EP3175443B1 - Determining a budget for lpd/fd transition frame encoding - Google Patents

Description

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

The invention is advantageously applied to the coding / decoding of sounds that can contain speech and music mixed or alternately.

To effectively code low-speed speech sounds, CELP (Code Excited Linear Prediction) techniques are recommended. To effectively encode musical sounds, transform coding techniques are preferred.

CELP coders are predictive coders. They aim to model the production of speech from various elements: a short-term linear prediction to model the vocal tract, a long-term prediction to model the vibration of vocal cords in voiced period, and an excitation derived from a fixed dictionary (white noise, algebraic excitation) to represent the "innovation" which could not be modeled.

Transform coders such as MPEG AAC, AAC-LD, AAC-ELD or ITU-T G.722.1 Annex C, for example, use critical-sampling transforms to compact the signal in the transformed domain. A "critical-sampling transform" is a transform for which the number of coefficients in the transformed domain is equal to the number of time samples in each frame analyzed.

One solution for efficiently coding a mixed speech / music content signal consists in selecting, over time, the best technique between at least two coding modes, one of the CELP type, the other of the transformed type.

This is the case, for example, of the 3GPP AMR-WB + and MPEG USAC codecs (for "Unified Speech Audio Coding" ). The applications targeted by AMR-WB + and USAC are not conversational, but correspond to broadcasting and storage services, without strong constraints on the algorithmic delay.

• Pour les signaux de type parole: modes LPD (pour "Linear Predictive Domain" en anglais) comprenant deux modes différents dérivés du codage AMR-WB+:
  • Un mode ACELP
  • Un mode TCX (Transform Coded eXcitation) appelé wLPT (pour "weighted Linear Predictive Transform" en anglais) utilisant une transformée de type MDCT (contrairement au codec AMR-WB+ qui utilise une transformée de Fourier rapide, FFT, pour « Fast Fourier Transform » en anglais).
• Pour les signaux de type musique: mode FD (pour "Frequency Domain" en anglais) utilisant un codage par transformée MDCT (pour "Modified Discrete Cosine Transform" en anglais) de type MPEG AAC (pour "Advanced Audio Coding" en anglais) sur 1024 échantillons.

The initial version of the USAC codec, called Reference Model 0 (RMO), is described in the M. Neuendorf et al., A Novel Scheme for Low Bitrate Unified Speech and Audio Coding - MPEG RM0, 7-10 May 2009, 126th AES Convention . This RMO codec alternates between several encoding modes:

• For speech type signals: LPD (for Linear Predictive Domain ) modes comprising two different modes derived from the AMR-WB + coding:
  • ACELP mode
  • A TCX (Transform Coded eXcitation) mode called wLPT (for "weighted Linear Predictive Transform" in English) using an MDCT type transform (unlike the AMR-WB + codec which uses a Fast Fourier Transform, FFT) in English).
• For music type signals: FD mode (for " Frequency Domain" in English) using a transform coding MDCT (for "Modified Discrete Cosine Transform " in English) MPEG AAC type (for "Advanced Audio Coding" in English) on 1024 samples.

In the USAC codec, the transitions between LPD and FD modes are crucial to ensure sufficient quality without switching faults, knowing that each mode (ACELP, TCX, FD) has a specific "signature" (in terms of artifacts) and that the FD and LPD modes are different in nature - the FD mode is based on transform coding in the signal domain, while the LPD modes use predictive linear coding in the perceptually weighted domain with filter memories to be handled correctly. The management of intermode switches in the USAC RMO codec is detailed in the article of J. Lecomte et al., "Efficient cross-fade windows for transitions between LPC-based and non-LPC based audio coding", 7-10 May 2009, 126th AES Convention . As explained in this article, the main difficulty lies in the transitions between LPD to FD modes and vice versa. We retain here only the case of transitions from ACELP to FD.

In order to understand how it works, the principle of transform coding MDCT is described here through a typical embodiment.

• Pondération du signal par une fenêtre appelé ici "fenêtre MDCT" de longueur 2M
• Repliement temporel (ou "time-domain aliasing" en anglais) pour former un bloc de longueur M
• Transformation DCT (pour "Discrete Cosine Transform" en anglais) de longueur M

At the coder, the MDCT transformation is typically divided into three steps, the signal being cut into frames of M samples before MDCT coding:

• Weighting of the signal by a window called here "MDCT window" of length 2M
• Time folding (or "time-domain aliasing" ) to form a block of length M
• Transformation DCT (for "Discrete Cosine Transform " in English) of length M

The MDCT window is divided into 4 adjacent portions of equal lengths M / 2, here called "quarters".

The signal is multiplied by the analysis window and folds are made: the first quarter (windowed) is folded (ie inverted in time and overlapped) on the second quarter and the fourth quarter is folded on the third.

More precisely, the temporal folding from one quarter to another is done in the following way: the first sample of the first quarter is added (or subtracted) to the last sample of the second quarter, the second sample of the first quarter is added (or subtracted ) the second-last sample of the second quarter, and so on until the last sample of the first quarter that is added (or subtracted) to the first sample of the second quarter.

We thus obtain, from 4 quarters, 2 folded quarters where each sample is the result of a linear combination of 2 samples of the signal to be coded. This linear combination induces a temporal folding.

The folded two quarters are then coded together after DCT (Type IV) transformation. For the next frame we shift one half of window (or 50% overlap), the third and fourth quarters of the previous frame then become the first and second quarter of the current frame. After folding, we send a second linear combination of the same pairs of samples as in the previous frame, but with different weights.

At the decoder, after inverse DCT transformation, the decoded version of these folded signals is thus obtained. Two consecutive frames contain the result of 2 different folds of the same quarters, ie for each pair of samples we have the result of 2 linear combinations with different but known weights: an equation system can be solved for to obtain the decoded version of the input signal, the temporal folding can thus be suppressed by using two consecutive decoded frames.

The resolution of the systems of equations mentioned is generally made implicitly by unfolding, multiplication by a wisely selected synthesis window then addition-recovery of the common parts. This overlap addition ensures at the same time the smooth transition (without discontinuity due to quantization errors) between two consecutive decoded frames, in fact this operation behaves like a crossfade. When the window for the first quarter or fourth quarter is zero for each sample, we are talking about an MDCT transformation without time folding in this part of the window. In this case the smooth transition is not ensured by the MDCT transformation, it must be done by other means such as for example an external crossfade.

It should be noted that implementation variants of the MDCT transformation exist, in particular on the definition of the DCT transform, on how to fold the block to be transformed temporarily (for example, the signs applied to the folded quarters can be reversed). left and right, or fold the second and third quarter over the first and fourth quarters respectively), etc. These variants do not change the principle of the MDCT analysis-synthesis with the reduction of the sample block by windowing, temporal folding then transformation and finally windowing, folding and addition-recovery.

• Une fenêtre de transition pour le mode FD est utilisée avec un recouvrement à gauche de 128 échantillons.

In the case of the USAC RMO coder described in the article by Lecomte et al., The transition between a frame coded by an ACELP coding and a frame coded by an FD coding is carried out as follows:

• A transition window for the FD mode is used with a left overlap of 128 samples.

The temporal folding on this overlapping zone is canceled by introducing an "artificial" time folding to the right of the reconstructed ACELP frame. The MDCT window serving at the transition has a size of 2304 samples and the DCT transformation operates on 1152 samples whereas normally FD mode frames are encoded with a window size of 2048 samples and a DCT transformation of 1024 samples. Thus the MDCT transformation of the normal FD mode is not directly usable for the transition window, the encoder must also integrate a modified version of this transformation which complicates the implementation of the transition for the FD mode.

This coding technique of the state of the art has an algorithmic delay of the order of 100 to 200 ms. This delay is incompatible with conversational applications for which coding delay is generally of the order of 20 to 25 ms for speech coders for mobile applications (eg GSM EFR, 3GPP AMR and AMR-WB) and in the order of 40 ms for conversational transform coders for videoconferencing (eg ITU-T G.722.1 Annex C and G.719). In addition, the fact of punctually increasing the DCT transformation size (2304 against 2048) generates a peak of complexity at the time of the transition.

To overcome these drawbacks, the international patent application WO2012 / 085451 , the contents of which are incorporated by reference in the present application, proposes a new method for encoding a transition frame. The transition frame is defined as the transform-encoded current frame that succeeds a previous frame encoded by predictive coding. According to the new method mentioned above, a part of the transition frame, for example a subframe of 5 ms, in the case of a CELP coding at 12.8 kHz, and two additional CELP frames of 4 ms each, in the case of a CELP coding at 16 kHz, are coded by a predictive coding restricted with respect to the predictive coding of the previous frame.

Restricted predictive coding consists in using the stable parameters of the preceding frame coded by a predictive coding, such as the coefficients of the linear prediction filter and coding only a few minimum parameters for the additional subframe in the transition frame.

Since the previous frame has not been coded with transform coding, the cancellation of the temporal folding in the first part of the frame is impossible. The patent application WO2012 / 085451 Moreover, it is proposed to modify the first half of the MDCT window so as not to have time folding in the first quarter normally folded. It is also proposed to integrate a part of the addition-overlap (also called "fade-in" or " overlap-add " in English) between the decoded CELP frame and the decoded MDCT frame by modifying the coefficients of the window. analysis / synthesis. With reference to the figure 4e of the aforementioned application, the mixed lines (lines alternating points and lines) correspond to the MDCT coding folding lines (top figure) and the MDCT decoding unfolding lines (bottom figure). In the figure above, the bold lines separate the frames of new samples at the input of the encoder. The coding of a new MDCT frame can be started when a so-defined frame of new input samples is fully available. It is It is important to note that these lines in bold at the coder do not correspond to the current frame but to the successive blocks of new samples arriving for each frame: the current frame is in fact delayed by 8.75 ms which corresponds to an anticipation, called " lookahead " . " in English. In the bottom figure, the bold lines separate the decoded frames at the output of the decoder.

At the encoder, the transition window is zero to the point of folding. Thus the coefficients of the left part of the folded window will be identical to those of the unfolded window. The portion between the folding point and the end of the CELP transition subframe (TR) corresponds to a sinusoidal half-window. At the decoder, after unfolding, the same window is applied to the signal. On the segment between the folding point and the beginning of the MDCT frame, the coefficients of the window correspond to a window of form sin ² . To ensure the overlap-addition between the decoded subframe CELP and the signal from the MDCT, it suffices to apply a window of cos ² type to the part of the subframe CELP in overlap and to summon the latter with the MDCT frame. The method is perfect reconstruction.

However, the demand WO2012 / 085451 provides for the allocation of a _trans B bit budget for the coding of the CELP subframe which would correspond to the budget required for the CELP coding of a conventional frame, reduced to a single sub-frame. The remaining budget for the coding of the transform transition frame is then insufficient and can lead to a drop in quality at low bit rate.

The present invention improves this situation.

• affectation d'un débit binaire pour le codage prédictif de la sous-trame de transition, le débit binaire étant égal au minimum entre le débit binaire du codage par transformée de la trame de transition et une première valeur prédéterminée de débit binaire ;
• détermination d'un premier nombre de bits alloués pour le codage prédictif de la sous-trame de transition pour le débit binaire ; et
• calcul d'un deuxième nombre de bits alloués pour le codage par transformée de la trame de transition à partir du premier nombre de bits et d'un nombre de bits disponibles pour le codage de le trame de transition.

For this purpose, a first aspect of the invention relates to a method for determining a coding bit distribution of a transition frame. This method is implemented in an encoder / decoder for encoding / decoding a digital audio signal. The transition frame is preceded by a preceding prediction coded frame and the coding of this transition frame comprises transform coding and predictive coding of a single subframe of the transition frame. The method comprises the following steps:

• assigning a bit rate for the predictive encoding of the transition subframe, the bit rate being at least between the bit rate of the transform coding of the transition frame and a first predetermined bit rate;
• determining a first number of bits allocated for the predictive coding of the transition subframe for the bit rate; and
• calculating a second number of bits allocated for the transform coding of the transition frame from the first number of bits and a number of bits available for the encoding of the transition frame.

The bit rate of the predictive coding is therefore limited by a maximum value. The number of bits allocated to the predictive coding depends on this bit rate. Since the lower the bit rate, the smaller the number of bits allocated to the coding, a minimum remaining budget is ensured for the coding of the transform transition frame.

In addition, the number of bits allocated to the predictive coding of the subframe is optimized with respect to the bit rate of the transform coding. Indeed, if the bit rate of the transform coding of the transition frame is less than the first predetermined value, the bit rate for the predictive coding and the bit rate for the transform coding is identical. The coherence of the signals thus generated is thus improved, which simplifies the subsequent steps of coding (channel coding) and the processing of the frames received at the decoder.

In another embodiment, the encoder / decoder comprises a first working core, for predictive coding / decoding of a signal frame, at a first frequency, and a second working heart, for predictive coding / decoding of a signal frame, at a second frequency. The first predetermined bit rate value depends on the core selected from the first and second cores for coding / decoding of the prediction encoded previous frame.

The operating frequency of the encoder / decoder core influences the number of bits required to correctly represent the input digital signal. For example, for certain operating frequencies, it is necessary to provide additional bits for coding frequency bands not directly processed by the core.

In one embodiment, when the first core has been selected for encoding / decoding of the prediction encoded previous frame, the assigned bit rate is furthermore equal to the maximum between the bit rate of the transform coded transition frame and the minus a second predetermined value of bit rate, the second value being less than the first value. Thus, a minimum bit rate is guaranteed in order to avoid excessively large bit rates between the different coded frames.

In another embodiment, the digital signal is decomposed into at least one low frequency band and one high frequency band. In this situation, the first calculated bit count is assigned to the predictive encoding of the transition subframe for the low frequency band. A third predetermined number of bits is then allocated to an encoding of the transition subframe for the high frequency band. In addition, the second number of bits allocated for the transform coding of the transition frame is then further determined from the third predetermined number of bits. Thus, it is possible to efficiently code the entire frequency spectrum of the input signal without sacrificing the quality of the signal reproduced at decoding.

In one embodiment, the number of bits available for encoding the transition frame is fixed. This reduces the complexity of the coding steps.

In another embodiment, the second number of bits is equal to the fixed number of coding bits of the transition frame minus the first number of bits minus the third number of bits. The final determination of the distribution of the bits in the transition frame is thus limited to a subtraction of integer values which simplifies the coding.

In a variant, the second number of bits is equal to the fixed number of coding bits of the transition frame minus the first number of bits minus the third number of bits minus a first bit minus a second bit. The first bit indicates whether low-pass filtering is performed when determining the parameters of the predictive coding of the transition subframe, the parameters being related to the pitch delay. The second bit indicates the frequency used by the encoder / decoder heart for the predictive coding / decoding of the transition subframe. Such an indication allows a more flexible coding.

• * codage d'une trame précédente d'échantillons du signal numérique selon un codage prédictif ;
• * codage d'une trame courante d'échantillons du signal numérique en une trame de transition, le codage de la trame de transition comprenant un codage par transformée et un codage prédictif d'une unique sous-trame de la trame de transition, le codage de la trame courante comportant les sous-étapes suivantes :
  • détermination de la répartition des bits par le procédé selon le premier aspect de l'invention ;
  • codage par transformée de la trame de transition sur le deuxième nombre de bits alloués ;
  • codage prédictif de la sous-trame de transition sur le premier nombre de bits alloués.

A second aspect of the invention relates to a method for coding an audio digital signal in an encoder able to code signal frames according to a predictive coding or according to a transform coding, comprising the following steps:

• coding of a previous frame of samples of the digital signal according to a predictive coding;
• coding of a current frame of samples of the digital signal into a transition frame, the coding of the transition frame comprising a transform coding and a predictive coding of a single subframe of the transition frame, the coding the current frame including the following substeps:
  • determining the bit distribution by the method according to the first aspect of the invention;
  • transform coding of the transition frame on the second number of allocated bits;
  • predictive encoding of the transition subframe on the first number of allocated bits.

The determination of the distribution of the bits included in the transition frame is thus determined before the coding. As described below, this determination of the distribution of the bits is reproducible to the decoder which avoids an explicit transmission of information on this distribution.

In addition, this coding ensures a balanced distribution between the predictive coding and the transform coding within the transition frame.

In one embodiment, the predictive coding comprises generating predictive coding parameters determined for the assigned bit rate when distributing the bits in the transition frame. The use of such predictive parameters optimizes the ratio between the bit rate assigned to the predictive coding and the remaining bit rate assigned to the transform coding, and thus optimizes the quality of the reconstructed signal. Indeed, at constant quality, the number of bits allocated to one or another predictive parameter can vary in non-linear proportions with respect to the bit rate assigned to the predictive coding.

In another embodiment, the predictive coding comprises generating predictive coding parameters restricted with respect to the predictive coding of the preceding frame by reusing at least one parameter of the predictive coding of the preceding frame. So, at decoding, additional information is extracted from the previous frame to complete the decoding of the transition subframe to be decoded. This reduces the number of bits that must be reserved for the predictive encoding of the transition subframe.

The combination of reuse of parameters of a previous frame with the allocation of the bit rate of the transform coding of the transition frame makes it possible to ensure a coherent transition at a lower cost.

• * décodage prédictif d'une trame précédente d'échantillons du signal numérique codée selon un codage prédictif ;
• * décodage d'une trame de transition codant une trame courante d'échantillons du signal numérique, le codage de la trame de transition comprenant un codage par transformée et un codage prédictif d'une unique sous-trame de la trame de transition, comportant les sous-étapes de :
  • détermination de la répartition des bits par le procédé selon le premier aspect de l'invention ;
  • décodage prédictif de la sous-trame de transition sur le premier nombre de bits alloués ;
  • décodage par transformée de la trame de transition sur le deuxième nombre de bits alloués.

A third aspect of the invention relates to a method for decoding an audio digital signal coded by predictive coding and by transform coding, comprising the steps of:

• predictive decoding of a previous frame of samples of the digital signal encoded according to a predictive coding;
• decoding of a transition frame coding a current frame of samples of the digital signal, the coding of the transition frame comprising a transform coding and a predictive coding of a single subframe of the transition frame, comprising the sub-steps of:
  • determining the bit distribution by the method according to the first aspect of the invention;
  • predictive decoding of the transition subframe on the first number of allocated bits;
  • decoding by transforming the transition frame on the second number of allocated bits.

As mentioned above, the method for determining the distribution of bits in the transition frame is directly reproducible at the decoder. Indeed, the bit distribution is determined solely from the bit rate of the portion of the transform coded transition frame. No additional bit is therefore necessary to implement this step of determining the distribution of the bits and a saving of bandwidth is therefore achieved.

A fourth aspect of the invention further provides a computer program comprising instructions for carrying out the method according to the aspects of the invention described above, when these instructions are executed by a processor.

• affectation d'un débit binaire pour le codage prédictif de la sous-trame de transition, ledit débit binaire étant égal au minimum entre le débit binaire du codage par transformée de la trame de transition et une première valeur prédéterminée de débit binaire, au ;
• détermination d'un premier nombre de bits alloués pour le codage prédictif de la sous-trame de transition pour le débit binaire;
• calcul d'un deuxième nombre de bits alloués pour le codage par transformée de la trame de transition à partir du nombre de bits requis pour le codage des paramètres de codage et le nombre fixe de bits de codage de la trame de transition.

A fifth aspect of the invention relates to a device for determining a coding bit distribution of a transition frame, this device being implemented in an encoder / decoder for coding / decoding of a digital audio signal, the transition frame being preceded by a preceding prediction coded frame, the coding of the transition frame comprising a transform coding and a predictive coding of a single sub-frame of the transition frame, the number of coding bits of the transition frame being fixed, the device comprising a processor arranged to perform the following operations:

• assigning a bit rate for the predictive coding of the transition subframe, said bit rate being at least between the bit rate of the transform coding of the transition frame and a first predetermined bit rate value, at;
• determining a first number of bits allocated for the predictive coding of the transition subframe for the bit rate;
• calculating a second number of bits allocated for the transform coding of the transition frame from the number of bits required for coding the coding parameters and the fixed number of coding bits of the transition frame.

• * un dispositif selon le cinquième aspect de l'invention ;
• * un codeur prédictif comportant un processeur agencé pour effectuer les opérations suivantes :
  • codage d'une trame précédente d'échantillons du signal numérique selon un codage prédictif ;
  • codage prédictif d'une unique sous-trame comprise dans une trame de transition codant une trame courante d'échantillons du signal numérique, le codage de la trame de transition comprenant un codage par transformée et un codage prédictif de la sous-trame, le processeur étant agencé pour effectuer l'opération de codage prédictif de la sous-trame de transition sur le premier nombre de bits alloués ;
• * un codeur par transformée comportant un processeur agencé pour effectuer un codage par transformée de la trame de transition sur le deuxième nombre de bits alloués.

A sixth aspect of the invention is further directed to an encoder capable of encoding frames of a digital audio signal according to a predictive or transform coding, comprising:

• a device according to the fifth aspect of the invention;
• a predictive encoder comprising a processor arranged to perform the following operations:
  • encoding a previous frame of samples of the digital signal according to a predictive coding;
  • predictive coding of a single sub-frame included in a transition frame coding a current frame of samples of the digital signal, the coding of the transition frame comprising a transform coding and a predictive coding of the sub-frame, the processor being arranged to perform the predictive coding operation of the transition subframe on the first number of allocated bits;
• a transform coder comprising a processor arranged to perform transform coding of the transition frame on the second number of allocated bits.

• * un dispositif selon le cinquième aspect de l'invention ;
• * un décodeur prédictif comportant un processeur agencé pour effectuer les opérations suivantes :
  • décodage prédictif d'une trame précédente d'échantillons du signal numérique codée selon un codage prédictif ;
  • décodage prédictif d'une unique sous-trame comprise dans une trame de transition codant une trame courante d'échantillons du signal numérique, le codage de la trame de transition comprenant un codage par transformée et un codage prédictif de la sous-trame, le processeur étant agencé pour effectuer l'opération de décodage prédictif de la sous-trame de transition sur le premier nombre de bits alloués ;
• * un décodeur par transformée comportant un processeur agencé pour effectuer un décodage par transformée de la trame de transition sur le deuxième nombre de bits alloués.

A seventh aspect of the invention is further directed to a decoder of a digital signal encoded by predictive coding and transform coding, comprising:

• a device according to the fifth aspect of the invention;
• a predictive decoder comprising a processor arranged to perform the following operations:
  • predictive decoding of a previous frame of samples of the coded digital signal according to a predictive coding;
  • predictive decoding of a single sub-frame included in a transition frame coding a current frame of samples of the digital signal, the coding of the transition frame comprising a transform coding and a predictive coding of the sub-frame, the processor being arranged to perform the predictive decoding operation of the transition subframe on the first number of allocated bits;
• a transform decoder comprising a processor arranged to carry out a decoding by transforming the transition frame on the second number of allocated bits.

• la figure 1 illustre un codeur audio selon un mode de réalisation de l'invention ;
• la figure 2 est un diagramme illustrant les étapes d'un procédé de codage, mis en oeuvre par le codeur audio de la figure 1, selon un mode de réalisation de l'invention ;
• la figure 3 présente une transition entre trames CELP et MDCT selon un mode de réalisation de l'invention ;
• la figure 4 est un diagramme illustrant les étapes d'un procédé de détermination d'une répartition de bits de codage d'une trame de transition, selon un mode de réalisation de l'invention ;
• la figure 5 illustre un décodeur audio selon un mode de réalisation de l'invention ;
• la figure 6 est un diagramme illustrant les étapes d'un procédé de décodage, mis en oeuvre par le décodeur audio de la figure 5, selon un mode de réalisation de l'invention ;
• la figure 7 illustre un dispositif de détermination de la répartition de bits dans une trame de transition selon un mode de réalisation de l'invention.

Other features and advantages of the invention will appear on examining the detailed description below, and the attached drawings in which:

• the figure 1 illustrates an audio coder according to an embodiment of the invention;
• the figure 2 is a diagram illustrating the steps of a coding method, implemented by the audio coder of the figure 1 according to one embodiment of the invention;
• the figure 3 has a transition between CELP and MDCT frames according to one embodiment of the invention;
• the figure 4 is a diagram illustrating the steps of a method for determining a coding bit distribution of a transition frame, according to an embodiment of the invention;
• the figure 5 illustrates an audio decoder according to an embodiment of the invention;
• the figure 6 is a diagram illustrating the steps of a decoding method, implemented by the audio decoder of the figure 5 according to one embodiment of the invention;
• the figure 7 illustrates a device for determining the distribution of bits in a transition frame according to an embodiment of the invention.

The figure 1 illustrates an audio encoder 100 according to one embodiment of the invention.

The figure 2 is a diagram illustrating the steps of a coding method, implemented by the audio coder 100 of the figure 1 according to one embodiment of the invention.

The encoder 100 comprises a reception unit 101 for receiving, at a step 201, an input signal sampled at a given frequency fs (for example 8, 16, 32 or 48 kHz) and decomposed into subframes, for example 20 ms.

On receipt of a current frame, a preprocessing unit 102 is able to select, in a step 202, the coding mode which is the most suitable for the coding of the current frame, among at least one LPD mode and one FD mode. In the following description, it is considered, by way of illustration, that an MDCT encoding is used for the FD mode and that a CELP encoding is used for the LPD mode. There are no restrictions on the coding techniques used for the LPD and FD modes, respectively. Thus, modes additional to the CELP and MDCT modes may for example be used, the CELP coding may be replaced by another type of predictive coding, the MDCT transform may be replaced by another type of transform.

It is assumed here that the frame type is transmitted explicitly via block 206, with for example a fixed length coding indicating the mode selected from a predefined list. In variants of the invention, this coding of the mode chosen in each frame may be of variable length. It is also expected that the CELP coding type (12.8 or 16 kHz) can be transmitted explicitly via a bit to facilitate the decoding of the transition frame.

A step 203 verifies that the CELP encoding has been selected in step 202. In the case where the LPD mode is selected, the signal frame is transmitted to a CELP encoder 103 for encoding a CELP frame at a step 204 The CELP coder may use two "cores" operating at two respective internal sampling frequencies, for example set at 12.8 kHz and 16 kHz, which require the use of a sampling of the input signal (at the frequency fs). at the internal frequency of 12.8 or 16 kHz. Such resampling may be implemented in a resampling unit in the preprocessing block 102 or in the CELP coder 103. The frame is then coded by prediction by the CELP coder 103 by deducting CELP parameters which generally depend on a classification of the signal. These CELP parameters typically include LPC coefficients, an adaptive and fixed gain vector, an adaptive dictionary vector, a fixed dictionary vector. This list can also be modified according to a signal class in the frame, as in ITU-T G.718 coding. The parameters thus calculated can then be quantized, multiplexed and transmitted at a step 206 to the decoder by a transmission unit 108. CELP coding parameters, such as the LPC coefficients, the adaptive and fixed gain vector, the adaptive dictionary vector , the fixed dictionary vector as well as states of the CELP decoder may also be stored, in a step 205, in a memory 107 in the case where the frame following the current frame would be a transition frame MDCT.

As explained below, a band extension may also be performed with associated coding of the high band when the current frame is of CELP type.

In the case where the MDCT coding has been selected by the unit 102 in the step 203, it is verified in a step 207 that the frame preceding the current frame has been coded by the MDCT transform. In the case where the frame preceding the current frame has been coded by MDCT transform, the current frame is transmitted to the MDCT encoder 105 directly, for a transform coding MDCT of the current frame at a step 208. The MDCT encoder can encode a frame covering 28.75 ms of un-resampled signal, including 20 ms of frame and 8.75 ms of lookahead for example. No restrictions are attached to the size of the MDCT window. In addition, a delay corresponding to the delay of the CELP coder due to resampling of the input signal is applied to the frame coded by the MDCT coder, so that the MDCT and CELP frames are synchronized. Such an encoder delay may be 0.9375 ms depending on the type of resampling before CELP coding. The transform coded frame MDCT is transmitted to the decoder at step 206.

In the case where the MDCT coding is selected by the unit 102, and in the case where the frame preceding the current frame has been prediction coded, the current frame is a transition frame and is transmitted to a transition unit 104. As described in the following, the MDCT transition frame comprises an additional CELP subframe.

• anticiper, à une étape 209, le budget de bits nécessaires au codage de la sous-trame CELP de transition de manière à définir le budget disponible pour le codage MDCT de la trame courante. Comme détaillé dans ce qui suit, le budget peut dépendre du débit de la trame courante. En outre, le budget peut être évalué en fonction du coeur CELP utilisé. Afin de conserver un budget de bits suffisant pour ne pas dégrader la qualité du codage MDCT, l'invention peut prévoir de borner le débit de codage de la sous-trame CELP. Elle comprend à cet effet un dispositif de détermination de la répartition de bits dans une trame de transition, tel que le dispositif 700 de la figure 7 ;
• modifier, à une étape 210, la fenêtre MDCT utilisée au codeur conformément à la figure 3 décrite ci-après ;
• mettre à zéro la mémoire de la transformation MDCT puisque la trame précédente était une trame CELP, à une étape 211 - de façon équivalente la mémoire MDCT pourra être ignorée dans le décodage MDCT.

The transition unit 104 is able to implement the following steps:

• anticipating, in a step 209, the budget of bits necessary for the coding of the CELP subframe to define the available budget for the MDCT encoding of the current frame. As detailed in the following, the budget may depend on the bit rate of the current frame. In addition, the budget can be evaluated according to the CELP core used. In order to maintain a sufficient bit budget so as not to degrade the quality of the MDCT coding, the invention can provide for limiting the coding rate of the CELP subframe. It comprises for this purpose a device for determining the distribution of bits in a transition frame, such as the device 700 of the figure 7 ;
• modify, at a step 210, the window MDCT used to the encoder according to the figure 3 described below;
• reset the memory of the MDCT transformation since the previous frame was a CELP frame, at a step 211 - in an equivalent manner the MDCT memory can be ignored in the MDCT decoding.

In one embodiment, at least one of these steps is performed by the transition frame coding unit 106 described hereinafter.

The transition MDCT frame is encoded by the MDCT encoder 105, at a step 212, as described in the following, and based on the bit budget allocated in step 209. The additional CELP subframe is also encoded by the CELP encoder 103, at a step 213, as described in the following with reference to the figure 3 , and according to the bit budget allocated in step 209. The CELP coding may be performed before or after MDCT coding.

The figure 3 presents the transition between CELP and MDCT frames to the coder, before coding, and to the decoder, before decoding.

A frame to be coded 301 is received at the encoder 100 and is encoded by the CELP encoder 103. A current frame 302 is then received at the input of the encoder 100 to be encoded by the MDCT transform. It is therefore a transition frame. The next frame 303 received at the input of the coder is also coded by MDCT transform. According to the invention, the next frame 303 could be coded by CELP coding and no restriction is attached to the coding used for the next frame 303.

An asymmetric MDCT window 304 may be used for encoding the current frame. This window 304 has a rising edge 307 of 14.375 ms, a bearing with a gain of 1 at 11.25 ms, a falling edge 309 of 8.75 ms corresponding to the lookahead, and a zero portion 310 of 5.265 ms. The addition of the zero part 310 reduces the lookahead and thus the corresponding delay. In one embodiment, the shape of this MDCT analysis window for the MDCT encoding is modified, for example, to further reduce the lookahead or to use a symmetrical window whose examples are given in the patent application. WO2012 / 085451 .

The dashed line 312 represents the middle of the MDCT window 304. On either side of the line 312, the 10 ms quarters of the MDCT window 212 are folded as described in FIG. introductory part. Continuous line 311 indicates the folding area between the first and second quarters of the MDCT window 304.

The MDCT window of the next frame 303 is referenced 306 and has a covering addition area with the MDCT window 304 corresponding to the falling edge 309 of the MDCT window 304.

An MDCT window 305 theoretically represents the window that would be applied to the previous frame if it had been coded by MDCT transform. However, since the preceding frame 301 is coded by the CELP coder 103, it is necessary, in order to allow the decoder to unfold the first part of the MDCT transformed coded frame, that the window be zero in the first quarter (since the second part of the previous MDCT frame is not available).

For this purpose, the MDCT window 304 is modified into an MDCT window 313 whose first quarter is zero, thus allowing time folding in the first part of the MDCT frame to the decoder.

At the level of the decoder, the analysis windows 304, 305, 306 and 313 respectively correspond to synthesis windows 324, 325, 326 and 327. This synthesis window is therefore inverted temporally with respect to the corresponding analysis window. In variants of the invention, the windows of analysis and synthesis may be identical, sinusoidal type or otherwise.

A first frame 320 of new samples encoded by CELP coding is received at the decoder. It corresponds to the coded version of the CELP frame 301. It is recalled here that the decoded frame is shifted by 8.75 ms with respect to the frame 320.

The coded version of the transition frame 302 is then received (references 321 and 322 forming a complete frame). Between the end of the CELP frame 320 and the beginning of the rising edge of the synthesis window 327 (corresponding to the fold line), a hole or GAP is created. In the particular example represented here, a quarter window MDCT being 10 ms and the zero part of the synthesis window MDCT 324 which covers the CELP frame 320 being 5.625 ms (corresponding to the part 310 of the analysis window MDCT 204), the hole is 4.375 ms. Further, in order to ensure a satisfactory overlay addition length with the beginning of the non-zero portion of the MDCT window 327, the delay between the CELP frame 320 and the start of the MDCT window 327 is extended by the required length. In the following example, we consider, by way of illustration, a satisfactory recovery addition length of 1.875 ms, the aforementioned delay (corresponding to a missing signal length) thus being increased to 6.25 ms, as represented by FIG. reference 321 on the figure 2 .

It will be noted that the signal frames represented in figure 3 may contain signals at different sampling rates of 12.8 or 16 kHz for CELP coding / decoding and fs for MDCT coding / decoding; however at the decoder, after resampling of the CELP synthesis and time shift of the MDCT synthesis the frames remain synchronized and the representation of the figure 3 remains accurate.

As mentioned above the request WO2012 / 085451 proposes to encode an additional CELP subframe of 5ms at the beginning of the MDCT transition frame, in the case of CELP coding at 12.8 kHz, and two additional CELP frames of 4ms each at the beginning of the MDCT transition frame, in the case of a CELP coding at 16 kHz.

In the case at 12.8 kHz, the 6.25 ms delay is not filled and the overlap is affected: the decoder is only 0.625 ms of overlap, which is insufficient.

In the 16 kHz case, two additional CELP subframes are encoded at the beginning of the transition frame, leaving very little budget for coding the transition MDCT frame and can lead to a significant lower quality drop. debit.

In order to overcome these drawbacks, the present invention can provide for encoding a single additional CELP subframe at 12.8 or 16 kHz by the CELP encoder 103. Additional samples are generated at the decoder, as detailed below, in order to generate the missing signal over the length of 6.25 ms above.

In order to encode the transition CELP subframe, the unit 106 may reuse at least one CELP parameter from the previous CELP frame. For example, the unit 106 can reuse the linear prediction coefficients A (z) of the preceding CELP subframe as well as the innovation energy of the preceding frame (stored in the memory 107 as previously described) so that to encode only the adaptive dictionary vector, the adaptive gain, the fixed gain and the fixed dictionary vector of the transition CELP subframe. Thus, the additional CELP subframe can be coded with the same core (12.8 kHz or 16 kHz) as the previous CELP frame.

A transition frame coding unit 106 provides the coding of a transition frame according to the invention. The invention may further provide for the insertion by the unit 106 into the bit stream of an additional bit indicating that the coded frame 322 is a transition frame, however in the general case this transition frame indication may also be transmitted in the global coding mode indication of the current frame, without taking any additional bits.

The invention may further provide that this unit 106 encodes the high band of the signal at steps 204 and 214 (so-called "band extension" method), when this is required, with a fixed budget since the frequency of sampling of the synthesis signal at the decoder is not necessarily identical to the frequency of the CELP core.

• filtrage de la trame précédente CELP et de la sous-trame CELP de la trame de transition par un filtre passe-haut afin de ne conserver que la partie haute du spectre (au-dessus de la fréquence qui correspond au coeur CELP utilisé, soit au dessus de 6,4 ou 8 kHz). Un tel filtrage peut être mis en oeuvre par un filtre à réponse impulsionnelle finie, FIR, du codeur CELP 103 ;
• recherche de corrélation entre la partie filtrée de la sous-trame CELP de transition originale et la trame CELP précédente filtrée afin d'estimer un paramètre de retard puis un gain (différence d'amplitude entre le signal correspondant à la sous-trame filtrée et le signal prédit par application du retard) ;
• codage du paramètre de retard et du gain précités en utilisant par exemple une quantification scalaire (par exemple, le retard peut être codé sur 6 bits et le gain sur 6 bits).

For this purpose, the transition frame coding unit 106 can implement the following steps:

• filtering of the previous CELP frame and the CELP subframe of the transition frame by a high-pass filter in order to keep only the upper part of the spectrum (above the frequency which corresponds to the CELP core used, ie above 6.4 or 8 kHz). Such a filtering can be implemented by a finite impulse response filter, FIR, of the CELP encoder 103;
• correlation search between the filtered part of the original transition CELP subframe and the filtered previous CELP frame in order to estimate a delay parameter and then a gain (difference in amplitude between the signal corresponding to the filtered subframe and the signal predicted by applying the delay);
• encoding the aforementioned delay and gain parameter using, for example, scalar quantization (for example, the delay may be 6-bit coded and the 6-bit gain).

Step 209 above is illustrated in more detail with reference to the figure 4 which is a diagram illustrating the steps of a method for determining a transition coding bit distribution according to an embodiment of the invention. The aforementioned method is implemented in the same manner to the encoder and the decoder, but is presented, for illustrative purposes only, on the coder side.

In a step 400, the total bit rate (in bits / s), denoted core_brate, which can be allocated to the coding of the current frame is set equal to the output rate of the MDCT encoder. The duration of the frame being considered in this example as being 20 ms, the number of frames per second is 50 and the total budget in bits is equal to core_brate / 50. The total budget can be fixed, in the case of a fixed rate encoder, or variable, in the case of a variable rate encoder when an adaptation of the coding rate is implemented. In the following, we use a variable num_bits, initialized to the value core_brate / 50.

At a step 401, the transition unit 104 determines the CELP core, among at least two CELP cores, which has been used for encoding the previous CELP frame. In the following example, two CELP cores operating at frequencies of 12.8 kHz and 16 kHz are considered. As a variant, a single CELP core is used for coding and / or decoding.

In the case where the CELP core used for the preceding CELP frame is at a frequency of 12.8 kHz, the method comprises a step 402 for assigning a bit rate, noted cbrate, for the CELP coding of the subframe transition, the bit rate being equal to the minimum between the bit rate of the MDCT coding of the transition frame and a first predetermined bit rate value. The first predetermined value can be set at 24.4 kbit / s, for example, which ensures a satisfactory bit budget for the transform coding.

So, cbrate = min (core_bitrate, 24400). This limitation amounts to restricting the operation of the restricted CELP coding limited to the supplementary subframe with CELP parameters coded as if they were encoded by a CELP coding of not more than 24.40 kbit / s.

• num_bits := num_bits - 1

At an optional step 403, the affected bit rate is compared to a CID rate of 11.60 kbit / s. If the assigned bit rate is higher, a bit may be reserved to encode an adaptive dictionary low pass filtering binary indication (as for example in the AMR-WB coding at rates greater than or equal to 12.65 kbit / s). The num_bits variable is updated:

• num_bits: = num_bits - 1

At a step 404, a first number of bits, noted budget1, is allocated for the predictive coding of the additional CELP subframe. The first number of bits budget1 represents the number of bits representing the CELP parameters used for the coding of the CELP subframe. As previously detailed, the coding of the CELP subframe can be restricted in that a small number of CELP parameters are used, some parameters used for the coding of the preceding CELP frame being reused advantageously.

For example, only the excitation can be modeled for the coding of the additional CELP subframe, and bits are thus reserved only for the fixed dictionary vector, for the adaptive dictionary vector and for the gain vector. The number of bits allocated to each of these parameters is deduced from the bit rate assigned to the coding of the additional CELP sub-frame in step 402. For example, Table 1 / G722.2 - Bit allocation of the algorithm of AMR-WB coding for 20ms frame, from the July 2003 release of the ITU-T G.722.2 standard, gives examples of bit allocations per CELP parameter as a function of the affected bit rate.

In the preceding example, where the coding of the sub-frame is restricted, budget1 corresponds to the sum of the bits allocated respectively to the adaptive dictionary, the fixed dictionary and the vector of gains. For example, for an assigned bit rate of 19.85 kbit / s, referring to the aforementioned Table / G722, 9 bits are allocated to the adaptive dictionary (tonal delay), 72 bits are allocated to the fixed dictionary (algebraic code), and 7 bits are allocated to the gain vector (directory gain). In this case, budget1 is equal to 88 bits.

• num_bits := num_bits - budg1

The num_bits variable can thus be updated:

• num_bits: = num_bits - budget1

The invention can also provide for taking into account frame classes in the bit allocation to the CELP parameters. For example, the ITU-T G.718 standard, in its June 2008 version, sections 6.8 and 8.1, gives the budgets to allocate to each CELP parameter according to classes, or modes such as unvoiced mode. (UC), the voiced mode (VC), the transition mode (TC) and the generic mode (GC), and according to the allocated bit rate (layerl or layer2, corresponding to rates of 8kbit / s and 8 respectively). +4 kbit / s). The G.718 encoder is a hierarchical coder, but it is possible to combine the CELP coding principles using a G.718 classification with the AMR-WB multi-bit allocation.

If it has been determined in step 401 that the CELP core used for the previous CELP frame is at a frequency of 16 kHz, the method comprises a step 405 of assigning a bit rate, noted cbrate, for coding CELP of the transition subframe, the bit rate being equal to the minimum between the bit rate of the MDCT coding of the transition frame and a first predetermined value of bit rate. In the case of the 16 kHz core, the first predetermined value can be fixed at 22.6 kbit / s, for example, which makes it possible to ensure a satisfactory bit budget for the transform coding. Thus, the first predetermined value depends on the CELP core used for encoding the previous CELP frame. In addition, for the 16 kHz core coding, threshold values can be applied when assigning a bit rate to the CELP coding. Thus, the assigned bit rate is furthermore equal to the maximum between the bit rate of the transform coded transition frame and at least a second predetermined bit rate value, the second value being less than the first value. The second predetermined value of bit rate may for example be equal to 14.8 kbit / s. Thus, if the bit rate of the transform encoded transition frame is less than 14.8 kbit / s, the bit rate assigned to the CELP encoding of the transition subframe may be 14.8 kbit / s.

In a complementary embodiment, if the bit rate of the transform coded transition frame is less than 8 kbit / s, the bit rate assigned may be 8 kbit / s.

 Si core_bitrate ≤ 8000
     cbrate = 8000
 Sinon si core_bitrate ≤ 14800
     cbrate = 14800
 Sinon
     cbrate = min(core_bitrate , 22600)
 Fin si

Thus, according to this complementary embodiment, the following algorithm is obtained:

 If core_bitrate ≤ 8000
     cbrate = 8000
 Otherwise if core_bitrate ≤ 14800
     cbrate = 14800
 If not
     cbrate = min (core_bitrate, 22600)
 End if 

• num_bits := num_bits - 1

At an optional step 407, the affected bit rate is compared to a CID rate of 11.60 kbit / s. If the assigned bit rate is higher, a bit may be reserved to encode an adaptive dictionary low pass filter bit indication. The num_bits variable is updated:

• num_bits: = num_bits - 1

In a step 408, in the same way as in step 404, a first number of bits budget1 is allocated for the predictive coding of the additional CELP subframe, and budget1 depends on the bit rate assigned to the CELP coding of the sub-frame. -Transition frame.

At a step 410 common to codings at different core frequencies, a second number of bits allocated for transform coding of the transition frame, denoted budg2, is calculated from the first number of bits budget1 and the total number of bits of the transition frame. Given the above calculations, budget2 is equal to the num_bits variable. In general, the mode of the current transition frame is assumed here imputed to the MDCT coding budget, so this information is not explicitly taken into account here.

The preceding steps may have been implemented for the coding of a low frequency band of the transition subframe, in the case where the audio signal is decomposed into at least one low band of frequencies and a high band of frequencies. . At a step 409 the method may include the allocation of a third predetermined number of bits, denoted budg3, to the coding of the high frequency band of the transitional subframe, prior to step 410, which is also common to codings at different core frequencies. . In this case, the second number of bits budg2 is calculated both from the first number of bits budg1 and the third number of bits budg3.

As previously explained, the coding of the high frequency band (or band extension) of the transition subframe may be based on a correlation between the previous frame of the audio signal and the transition subframe. For example, the coding of the high frequency band can be broken down into two stages.

In a first step, the previous frame and the current frame of the audio signal are filtered by a high-pass filter to keep only the upper part of the spectrum. The upper part of the spectrum may correspond to frequencies higher than that of the CELP core used. For example, if the CELP core used is the 12.8 kHz CELP core, the high band is the audio signal for which frequencies below 12.8 kHz have been filtered. Such filtering can be implemented by means of a FIR filter.

In a second step, a correlation search between the filtered parts of the previous frame and the current frame is performed. Such a correlation search makes it possible to estimate a delay parameter and then a gain. The gain corresponds to the amplitude ratio between the filtered part of the current frame and the predicted signal by applying the delay.

For example, 6 bits can be allocated for gain and 6 bits for delay. The third number of bits budget3 is then equal to 12.

• num_bits := num_bits - budg3.

The num_bits variable can then be updated:

• num_bits: = num_bits - budget3.

The second number of bits budg2 is then equal to the variable num_bits update.

The figure 5 illustrates an audio decoder 500 according to one embodiment of the invention and the figure 6 is a diagram illustrating the steps of a decoding method according to an embodiment of the invention, implemented in the audio decoder 500 of the figure 5 .

The decoder 500 comprises a reception unit 501 for receiving, at a step 601, the coded digital signal (or bit stream) coming from the coder 100 of the figure 1 . The bit stream is subjected to a classification unit 502 able to determine, in a step 602 whether the current frame is a CELP frame, an MDCT frame or a transition frame. For this purpose, the classification unit 502 is able to deduce from the bit stream information indicating whether the current frame is a transition frame or not, and information indicating the CELP core to be used for decoding a CELP frame or a CELP transition subframe.

At a step 603, it is verified that the current frame is a transition frame.

If the current frame is not a transition frame, it is verified at a step 604 that the current frame is a CELP frame. If this is the case, the frame is transmitted to a decoder CELP 504 capable of decoding a CELP frame at a step 605, at the core frequency indicated by the classification unit 502. Following the decoding of a CELP frame, the The decoder CELP 504 can store, in a step 606, in a memory 506 parameters such as the coefficients of the linear prediction filter A (z) and internal states such as the predictive energy in the case where the following frame would be a frame of transition.

At the output of the CELP decoder 504, the signal can be resampled, at a step 607, at the output frequency of the decoder 500 by a resampling unit 505. In one embodiment of the invention, the unit resampling includes an FIR filter and resampling introduces a delay of (for example) 1.25 ms. In one embodiment, postprocessing may be applied to the CELP decoding before or after resampling.

As mentioned above, in one embodiment, a band extension can also be performed by a band extension management unit 5051 at steps 6071 and 6151, with associated decoding of the high band when the current frame is of type CELP. The high band is then combined with the CELP coding with possibly additional delay applied to the low band CELP synthesis.

The signal decoded by the decoder CELP and resampled, possibly post-processed before or after resampling, is transmitted to an output interface 510 of the decoder at a step 608.

The decoder 500 further comprises an MDCT decoder 507. In the case where it has been determined in the step 604 that the current frame is an MDCT frame, the decoder MDCT 507 is able to decode the frame MDCT in a conventional way at one stage. 609. In addition, a delay corresponding to the delay required for the application of resampling of the signal from the CELP decoder 504 is applied to the output of the decoder by a delay unit 508, so as to synchronize the synthesis MDCT with the synthesis. CELP, at step 610. The MDCT decoded and delayed signal is transmitted to the output interface 510 of the decoder at step 608.

In the case where the current frame is determined to be a transition frame at the end of step 603, a device for determining the bit distribution 503 is able to determine, in a step 611, the first number of bits. budget1 allocated to the CELP encoding of the transition subframe and the second number of budget bits3 allocated to the transform encoding of the transition frame. The device 503 may correspond to the device 700 described in detail with reference to the figure 7 .

The MDCT decoder 507 uses the third number of bits budget3 calculated by the determination unit 503 to adjust the rate necessary to decode the transition frame. The MDCT decoder 507 also sets the memory of the MDCT transformation to zero and decodes the transition frame at a step 612. The signal from the MDCT decoder is then delayed by the delay unit 508 at a step 613.

In parallel, the CELP decoder 504 decodes the transition CELP subframe according to the first number of bits budget1, at a step 614. The decoder CELP 504 decodes for this purpose the CELP parameters which can depend on the class of the current frame , and which include for example the pitch values of the adaptive dictionary, fixed dictionary and gains of the CELP subframe and uses the coefficients of the linear prediction filter. In addition, the CELP decoder 504 updates the CELP decoding states. These states may typically include the predictive energy of the innovation from the previous CELP frame to generate the 4ms or 5ms signal subframe depending on whether the CELP core at 12.8 kHz or 16 kHz is used (in the case of a restricted coding of the transition CELP subframe).

As mentioned above, the demand WO2012 / 085451 provides additional coding of a 5ms subframe for the 12.8 kHz CELP core and two additional 4ms subframes for the 16 kHz CELP core.

As explained with reference to the figure 3 in the case at 12.8 kHz, the 6.25 ms delay is not filled and the overlap is impacted: the decoder is only 0.625 ms of overlap, which is insufficient.

In the case of 16 kHz, two additional CELP subframes are coded at the beginning of the transition frame, which leaves very little budget for encoding the transition MDCT frame and can lead to a drop in quality compared to a "full rate" MDCT coding in the current frame.

Thus, the solution of international demand WO2012 / 085451 is not satisfactory.

An independent aspect of the invention provides, from a single additional CELP transition sub-frame, the partial generation of a second sub-frame by reusing coding parameters used for the coding of the CELP sub-frame of transition. The delay is thus bridged, ensuring a sufficient addition-recovery, and without impacting the rate of the MDCT coding of the transition frame.

• réception, à l'étape 501, d'un premier ensemble de paramètres de codage prédictif codant une première trame du signal numérique ;
• décodage prédictif, à l'étape 605, de la première trame sur la base du premier ensemble de paramètres de codage prédictif ;
• réception, à l'étape 501 pour une nouvelle trame, d'un deuxième ensemble de paramètres de codage prédictif d'une première sous-trame de transition d'une trame de transition codée par transformée :
• décodage, à l'étape 614, de la première sous-trame de transition sur la base du deuxième ensemble de paramètres de codage prédictif ;
• génération, à l'étape 614, d'échantillons d'une deuxième sous-trame de transition, à partir d'au moins un paramètre de codage prédictif du deuxième ensemble.

For this purpose, the invention also provides a method for decoding P of a coded digital signal, in a decoder 500 capable of decoding signal frames according to a predictive decoding or according to a transform decoding, comprising the following steps:

• receiving, in step 501, a first set of predictive coding parameters encoding a first frame of the digital signal;
• predictive decoding, in step 605, of the first frame based on the first set of predictive coding parameters;
• receiving, in step 501 for a new frame, a second set of predictive coding parameters of a first transition subframe of a transform coded transition frame:
• decoding, in step 614, the first transition subframe based on the second set of predictive coding parameters;
• generating in step 614 samples of a second transition subframe from at least one predictive coding parameter of the second set.

The invention also aims at the decoder 500 for implementing the decoding method P, as well as a computer program comprising instructions for implementing the decoding method P, when these instructions are executed by a processor.

The CELP parameters reused for the generation of the second sub-frame can be the vector of gains, the vector of the adaptive dictionary and the vector of the fixed dictionary.

According to an embodiment of the decoding method P, a minimum overlap value can be predefined for the transform decoding and the number of generated samples of the second subframe is determined according to the minimum overlap value. This latter sub-frame can be generated without additional information by extending the CELP synthesis by repeating the pitch prediction with the same pitch delay and the same adaptive dictionary gain as in the first subframe, and performing a LPC filtering of synthesis with the same LPC coefficients and a de-emphasis or de-emphasis.

The second CELP sub-frame can then be truncated so as to retain only 1.25 ms of signal in the case of the CELP core at 12.8 kHz, and 2.25 ms of signal in the case of the 16 kHz CELP core. . The first sub-frame CELP is thus completed so as to have 6.25 ms of additional signal making it possible to fill the hole and to ensure a satisfactory addition-recovery (minimum value of overlap, for example of 1.875 ms) with the frame of MDCT transition. In one embodiment, the additional CELP subframe has a length extended to 6.25 ms for CELP cores at 12.8 and 16 kHz, which implies modifying the "normal" CELP coding to have such a extended subframe length, especially for the fixed dictionary.

In addition to the previous embodiment of the decoding method P, the method P may further comprise a re-sampling step 615 implemented by a finite impulse response filter. As previously explained, the FIR filter can be integrated in the resampling unit 505. The resampling uses the memory of the FIR filter of the previous CELP frame and the processing induces an additional delay of 1.25 ms in this example. .

The method P may further comprise a step of adding an additional signal obtained from samples stored in the memory of the finite impulse response filter, to fill the delay introduced by the resampling step. Thus, 1.25 ms of signal, in addition to the 6.25 ms of additional signal previously generated, are generated by the decoder 500, these samples advantageously making it possible to bridge the delay introduced by the resampling of the 6.25 ms of signal. additional.

For this purpose, the memory of the FIR filter of the resampling unit 505 can be saved at each frame after CELP decoding. The number of samples in this memory corresponds to 1.25 ms at the frequency of the CELP core considered (12.8 or 16 kHz).

According to a complementary embodiment of the method P, the resampling of the stored samples is performed by an interpolation method introducing a second delay less than the first delay of the finite impulse response filter, which can be considered as zero. Thus, the 1.25 ms of signal generated from the memory of the FIR filter, are resampled according to a method involving a lesser delay. For example, a re-sampling of the 1.25 ms of signal generated by the memory of the FIR filter can be implemented by cubic interpolation, which involves a delay of only two samples, less delay compared to the delay of the FIR filter. Thus, two additional signal samples are required to re-sample the aforementioned 1.25 ms of signal: these two additional samples can be obtained by repeating the last value of the re-sampling memory of the FIR filter.

The decoder may further decode the high frequency part of the 6.25 msec CELP signal obtained from the first and second transition subframes. For this purpose, the decoder CELP 504 can use the adaptive gain and the vector of the fixed dictionary of the last subframe of the previous CELP frame.

The decoder 500 further comprises a summation-recovery unit 509 capable of providing the overlap-addition, at a step 616, between the decoded and resampled CELP transition subframes, the samples resampled by the cubic interpolation, and the decoded signal of the transition frame from the MDCT decoder 507.

For this purpose, the unit 509 applies the modified window 327 of synthesis of the figure 3 . Thus, before the MDCT fold point for the first two quarters, the windowed samples are set to zero. After the aforementioned folding point, the windowed samples are divided by the unmodified window 324 of the figure 3 , and multiplied by a sine-type window so that, combined with the window applied to the encoder, the total window is sin ² . In the overlap-addition part, samples from the CELP and 0-delay resampling (for example by cubic interpolation) are weighted by a cos ² window.

The transition frame thus obtained is transmitted to the output interface 510 of the decoder at step 608.

The figure 7 represents an example of a device 700 for determining the distribution of the bits of a transition frame.

The device comprises a random access memory 704 and a processor 703 for storing instructions for carrying out the method for determining the distribution of the bits of a transition frame described above. The device also has a mass memory 705 for storing data to be retained after the application of the method. The device 700 further comprises an input interface 701 and an output interface 706 respectively intended to receive the frames of the digital signal and to issue the detail of the budget allocated to these different frames.

The device 700 may further comprise a digital signal processor (DSP) 702. This DSP 702 receives the digital signal frames for shaping, demodulating and amplifying, in a manner known per se, these frames.

The present invention is not limited to the embodiments described above as examples; it extends to other variants.

Thus, an embodiment has been described above in which the compression or decompression devices are entities in their own right. Of course, these devices can be embedded in any type of larger device such as a digital camera, a camera, a mobile phone, a computer, a movie projector, etc.

In addition, an embodiment has been described proposing a particular architecture of compression, decompression and comparison devices. These architectures are given for illustrative purposes only. Thus, an arrangement of the components and a different distribution of the spots assigned to each of these components is also conceivable. For example, the tasks performed by the digital signal processor (DSP) can also be performed by a conventional processor.

Claims (15)

1. Method for determining a distribution of coding bits of a transition frame (321; 322), said method being implemented in a coder/decoder (100; 500) for coding/decoding a digital audio signal, the transition frame being preceded by a preceding predictively coded frame (320), the coding of the transition frame comprising transform coding and predictive coding of a single subframe of the transition frame, the method being characterized by the following steps:

   - assigning (402; 405) a bit rate for the predictive coding of the transition subframe, said bit rate being equal to the minimum between the bit rate of the transform coding of the transition frame and a first predetermined bit rate value;

   - determining (404; 408) a first number of bits allocated for the predictive coding of the transition subframe for said bit rate; and

   - calculating (410) a second number of bits allocated for the transform coding of the transition frame on the basis of the first number of bits and of a number of bits available for the coding of the transition frame.
2. Method according to Claim 1, wherein the coder/decoder (100; 500) comprises a first core operating at a first frequency for the predictive coding/decoding of a signal frame, and a second core operating at a second frequency for the predictive coding/decoding of a signal frame,
   wherein the first predetermined bit rate value depends on the core selected from among the first and second cores for the coding/decoding of the preceding predictively coded frame (320).
3. Method according to Claim 2, when the first core has been selected for the coding/decoding of the preceding predictively coded frame (320), the assigned bit rate is furthermore equal to the maximum between the bit rate of the transform-coded transition frame (322) and at least one second predetermined bit rate value, the second value being lower than the first value.
4. Method according to one of the preceding claims, wherein the digital audio signal is decomposed into at least one frequency low band and one frequency high band,
   wherein the calculated first number of bits is assigned to the predictive coding of the transition subframe (321) for the frequency low band, and wherein a predetermined third number of bits is allocated to a coding of the transition subframe for the frequency high band,
   and wherein the second number of bits allocated for the transform coding of the transition frame (322) is furthermore determined on the basis of the predetermined third number of bits.
5. Method according to Claim 4, wherein the number of bits available for the coding of the transition frame (321; 322) is fixed.
6. Method according to Claim 5, wherein the second number of bits is equal to the fixed number of coding bits of the transition frame (321; 322) minus the first number of bits minus the third number of bits.
7. Method according to Claim 5, wherein the second number of bits is equal to the fixed number of coding bits of the transition frame (321; 322) minus the first number of bits minus the third number of bits minus a first bit minus a second bit,
   the first bit indicating whether low-pass filtering is carried out when determining the parameters of the predictive coding of the transition subframe, said parameters relating to the tonal delay,
   the second bit indicating the frequency used by the core of the coder/decoder for the predictive coding/decoding of the transition subframe.
8. Method for coding a digital audio signal in a coder (100) able to code signal frames using predictive coding or using transform coding, including the following steps:

   * coding a preceding frame (301) of samples of the digital audio signal using predictive coding;

   * coding a current frame (302) of samples of the digital audio signal into a transition frame (321; 322), the coding of the transition frame (321; 322) comprising transform coding and predictive coding of a single subframe (321) of the transition frame, said coding of the current frame (302) including the following sub-steps:

   - determining (209) the distribution of the bits according to one of the preceding claims;

   - transform coding (212) the transition frame (322) on the second number of allocated bits;

   - predictively coding (213) the transition subframe (321) on the first number of allocated bits.
9. Coding method according to Claim 8, wherein the predictive coding comprises generating determined predictive coding parameters for said allocated bit rate.
10. Coding method according to either of Claims 8 and 9, wherein the predictive coding comprises generating predictive coding parameters that are limited with respect to the predictive coding of the preceding frame (320) by reusing at least one parameter of the predictive coding of the preceding frame.
11. Method for decoding a coded digital audio signal, implemented in a decoder (500) able to decode signal frames using predictive coding or using transform decoding, including the steps of:

    * predictively decoding (605) a preceding frame of samples of the digital audio signal coded using predictive coding;

    * decoding a transition frame (321; 322) coding a current frame of samples of the digital audio signal, the coding of the transition frame comprising transform coding and predictive coding of a single subframe (321) of the transition frame, including the sub-steps of:

    - determining (611) the distribution of the bits according to one of Claims 1 to 7;

    - predictively decoding (614) the transition subframe (321) on the first number of allocated bits;

    - transform decoding (612) the transition frame (322) on the second number of allocated bits.
12. Computer program including instructions for implementing the method according to any one of the preceding claims when these instructions are executed by a processor.
13. Device for determining a distribution of coding bits of a transition frame (321; 322), said device (104; 503) being implemented in a coder/decoder for coding/decoding a digital audio signal, the transition frame being preceded by a preceding predictively coded frame (320), the coding of the transition frame comprising transform coding and predictive coding of a single subframe (321) of the transition frame, the number of coding bits of the transition frame being fixed, the device being characterized by a processor designed to carry out the following operations:

    - assigning a bit rate for the predictive coding of the transition subframe, said bit rate being equal to the minimum between the bit rate of the transform coding of the transition frame and a first predetermined bit rate value;

    - determining a first number of bits allocated for the predictive coding of the transition subframe for said bit rate, and;

    - calculating a second number of bits allocated for the transform coding of the transition frame on the basis of the number of bits required for the coding of the coding parameters and the fixed number of coding bits of the transition frame.
14. Coder able to code frames of a digital audio signal using predictive coding or using transform coding, including:

    * a device (104) according to Claim 13;

    * a predictive coder (103) including a processor designed to carry out the following operations:

    - coding a preceding frame of samples of the digital audio signal using predictive coding;

    - predictively coding a single subframe contained in a transition frame coding a current frame of samples of the digital audio signal, the coding of the transition frame comprising transform coding and predictive coding of said subframe, the processor being designed to carry out the predictive coding of the transition subframe on the first number of allocated bits;

    * a transform coder (105) including a processor designed to carry out the operation of transform coding the transition frame on the second number of allocated bits.
15. Decoder for a digital audio signal coded by way of predictive coding and by way of transform coding, including:

    * a device (503) according to Claim 13;

    * a predictive decoder (504) including a processor designed to carry out the following operations:

    - predictively decoding a preceding frame (320) of samples of the digital audio signal coded using predictive coding;

    - predictively decoding a single subframe (321) contained in a transition frame coding a current frame of samples of the digital audio signal, the coding of the transition frame comprising transform coding and predictive coding of said subframe, the processor being designed to carry out the operation of predictively decoding the transition subframe on the first number of allocated bits;

    * a transform decoder (507) including a processor designed to carry out the operation of transform decoding the transition frame (322) on the second number of allocated bits.

Images