d _n (n) is assigned the logical value 0, that is to say the symbolic value "fwd", step 128.

On the contrary, on a negative response to test 125, a new test is carried out, in a step 127, this test consisting in verifying the conditions of comparison of the "back" LPC filtering gain Gpb with the LPC filtering prediction gain "before" decreased of the threshold value S_PRED (n), of comparing the superiority of the intermediate LPC filtering prediction gain Gpi to the value of the LPC filtering prediction gain " before " minus the aforementioned threshold value S_PRED (n) and of comparison of superiority of the “backward” filtering prediction gain Gpb to the threshold value G ₁ , as well as of comparing the value of the intermediate filtering prediction gain Gpi to the threshold value G ₁ .

We indicate that the negative response to test 123 previously mentioned in the description also brings when carrying out the above test 127.

On a positive response to the test 127 previously mentioned, the value of analysis choice d _n (n) is assigned the logical value 1, that is to say the symbolic value "bwd" in step 129, while qu 'to the negative response to the above test 127, to the analysis choice value d _n (n), on the contrary, is assigned the logical value 0, that is to say the symbolic value "fwd" in step 128 .

We thus have, thanks to the implementation of the decision function D _n , the value of choice of analysis d _n (n) obtained with the aforementioned logical values 1 or 0, these logical values being however linked to a value of priority or lack of priority to return to the "rear" or "front" filtering mode depending on the value of the stationarity parameter.

A more detailed description of a device for coding a digital audio signal by double analysis on the LPC analysis selection criterion " front " respectively "rear" into a coded transmitted signal, in accordance with the object of the present invention, will now be given in connection with FIG. 3.

In a practical way, it indicates that the signal digital to be coded is subdivided into frames consisting of successive blocks of samples, each block comprising a given number N of samples for example.

In FIG. 3, the mode of constitution of the digital audio-frequency signal to be coded in successive blocks of samples B _n has not been shown because this operating mode is perfectly known from the state of the art and can be carried out from 'a simple buffer memory, for example read periodically at the frame frequency and the block frequency.

As also shown in FIG. 3 above, the coding device which is the subject of the invention comprises a " front " LPC analysis filter, bearing the reference 1A, and a "rear" LPC analysis filter , bearing the reference 1B, in order to enable a transmitted coded signal consisting of LPC filtering parameters accompanied by an analysis decision indication to be delivered, as well as parameters Pr _n relating to the harmonic analysis and to the CELP excitation signal.

In general, it is indicated that the analysis decision indication corresponds to the analysis choice value d _n (n) as mentioned previously in the description. As regards the LPC filtering parameters, it is indicated that these correspond to specific parameters, in accordance with the mode of implementation of the coding method which is the subject of the present invention, as will be described below in the description.

In FIG. 3, the existence of an adaptive filter as a function of the value of the stationarity parameter has also been shown in the coding device according to the invention, this adaptive filter bearing the reference 1E. This adaptive filter 1E naturally receives the original digital signal, noted s _{n (t)} , that is to say the current block B _n . The 1E filter uses the LPC filtering parameters to calculate the residual signal which will then be coded by the 1F module. These LPC parameters, as well as the filter decision indication constitute a part of the coded signal which is transmitted to the decoder.

In addition, as shown in FIG. 3, the coding device which is the subject of the present invention comprises a coding means, bearing the reference 1F, of a residual coding signal not transmitted, the residual coding signal, designated by res _{n (t)} is directly available at the output of the adaptive filter 1E, this signal thus being delivered at the input with the digital audio frequency signal to the coding module of the non-transmitted coding residue signal, to generate a synthesis residue signal, res_syn _n (t).

A reverse filtering module, bearing the reference 1G, receives the synthesis residue signal and makes it possible to deliver a synthesis signal referenced s_syn _{n (t)} .

A storage module 1H receives the aforementioned synthesis signal s_syn _{n (t)} to deliver the aforementioned synthesis signal for the block prior to the current block B _n , the synthesis signal thus obtained being designated by s_syn _n-1 (t). This synthesis signal is delivered to the "rear" LPC analysis filter bearing the reference 1B in FIG. 3 above.

The coding device, object of the present invention, as shown in FIG. 3, makes it possible to carry out coding of the digital audio frequency signal on the aforementioned digital audio signal using the " front " LPC filter for the non-stationary zones and on the aforementioned synthesis signal s_syn _n-1 (t) from the “rear” LPC filter 1B for the stationary zones, as will be described below.

As will be observed in FIG. 3 above, the device which is the subject of the invention comprises for this purpose, for each current LPC block B _n , a module 1C for calculating the degree of stationarity of the digital audio signal according to a parameter of stationarity whose value is between a maximum stationarity value and a minimum stationarity value. Of course, the stationarity parameter is the STAT (n) parameter previously described in the description in accordance with the coding method which is the subject of the present invention. The maximum and minimum stationarity values are also defined above.

As shown in addition in FIG. 3, the coding device which is the subject of the invention comprises a module, denoted 1D ₁ , for establishing from the above-mentioned stationarity parameter STAT (n) a function of decision and an LPC analysis choice value, the decision function being denoted D _n as mentioned previously in the description, and the LPC analysis choice value being of course and corresponding to the choice value of LPC analysis noted d _n (n) previously described in the description. It is recalled that the value of choice of analysis d _n (n) can take the values 0 or 1, logical values, which correspond to the symbolic value of choice of analysis "fwd" and "bwd" for LPC analysis " front " and" rear "respectively.

It is understood in particular that as regards the means for establishing the decision function D _n , this corresponds to a software embodiment for example, as described above in connection with FIG. 2f. In addition, the coding device according to the invention as shown in FIG. 3 comprises an LPC filtering analysis discrimination module, denoted 1D ₂ , this module receiving the analysis choice value d _n (n) and allowing to deliver, for the current LPC block B _n, the value of the LPC filtering parameters "rear" respectively "front" according to the above-mentioned analysis choice value. It will of course be understood that the " back " LPC filtering analysis parameters as well as the " front " LPC analysis filtering parameters are of course available in digital form at the filters bearing the reference 1B and 1A respectively in the figure. 3. These parameters are designated respectively Af _n (z) for the LPC filter analysis parameters "before" with regard to the LPC analysis filter " before ", bearing the reference 1A, and by Ab _n (z) for the "rear" LPC analysis parameters with regard to the "rear" LPC analysis filter bearing the reference 1B. These parameters are delivered to the 1D ₁ module and to the 1D ₂ module respectively.

With regard to the hardware embodiment of the discrimination module 1D ₂ , it is indicated that it can for example, in a non-limiting embodiment, consist of two distinct memory zones allowing the memorization of the filtering parameters Af _n (z) and Ab _n (z) respectively, the analysis choice value d _n (n) as a function of its logical current value, 0 or 1, allowing the addressing in reading of the values of filtering parameters stored by the module 1D ₂ for example and the transmission of these filtering parameters by the latter.

Finally, as shown in FIG. 3, it is indicated that the coding device in accordance with the object of the present invention, for producing the adaptive filter as a function of the stationarity value carrying the reference 1E, can be produced by a filter element whose transfer function, denoted A (z), is established from the values of filter parameters delivered by the discrimination module 1D ₂ previously mentioned.

It is thus understood that the adaptive filtering module 1E can be produced by a filter with adjustable coefficients, to the value of the coefficients of the latter being assigned the values of filtering parameters delivered by the discrimination module 1D ₂ previously mentioned. The filtering performed by the module 1E is thus of the adaptive type as a function of the degree of stationarity of the digital audio frequency signal to be coded. The module 1E thus delivers, from the original digital audio signal s _{n (t)} , the residual filtering signal LPC designated by res _n (t) to the coding module for the residue 1F, which then makes it possible to deliver the residual signal LPC synthesis designated by res_syn _n (t).

Finally, the module 1G is a filtering module whose transfer function is the inverse of the transfer function of the module 1E obtained from the memorized parameters of the latter. It receives the LPC synthesis residue signal res_syn _n (t) delivered by the coding module from the coding residue delivered by the module 1F. It is thus understood that the coding of the digital audio signal s _n (t) is carried out at the level of the module 1E by virtue of the LPC analysis "front", respectively "rear" carried out by the LPC analysis filters "before" 1A and d analysis LPC "back" 1B, the coded signal s_c _n (t) consisting of the transmission of filtering parameters LPC "before" when the value of analysis choice d _n (n) has the symbolic value "fwd" as well as the indication of the choice of analysis, that is to say of the value of the choice of analysis previously cited. This operating mode makes it possible to carry out the coding of the digital audio-frequency signal and to favor the maintenance in one of the LPC filtering modes "front", respectively "rear", according to the degree of stationarity of the digital signal and to further limit the number of switches from one to the other of the filtering modes considered.

A device for decoding a digital audio signal coded in double analysis on the criterion of choice of LPC analysis "front", respectively "rear", into a coded signal transmitted in accordance with the coding method object of the present invention, and thanks to the implementation of a coding device as shown in FIG. 3 for example, will now be described in conjunction with FIG. 4.

In general, it is indicated that the transmitted coded signal s_c _n (t) consists for each analysis block LPC in the value of choice of analysis mentioned above and, in the case where the value of choice of analysis corresponds for the LPC analysis block considered in a " before " LPC analysis, in the "before" LPC filtering parameters as well as the LPC filtering residue coding parameters, parameters Pr _n , that is to say of the res signal _n (t) into a synthesis residue signal res_syn (t) by the coding module of the residue 1F.

As shown in FIG. 4, it is indicated that the decoding device comprises at least one synthesis module, referenced 2A, of the filtering residue signal receiving the coding parameters of the LPC residue delivered by the module 1F. The module 2A decodes the coding parameters supplied by the module 1F and consequently delivers a synthesis residue signal, which is referenced in FIG. 4 res_syn _n (t).

The decoding device as shown in FIG. 4 also includes a module, bearing the reference 2B, of adaptive reverse filtering as a function of the degree of stationarity, receiving the above-mentioned synthesis residue signal, delivered by the module 2A, and allowing d 'generating a synthesis signal s_syn _n (t) representative of the digital audio frequency signal, this signal constituting in fact the decoded signal.

It will of course be understood that the reverse filtering module 2B implements the filtering parameters received by the decoder due to the transmission, ie the LPC analysis parameters " before " when these are transmitted and the decision to analysis corresponds to a “ forward ” LPC analysis or, on the contrary, the “rear” filtering analysis parameters as will be described below.

To this end, the decoding device which is the subject of the present invention of course comprises a "rear" LPC filter module , carrying the 2D reference, receiving the synthesis signal, that is to say the signal referenced s_syn _n (t ) for the LPC block prior to the current LPC block, this synthesis signal thus being referenced s_syn _n-1 (t) in FIG. 4. It is understood for this purpose that the synthesis signal relating to the current block B _n and referenced s_syn _n (t) can then be delivered to the 2D "rear" LPC filtering module by means of a storage module, bearing the reference 2E, making it possible in fact, by addressing in read-appropriate mode, to offset the reading of the signal from synthesis to that corresponding to the block preceding the current block B _n .

Finally, and to ensure the above-mentioned operating mode, the decoding device which is the subject of the present invention, as shown in FIG. 4, finally comprises a discriminator module bearing the reference 2C, making it possible to discriminate the LPC analysis "before ", respectively " rear ". The module 2C receives, on the one hand, for a discrimination command, the value of analysis choice received, that is to say the value d _n (n), and, on the other hand, the filtering parameters " Front " LPC, ie the parameters Af _n (z) transmitted, as well as the "rear" LPC filtering parameters Ab _n (z) obtained by means of the 2D module. The module 2C thus makes it possible to deliver, as a function of the value of choice of analysis, that is to say of the value d _n (n), that is to say the filtering parameters LPC "before" Af _n (z), or the "rear" LPC filtering parameters Ab _n (z) to the adaptive reverse filtering module 2B as a function of the degree of stationarity.

As regards the hardware embodiments of the modules 2C and 2B, it is indicated that these can simply consist of modules substantially identical to the modules 1D ₂ and 1E or, more particularly, 1G of FIG. 3.

Regarding the effective realization of a coding device in accordance with the subject of this invention, allowing the implementation of the method as described previously in the description, two modes of specific achievements have been made.

* CELP coder in telephone band, according to broadband extension of the ITU-T standard to 8 kb / s :

The actual coder consisted of a telephone band from 300 to 3400 Hz, at a speed of 12 kb / s CELP type. The frames were formed over a period 10 ms for excitation provided by dictionary algebraic according to the technique called ACELP previously mentioned in the description.

The " forward " LPC analysis was a 10-order analysis and the "back" LPC analysis a 30-order analysis every 80 samples.

A separation for coding the residue into two sub-blocks of 40 samples was carried out. Each block B _n contained 80 samples.

Adaptation of the stationary parameter STAT (n)

The above-mentioned stationarity parameter varies between two extreme values 0 and 100, the aforementioned values STAT _m and STAT _M.

Dans ces relations, x = DGfb.The adaptation functions previously described in the description, and in particular the functions f _a (N_BWD) and f _b (N_BWD) were such that:

In these relationships, x = DGfb.

As regards the function f _r making it possible to establish the refinement value ΔS previously mentioned in the description, this is a step function of the variable x, with x = Gpb - Gpf and ΔS = f _r (x ) and having the value:

The refining of STAT (n) is also subject to the following conditions mentioned above in relation to FIG. 2c:
If ΔS> 0:
If STAT * (n) <STAT _i STAT (n) = STAT * (n) + ΔS
Otherwise STAT (n) = STAT * (n)
If not :
STAT (n) = STAT * (n)
with STAT _i = 40.6.

The other test conditions referenced 1121d, 1121c and 1121f in Figure 2c were not used in this embodiment.

Adaptation of decision thresholds

La valeur du seuil S_STAT utilisée en cas de stationnarité des filtres LPC mesurée à l'aide du seuil S_LSP_L a été fixée à 4,0 dB.
Le seuil S_LSP_H n'a pas été utilisé dans ce mode de réalisation.
La valeur du seuil G₁ a été fixée à OdB.
En ce qui concerne la valeur d'énergie caractérisant une trame de silence ENER_SIL, cette valeur a été fixée à 40 dB mesurée sur les 80 échantillons s(i) du bloc courant B_n :

Regarding decision thresholds:
S_PRED is adapted as follows:
S_PRED (n) = 0.03.STAT (n) + 1.0
S_PRED ∈ [S_PRED _m , S_PRED _M ], S_PRED _m = 1.03 and S_PRED _M = 4;
The threshold S_LSP_L is adapted using the following staircase function:

The value of the threshold S_STAT used in case of stationarity of the LPC filters measured using the threshold S_LSP_L has been fixed at 4.0 dB.
The threshold S_LSP_H was not used in this embodiment.
The value of the G ₁ threshold has been set at OdB.
Regarding the energy value characterizing an ENER_SIL silence frame, this value was set at 40 dB measured on the 80 samples s (i) of the current block B _n :

With regard to the value of the threshold S _FWD mentioned above and intended to further limit the risk of switching over by imposing the LPC filtering mode " before " when the value STAT (n) is less than this threshold, this value S _FWD has was set at 40.6.

* A second embodiment of a CELP type encoder in a band extended to two 16/24/32 kb / s sub-bands was carried out under the following conditions :

• wideband encoder from 0 to 7000 Hz in two subbands. A main strip was coded with the CELP technique, frame of 120 samples, excitation created by algebraic dictionaries, and transmission of certain energy and spectrum characteristics of a host band between 6000 Hz and 7000 Hz.
• " front " LPC analysis with 14 coefficients and " rear " LPC analysis with 50 coefficients every 120 samples. In LPC analysis mode " before ", separation into two LPC sub-blocks of 60 samples, the filter used for the first sub-block being interpolated from the current filter and the previous filter.

Calculation of the stationary parameter STAT (n)

In this embodiment, the above-mentioned stationarity parameter varies between the two extreme values 0 and 120, the aforementioned values STAT _m and STAT _M.

As regards the adaptation of the value of the stationary parameter STAT (n), the values of the functions f _a (N_BWD) and f _c (N_BWD) are such that:

With regard to the function f _r making it possible to establish the refinement value ΔS previously mentioned in the description, this is a step function of the variable x, with x = Gpb / Gpf and ΔS = f _r (x ) and having the value:

The refining of STAT (n) is also subject to the following conditions mentioned above in relation to FIG. 2c:
If ΔS> 0:
If Gpb> If
If STAT * (n) <STAT _i STAT (n) = STAT * (n) + ΔS
Otherwise STAT (n) = STAT * (n)
Otherwise STAT (n) = STAT * (n)
If not :
If STAT * (n) <STAT _i STAT (n) = STAT * (n) + ΔS
Otherwise STAT (n) = STAT * (n)
with STAT _i = 80, S _i = 0dB.

The other test conditions referenced 1121h and 1121d of figure 2c were not used in this mode of achievement.

Adaptation of decision thresholds

Le seuil S_LSP_H est adapté à l'aide de la fonction en escalier suivante :

La valeur du seuil S_TRANS utilisée en cas de transition des filtres LPC mesurée à l'aide du seuil S_LSP_H a été fixée à 0 dB.
La valeur du seuil S_STAT utilisée en cas de stationnarité des filtres LPC mesurée à l'aide du seuil S_LSP_L a été fixée à 2,5 dB.
La valeur du seuil G₁ a été fixée à 0dB.
En ce qui concerne la valeur d'énergie caractérisant une trame de silence ENER_SIL, cette valeur a été fixée à 50 dB mesurée sur les 120 échantillons s(i) du bloc courant B_n :

Regarding decision thresholds:
S_PRED is adapted as follows:
S_PRED (n) = 0.03.STAT (n) - 0.5 bounded in the interval
[S_PRED _m , S_PRED _M ],
with S_PRED _m = 0.5 and S_PRED _M = 2.5.
The threshold S_LSP_L is adapted using the following staircase function:

The threshold S_LSP_H is adapted using the following staircase function:

The value of the threshold S_TRANS used in the event of transition of the LPC filters measured using the threshold S_LSP_H was fixed at 0 dB.
The value of the threshold S_STAT used in case of stationarity of the LPC filters measured using the threshold S_LSP_L was fixed at 2.5 dB.
The value of the threshold G ₁ has been set at 0dB.
Regarding the energy value characterizing an ENER_SIL silence frame, this value was fixed at 50 dB measured on the 120 samples s (i) of the current block B _n :

With regard to the value of the threshold S _FWD mentioned above and intended to further limit the risk of switching over by imposing the LPC filtering mode " before " when the value STAT (n) is less than this threshold, this value S _FWD has was set at 60.

Claims (14)

1. A method for encoding a digital audio signal by dual analysis according to a choice criterion of LPC "forward' and "backward' analysis respectively into a transmitted encoded signal consisting of LPC filtering parameters accompanied by analysis decision information, and into a residue encoding signal, not transmitted, said digital audio signal being subdivided into frames, a succession of blocks of a specified number of samples, the encoding of said digital audio signal being carried out on this signal through a "forward' LPC filtering for non-stationary zones respectively on a synthesis signal, obtained from said residue encoding signal, through a "backward" LPC filtering for stationary zones, said choice criterion consisting, on each current block of said succession of current blocks constituting a current frame:

   in defining the degree of stationarity (STAT(n)) of the digital audio signal according to a stationarity parameter, the value of which lies between a maximum stationarity value and a minimum stationarity value;

   in establishing an analysis choice value (d_n(n)), by applying a decision function (D_n) adaptive to said stationarity parameter;

   in applying said analysis choice value to the "forward' LPC filtering so as to carry out the encoding of the said digital audio signal by "forward" LPC filtering for non-stationary zones on said digital audio signal, and by "backward' LPC filtering respectively for stationary zones on said synthesis signal, which makes it possible to favor the maintenance of the digital audio signal in one of the "forward" and "backward" filtering modes respectively in relation to the degree of stationarity and to limit the amount of switching from one to the other and vice versa of the filtering modes.
2. A method according to claim 1, characterized in that said decision function is an adaptive function, actualized for each current block from the stationarity parameter, said actualization of said adaptive function making it possible to favor the maintenance of the digital audio signal filtering in one of the "forward" and "backward" filtering modes respectively as a function of the degree of stationarity of said digital audio signal and thus to limit the amount of switching from one to the other and vice versa of the filtering modes.
3. A method according to one of claims 1 or 2, characterized in that said analysis choice value established from said decision function corresponds to a "forward" LPC filtering mode priority value and a "backward' LPC filtering mode priority value respectively.
4. A method according to one of claims 1 to 3, characterized in that the stage consisting in specifying the degree of stationarity of each current block of said digital audio signal consists, starting from an arbitrary starting value of said stationarity parameter:

   in calculating for the said current block an intermediate stationarity parameter value, as a function of a specified number of analysis choice values, obtained for different successive blocks prior to the said current block of the said succession of blocks, and of the stationarity parameter value of the block preceding the said current block;

   in tuning said intermediate stationarity parameter value as a function of the value of prediction gains of the "forward" and "backward' LPC filterings of the frame preceding said current frame.
5. A method according to claim 4, characterized in that the stage consisting, for each current block, in calculating an intermediate stationarity parameter value consists:

   in discriminating between the "forward' LPC or "backward' LPC analysis mode of the block preceding the said current block; and

   for any previous block analyzed by "backward' LPC analysis mode:

   in specifying the number of previous frames consecutively analyzed in "backward" LPC analysis mode,

   in comparing, according to a superiority comparison criterion, said number of previous frames with a first arbitrary value representative of a number of successive frames analyzed in "backward" LPC mode, and on positive response to this superiority comparison,

   attributing to said intermediate stationarity parameter value the stationarity parameter value of the block preceding the said current block, augmented by a specified value function of said first arbitrary value, and on negative response to this superiority comparison,

   attributing to said intermediate stationarity parameter value the stationarity parameter value of the block preceding the current block; and

   for any previous block analyzed in "forward' LPC analysis mode,

   in specifying according to a test criterion the occurrence of a transition from "backward" LPC analysis mode to "forward' LPC analysis mode between the block prior to said preceding block and said preceding block, and on positive response to the said test of occurrence,

   in comparing, according to an inferiority comparison criterion, said number of previous frames with a second arbitrary value representative of a number of successive frames analyzed in "backward" LPC mode preceding said preceding block, and on positive response to said inferiority comparison,

   attributing to said intermediate stationarity parameter value the stationarity parameter value of said block preceding the current block, reduced by a specified value which is a function of said second arbitrary value, and on negative response to said inferiority comparison,

   attributing to said intermediate stationarity parameter value the stationarity parameter value of said preceding block.
6. A method according to claim 4 or 5, characterized in that the stage consisting for each current block in tuning said intermediate stationarity parameter value consists:

   in distinguishing between prediction gains of the "forward' LPC filtering and "backward' LPC filtering;

   in modifying the intermediate stationarity parameter value of a refining value function of the relative value of prediction gains of "forward' and "backward' LPC filtering, the modification, increase or reduction, of the intermediate stationarity parameter value being proportional to said refining value.
7. A method according to claim 6, characterized in that the stage of increase proportional to said refining value of the intermediate stationarity parameter value is moreover subject to a condition of superiority of said value of "backward' LPC filtering gain relative to a first specified positive value and to a condition of inferiority of the value of said intermediate stationarity parameter value relative to a second specified positive value.
8. A method according to claim 6 or 7, characterized in that the stage of reduction proportional to said refining value of the intermediate stationarity parameter value is moreover subject to a condition of inferiority of said value of "backward' LPC filtering gain relative to a third specified positive value and to a condition of superiority of the value of said intermediate stationarity parameter value relative to a fourth specified positive value.
9. A method according to one of claims 6 to 8, characterized in that said relative value of the prediction gains of "forward' and "backward" LPC filtering consists in the ratio or the difference between prediction gains of "forward" and "backward" LPC filtering.
10. A method according to one of claims 1 to 3, characterized in that it consists in addition, for each successive current block:

    in establishing the average energy of said digital audio signal,

    in comparing, according to an inferiority comparison criterion, said average energy with a specified threshold value representative of a silence frame, and on positive response to said inferiority comparison,

    in attributing to said stationarity parameter of the current block the stationarity parameter value of the preceding block.
11. A method according to one of claims 2 to 10, characterized in that, for a degree of stationarity represented by a stationarity parameter between a minimum value and a maximum value, said minimum value representing the degree of stationarity of a substantially non-stationary digital signal and said maximum value representing the degree of stationarity of a substantially stationary signal, said adaptive function constituting the decision function is an increasing function of the priority value of the "backward' LPC filtering mode according to the increasing degree of stationarity of said digital signal.
12. An encoding device for a digital audio signal by dual analysis according to a choice criterion of "forward' and "backward" LPC analysis respectively into a transmitted encoded signal, said digital signal being subdivided into frames constituted by successive blocks comprising a specified number of samples, said encoding device comprising a "forward" LPC analysis filter and a "backward' LPC filter enabling delivery of a transmitted encoded signal consisting of LPC filtering parameters accompanied by an analysis decision indication and a means of encoding an encoding residue signal, not transmitted, enabling generation of a synthesis residue signal, the encoding of said digital audio signal being carried out on this digital audio signal from the "forward' LPC filter for non-stationary zones and on this synthesis signal, from the "backward" LPC filter respectively for stationary zones, said encoding device comprising in addition, for each current LPC block;

    calculation means of the degree of stationarity (STAT(n)) of said digital audio signal, according to a stationarity parameter the value of which is between a maximum stationarity value and a minimum stationarity value;

    setting means, from a stationarity parameter, of an adaptive decision function (D_n) enabling an LPC analysis choice value to be set (d_n(n));

    discrimination means of LPC analysis receiving said analysis choice value and enabling delivery, for said LPC current block, of the value of the "backward" and "forward" LPC filtering parameters respectively as a function of said analysis choice value;

    adaptive filtering means as a function of the degree of stationarity receiving said digital audio signal and the value of the "forward" and "backward' LPC filtering parameters respectively as a function of said analysis choice value and delivering the encoding residue signal to said encoding means of the encoding residue signal, which makes it possible to encode said digital audio signal and to favor the maintenance of said digital audio signal in one of the "forward" and "backward" filtering modes respectively in relation to the degree of stationarity of said digital signal and to limit the amount of switching from one to the other and vice versa of the filtering modes.
13. An encoding device according to claim 12, characterized in that said transmitted encoded signal consists, for each LPC analysis block, of:

    said analysis value,
       and in the case where the analysis choice value corresponds for LPC analysis block considered, to a "forward" LPC analysis;

    said "forward' LPC filtering parameters.
14. A decoding device of a digital audio signal encoded by dual analysis according to a choice criterion of "forward" and "backward' LPC analysis respectively, by an encoding device according to one of the claims 12 or 13, into a transmitted encoded signal consisting of LPC filtering parameters accompanied by an analysis decision value, said transmitted encoded signal consisting for each LPC analysis block of said analysis choice value and corresponding for an LPC analysis block considered to a "forward' LPC analysis in "forward' LPC filtering parameters, and for an LPC analysis block considered to a "backward' LPC analysis in "backward' LPC filtering parameters,
       said decoding device comprising at least:

    synthesis means for the filtering residue signal receiving the said encoding parameters of the LPC residue and delivering a synthesis residue signal,

    reverse filtering adaptive means as a function of the degree of stationarity, receiving the synthesis residue signal and enabling generation of a synthesis signal representative of the digital audio signal and constituting the decoded signal,

    "backward' LPC analysis means receiving said synthesis signal and enabling generation of "backward' LPC filtering parameters,

    discriminating means between "forward' LPC analysis and "backward" LPC analysis respectively receiving, on the one hand, for discrimination control said analysis choice value and, on the other hand, the "forward" LPC filtering parameters and the "backward' LPC filtering parameters and enabling delivery as a function of said analysis choice value, of either "forward" LPC filtering parameters, or "backward" LPC filtering parameters to said reverse filtering adaptive means as a function of the degree of stationarity.

Images