EP0801790B1 - Speech coding method using synthesis analysis - Google Patents

Abstract

PCT No. PCT/FR96/00006 Sec. 371 Date Oct. 22, 1997 Sec. 102(e) Date Oct. 22, 1997 PCT Filed Jan. 3, 1996 PCT Pub. No. WO96/21220 PCT Pub. Date Jul. 11, 1996A linear prediction analysis is performed for each frame of a speech signal to determine the coefficients of a short-term synthesis filter and an open-loop analysis is performed to determine a degree of frame voicing. At least one closed-loop analysis is performed for each sub-frame to determine an excitation sequence which, when applied to the short-term synthesis filter, generates a synthetic signal representative of the speech signal. Each closed-loop analysis uses the impulse response of a filter consisting of the short-term synthesis filter and a perceptual weighting filter, by truncating the impulse response to a truncation length that is no greater than the number of samples per sub-frame and is dependent on the energy distribution of the response and the degree of voicing of the frame.

Description

The present invention relates to speech coding using synthetic analysis.

The Applicant has in particular described such coders of speech that she developed in her patent applications European 0 195 487, 0 347 307 and 0 469 997.

In a speech coder with synthesis analysis, a linear prediction of the speech signal is carried out to obtain the coefficients of a short-term synthesis filter modeling the transfer function of the vocal tract. These coefficients are transmitted to the decoder, as well as parameters characterizing an excitation to be applied to the short-term synthesis filter. In most of the current coders, further research is carried out on the longer-term correlations of the speech signal in order to characterize a long-term synthesis filter accounting for the pitch of the speech. When the signal is voiced, the excitation indeed has a predictable component which can be represented by the past excitation, delayed by TP samples of the speech signal and affected by a gain g _P. The long-term synthesis filter, also reconstituted at the decoder, then has a transfer function of the form 1 / B (z) with B (z) = 1-g _P .z ^-TP . The remaining, unpredictable part of the excitation is called stochastic excitation. In so-called CELP ("Code Excited Linear Prediction") coders, the stochastic excitation consists of a vector sought in a predetermined dictionary. In so-called Multi-Pulse Linear Prediction Coding (MPLPC) coders, the stochastic excitation comprises a certain number of pulses whose positions are sought by the coder. In general, CELP coders are preferred for low transmission rates, but they are more complex to implement than MPLPC coders.

To determine the long-term prediction lag, we frequently use a closed loop analysis contributing directly to minimize the perceptually weighted deviation between the speech signal and the synthetic signal. The downside of this closed loop analysis is that it is demanding in terms of volume of calculations, because the selection of a delay implies the evaluation of a certain number of delays candidates and each delay assessment requires calculations of convolutional products between delayed excitation and the impulse response of the weighted synthesis filter perceptually. The above disadvantage also exists for looking for stochastic excitement, which is also a closed loop process involving products of convolution with this impulse response. The excitement varies faster than spectral parameters characteristics of the short-term synthesis filter. The excitement (predictable and stochastic) is typically determined once per 5ms subframe, while the parameters spectral are once per 20 ms frame. The complexity and the frequency of the closed-loop search for excitement make it the most critical step in how quickly calculations required in a speech coder.

A main object of the invention is to propose a speech coding method of reduced complexity with respect to concerns the closed-loop analysis (s).

The invention thus provides an analysis coding method by synthesizing a speech signal digitized in frames successive subdivided into subframes with a number determined of samples, in which one performs for each frame a linear prediction analysis of the speech signal to determine the coefficients of a synthesis filter at short term, and an open loop analysis to determine a degree of voicing of the frame, and we perform for each subframe at least one closed loop analysis to determine an excitation sequence which, subject to the short-term synthesis, produces a representative synthetic signal of the speech signal. Each closed loop analysis uses the impulse response of a filter composed of the filter short-term summary and a weighting filter perceptual. During each closed loop analysis, we uses said impulse response by truncating it to a truncation length at most equal to the number of samples by subframe and dependent on energy distribution of said response and the degree of voicing of the frame.

In general, the length of truncation will be as much larger than the frame is seen. We can thus reduce significantly the complexity of closed loop analyzes without lose in quality of coding, thanks to an adaptation to signal voicing characteristics.

• la figure 1 est un schéma synoptique d'une station de radiocommunication incorporant un codeur de parole mettant en oeuvre l'invention ;
• la figure 2 est un schéma synoptique d'une station de radiocommunication apte à recevoir un signal produit par celle de la figure 1 ;
• les figures 3 à 6 sont des organigrammes illustrant un processus d'analyse LTP en boucle ouverte appliqué dans le codeur de parole de la figure 1 ;
• la figure 7 est un organigramme illustrant un processus de détermination de la réponse impulsionnelle du filtre de synthèse pondéré appliqué dans le codeur de parole de la figure 1 ;
• les figures 8 à 11 sont des organigrammes illustrant un processus de recherche de l'excitation stochastique appliqué dans le codeur de parole de la figure 1.

Other features and advantages of the invention will appear in the description below of preferred, but non-limiting, examples of embodiment, with reference to the appended drawings, in which:

• FIG. 1 is a block diagram of a radiocommunication station incorporating a speech coder implementing the invention;
• Figure 2 is a block diagram of a radio station capable of receiving a signal produced by that of Figure 1;
• Figures 3 to 6 are flowcharts illustrating an open loop LTP analysis process applied in the speech coder of Figure 1;
• Figure 7 is a flow diagram illustrating a process for determining the impulse response of the weighted synthesis filter applied in the speech coder of Figure 1;
• FIGS. 8 to 11 are flowcharts illustrating a process for finding the stochastic excitation applied in the speech coder of FIG. 1.

A speech coder implementing the invention is applicable in various types of speech transmission and / or storage systems using a digital compression technique. In the example of Figure 1, the speech coder 16 is part of a mobile radio station. The speech signal S is a digital signal sampled at a frequency typically equal to 8 kHz. The signal S comes from an analog-digital converter 18 receiving the amplified and filtered output signal from a microphone 20. The converter 18 puts the speech signal S in the form of successive frames themselves subdivided into nst sub-frames of 1st samples. A frame of 20 ms typically comprises nst = 4 sub-frames of lst = 40 samples of 16 bits at 8 kHz. Upstream of the encoder 16, the speech signal S can also be subjected to conventional shaping treatments such as Hamming filtering. The speech coder 16 delivers a binary sequence with a significantly lower bit rate than that of the speech signal S, and addresses this sequence to a channel coder 22 whose function is to introduce redundancy bits into the signal in order to allow detection and / or a correction of any transmission errors. The output signal from the channel encoder 22 is then modulated on a carrier frequency by the modulator 24, and the modulated signal is transmitted on the air interface.

The speech coder 16 is a synthesis analysis coder. The coder 16 determines on the one hand parameters characterizing a short-term synthesis filter modeling the speaker's vocal tract, and on the other hand an excitation sequence which, applied to the short-term synthesis filter, provides a synthetic signal constituting an estimate of the speech signal S according to a perceptual weighting criterion.

The short-term synthesis filter has a function transfer form 1 / A (z), with:

The coefficients a _i are determined by a module 26 for short-term linear prediction analysis of the speech signal S. The a _i are the linear prediction coefficients of the speech signal S. The order q of the linear prediction is typically of the order of 10. The methods applicable by module 26 for short-term linear prediction are well known in the field of speech coding. Module 26, for example, implements the Durbin-Levinson algorithm (see J. Makhoul: "Linear Prediction: A tutorial review", Proc. IEEE, Vol.63, N ° 4, April 1975, p. 561-580 ). The coefficients a _i obtained are supplied to a module 28 which converts them into spectral line parameters (LSP). The representation of the prediction coefficients a _i by LSP parameters is frequently used in speech coders with analysis by synthesis. The LSP parameters are the q numbers cos (2πf _i ) arranged in descending order, the q spectral line frequencies (LSF) normalized f _i (1≤i≤q) being such that the complex numbers exp (2πf _i ), with 1 = 1.3, ..., q-1, q + 1 and f _{q + 1} = 0.5, let the roots of the polynomial Q (z) defined by Q (z) = A (z) + z ^{- ( q + 1)} .A (z ^-1 ) and that the complex numbers exp (2πjf _i ), with i = 0,2,4, ..., q and f ₀ = 0, are the roots of the polynomial Q * ( z) defined by Q * (z) = A (z) -z ^{- (q + 1)} .A (z ^-1 ).

et A(z) =[Q(z)+Q*(z)]/2 The LSP parameters can be obtained by the conversion module 28 by the classical method of Chebyshev polynomials (see P. Kabal and RP Ramachandran: "The computation of line spectral frequencies using Chebyshev polynomials", IEEE Trans. ASSP, Vol.34, N ° 6, 1986, pages 1419-1426). These are quantization values of the LSP parameters, obtained by a quantization module 30, which are transmitted to the decoder so that the latter finds the coefficients a _i of the short-term synthesis filter. The coefficients a _i can be found simply, given that:

and AT (( z ) = [ Q (( z ) + Q * ( z )] / 2

To avoid sudden variations in the transfer function of the short-term synthesis filter, the LSP parameters are interpolated before the prediction coefficients a _{i are} deduced therefrom. This interpolation is performed on the first sub-frames of each frame of the signal. For example, if LSP _t and LSP _t-1 respectively designate a LSP parameter calculated for frame t and for the previous frame t-1, we take: LSP _t (0) = 0.5. LSP _t-1 +0, 5.LSP _t , LSP _t (1) = 0.25.LSP _t-1 + 0.75.LSP _t and LSP _t (2) = ... = LST _t (nst-1) = LSP _t for sub -frames 0,1,2, ..., nst-1 of the frame t. The coefficients a _i of the filter 1 / A (z) are then determined, sub-frame by sub-frame from the interpolated LSP parameters.

The non-quantified LSP parameters are supplied by the module 28 to a module 32 for calculating the coefficients of a perceptual weighting filter 34. The perceptual weighting filter 34 preferably has a transfer function of the form W (z) = A (z / γ ₁ ) / A (z / γ ₂ ) where γ ₁ and γ ₂ are coefficients such that γ ₁ > γ ₂ > 0 (for example γ ₁ = 0.9 and γ ₂ = 0.6). The coefficients of the perceptual weighting filter are calculated by the module 32 for each subframe after interpolation of the LSP parameters received from the module 28.

The perceptual weighting filter 34 receives the speech signal S and delivers a perceptually weighted SW signal which is analyzed by modules 36, 38, 40 for determine the excitation sequence. The excitation sequence of the short-term filter consists of an excitation predictable by a long-term synthetic filter modeling the pitch of the speech, and an excitement unpredictable stochastic, or innovation sequence.

Module 36 performs long-term prediction (LTP) in open loop, i.e. it does not contribute directly to the minimization of the weighted error. In the case shown, the weighting filter 34 intervenes in upstream of the open loop analysis module, but it could otherwise: module 36 could operate directly on the speech signal S or on the signal S cleared of its short-term correlations by a filter transfer function A (z). In contrast, modules 38 and 40 operate in a closed loop, i.e. they directly contribute to minimizing the error perceptually weighted.

The long-term synthesis filter has a transfer function of the form 1 / B (z) with B (z) = 1-g _P .z ^-TP where g _P denotes a long-term prediction gain and TP denotes a long-term prediction delay. The long-term prediction delay can typically take N = 256 values between rmin and rmax samples. A fractional resolution is provided for the smallest delay values so as to avoid discernible differences in terms of voicing frequency. For example, use a resolution 1/6 between rmin = 21 and 33 + 5/6, a resolution 1/3 between 34 and 47 + 2/3, a resolution 1/2 between 48 and 88 + 1/2, and a integer resolution between 89 and rmax = 142. Each possible delay is thus quantified by an integer index between 0 and N-1 = 255.

Long-term prediction lag is determined in two steps. In the first step, the analysis module 36 Open loop LTP detects voiced frames from the speech signal and determines, for each voiced frame, a degree of voicing MV and a delay search interval long-term prediction. The degree of voicing MV of a voiced frame can take three values: 1 for frames weakly voiced, 2 for moderately voiced frames, and 3 for very close fields. In the notations used below, we take a degree of voicing MV = 0 for the frames not seen. The search interval is defined by a central value represented by its quantification index ZP and by a width in the field of quantification indexes, depending on the degree of voicing MV. For frames weakly or moderately voiced (MV = 1 or 2) the width of the search interval is N1 index, i.e. long term prediction lag index will be searched between ZP-16 and ZP + 15 if N1 = 32. For highly visible frames (MV = 3), the width of the search interval is N3 index, i.e. the index of the prediction delay at long term will be sought between ZP-8 and ZP + 7 if N3 = 16.

Once the degree of voicing MV of a frame has been determined by the module 36, the module 30 operates the quantization of the LSP parameters which have previously been determined for this frame. This quantization is for example vectorial, that is to say it consists in selecting, from one or more predetermined quantization tables, a set of quantized parameters LSP _Q which has a minimum distance from the set of parameters LSP provided by the module 28. In known manner, the quantification tables differ according to the degree of voicing MV provided to the quantization module 30 by the open-loop analyzer 36. A set of quantization tables for a degree of voicing MV is determined, during prior tests, so as to be statistically representative of frames having this degree MV. These sets are stored both in the coders and in the decoders implementing the invention. The module 30 delivers the set of quantized parameters LSP _{Q as} well as its index Q in the applicable quantification tables.

The speech coder 16 further comprises a module 42 for calculating the impulse response of the compound filter short-term summary filter and weighting filter perceptual. This compound filter has the function of transfer W (z) / A (z). For the calculation of its impulse response h = (h (0), h (1), ..., h (1st-1)) over the duration of a subframe, module 42 takes for the weighting filter perceptual W (z) that corresponding to the LSP parameters interpolated but not quantified, i.e. the one whose coefficients were calculated by module 32, and for the synthesis filter 1 / A (z) the one corresponding to the parameters Quantified and interpolated LSP, i.e. the one that will actually reconstructed by the decoder.

où x(i) désigne le signal de parole pondéré SW de la sous-trame auquel on a soustrait la mémoire du filtre de synthèse pondéré (c'est-à-dire la réponse à un signal nul, due à ses états initiaux, du filtre dont la réponse impulsionnelle h a été calculée par le module 42), et y_T(i) désigne le produit de convolution :

u(j-T) désignant la composante prédictible de la séquence d'excitation retardée de T échantillons, estimée par la technique bien connue du répertoire adaptatif ("adaptive codebook"). Pour les retards T inférieurs à la longueur d'une sous-trame, les valeurs manquantes de u(j-T) peuvent être extrapolées à partir des valeurs antérieures. Les retards fractionnaires sont pris en compte en suréchantillonnant le signal u(j-T) dans le répertoire adaptatif. Un suréchantillonnage d'un facteur m est obtenu au moyen de filtres polyphasés interpolateurs.In the second step of determining the long-term prediction delay TP, the closed-loop LTP analysis module 38 determines the delay TP for each sub-frame of the voiced frames (MV = 1, 2 or 3). This delay TP is characterized by a differential value DP in the field of quantification indexes, coded on 5 bits if MV = 1 or 2 (N1 = 32), and on 4 bits if MV = 3 (N3 = 16). The TP delay index is ZP + DP. As is known, closed-loop LTP analysis consists in determining, in the search interval for long-term prediction delays T, the delay TP which maximizes, for each sub-frame of a voiced frame, the normalized correlation :

where x (i) denotes the weighted speech signal SW of the subframe from which the memory of the weighted synthesis filter has been subtracted (i.e. the response to a zero signal, due to its initial states, of the filter whose impulse response has been calculated by module 42), and y _T (i) denotes the convolution product:

u (jT) designating the predictable component of the delayed excitation sequence of T samples, estimated by the well-known technique of the adaptive codebook. For delays T less than the length of a subframe, the missing values of u (jT) can be extrapolated from the previous values. Fractional delays are taken into account by oversampling the signal u (jT) in the adaptive repertoire. An oversampling of a factor m is obtained by means of polyphase interpolating filters.

Toutefois, dans une version préférée de l'invention, le gain g_P est calculé par le module d'analyse stochastique 40.The gain g _P of long-term prediction could be determined by the module 38 for each sub-frame, by applying the known formula:

However, in a preferred version of the invention, the gain g _P is calculated by the stochastic analysis module 40.

The stochastic excitation determined for each subframe by the module 40 is of the multi-pulse type. An innovation sequence of 1st samples includes np pulses of positions p (n) and of amplitude g (n). Other says, the pulses have an amplitude of 1 and are associated to respective gains g (n). Since the LTP delay is not determined for the subframes of the non we can take a higher number of pulses for the stochastic excitation relating to these subframes, for example np = 5 if MV = 1, 2 or 3 and np = 6 if MV = 0. The positions and gains calculated by the analysis module 40 stochastics are quantified by a module 44.

• l'index Q des paramètres LSP quantifiés pour chaque trame ;
• le degré MV de voisement de chaque trame ;
• l'index ZP du centre de l'intervalle de recherche des retards LTP pour chaque trame voisée ;
• l'index différentiel DP du retard LTP pour chaque sous-trame d'une trame voisée, et le gain associé g_P ;
• les positions p(n) et les gains g(n) des impulsions de l'excitation stochastique pour chaque sous-trame.

A bit scheduling module 46 receives the various parameters which will be useful to the decoder, and constitutes the binary sequence transmitted to the channel coder 22. These parameters are:

• the index Q of the LSP parameters quantized for each frame;
• the degree MV of voicing of each frame;
• the ZP index of the center of the LTP delay search interval for each voiced frame;
• the differential index DP of the delay LTP for each sub-frame of a voiced frame, and the associated gain g _P ;
• the positions p (n) and the gains g (n) of the pulses of the stochastic excitation for each sub-frame.

Some of these parameters may be important particular in the quality of restitution of the speech or a particular sensitivity to errors of transmission. A module 48 is thus provided in the encoder which receives the different parameters and which adds to some of them redundancy bits to detect and / or correct any transmission errors. Through example, the degree of voicing MV coded on two bits being a critical parameter, we want it to reach the decoder with as few errors as possible. For this reason, redundancy bits are added to this parameter by module 48. We can for example add a bit of parity to the two bits encoding MV and repeat the three once bits thus obtained. This example of redundancy allows detect all single or double errors and correct all single errors and 75% of double errors.

The allocation of bit rate per 20 ms frame is for example that indicated in table I.

In the example considered here, the channel coder 22 is that used in the pan-European system of radiocommunication with mobiles (GSM). This channel coder, described in detail in Recommendation GSM 05.03, was developed for a 13 kbit / s speech coder of RPE-LTP type which also produces 260 bits per 20 ms frame. The sensitivity of each of the 260 bits was determined from listening tests. The bits from the source encoder have been grouped into three categories. The first of these categories IA groups 50 bits which are coded convolutionally on the basis of a generator polynomial giving a half redundancy with a constraint length equal to 5. Three parity bits are calculated and added to the 50 bits of the category IA before convolutional coding. The second category (IB) has 132 bits which are protected at a rate of a half by the same polynomial as the previous category. The third category (II) contains 78 unprotected bits. After application of the convolutional code, the bits (456 per frame) are subjected to interleaving. The scheduling module 46 of the new source coder implementing the invention distributes the bits in the three categories according to the subjective importance of these bits. quantified parameters MV = 0 MV = 1 or 2 MV = 3 LSP 34 34 34 MV + redundancy 6 6 6 ZP - 8 8 DP - 20 16 g _TP - 20 24 pulse positions 80 72 72 pulse gains 140 100 100 Total 260 260 260

A mobile radio station capable of receiving the speech signal processed by the source encoder 16 is shown schematically in Figure 2. The radio signal received is first processed by a demodulator 50 then by a channel 52 decoder which performs dual operations of those of modulator 24 and channel encoder 22. The decoder channel 52 provides the speech decoder 54 with a binary sequence which, in the absence of transmission errors or when any errors have been corrected by the decoder channel 52, corresponds to the binary sequence delivered by the scheduling module 46 at the coder 16. The decoder 54 includes a module 56 which receives this binary sequence and which identifies the parameters relating to the different frames and subframes. The module 56 also performs some checks on the parameters received. In particular, module 56 examines the redundancy bits introduced by the encoder module 48, to detect and / or correct the errors affecting the parameters associated with these bits of redundancy.

For each speech frame to be synthesized, a module 58 of the decoder receives the degree of voicing MV and the index of Q for quantizing the LSP parameters. The module 58 finds the quantized LSP parameters in the tables corresponding to the value of MV, and, after interpolation, converts them into coefficients a _i for the short-term synthesis filter 60. For each speech sub-frame to be synthesized, a pulse generator 62 receives the positions p (n) of the np pulses of the stochastic excitation. The generator 62 delivers pulses of unit amplitude which are each multiplied by 64 by the associated gain g (n). The output of amplifier 64 is addressed to the long-term synthesis filter 66. This filter 66 has an adaptive directory structure. The output samples u of the filter 66 are stored in the adaptive directory 68 so as to be available for the subsequent subframes. The delay TP relative to a sub-frame, calculated from the quantization indices ZP and DP, is supplied to the adaptive repertoire 68 to produce the signal u suitably delayed. The amplifier 70 multiplies the signal thus delayed by the gain g _P of long-term prediction. The long-term filter 66 finally comprises an adder 72 which adds the outputs of amplifiers 64 and 70 to provide the excitation sequence u. When the LTP analysis has not been carried out at the coder, for example if MV = 0, a prediction gain g _P zero is imposed on the amplifier 70 for the corresponding sub-frames. The excitation sequence is addressed to the short-term synthesis filter 60, and the resulting signal can also, in known manner, be subjected to a post-filter 74 whose coefficients depend on the synthesis parameters received, to form the signal of synthetic speech S '. The output signal S 'of the decoder 54 is then converted into analog by the converter 76 before being amplified to control a loudspeaker 78.

We will now describe, with reference to Figures 3 at 6, the open loop LTP analysis process implemented work by the encoder module 36 according to a first aspect of the invention.

In a first step 90, the module 36 calculates and stores, for each sub-frame st = 0.1, ..., nst-1 of the current frame, the autocorrelations c _st (k) and the delayed energies G _st ( k) of the weighted speech signal SW for the whole delays k comprised between rmin and rmax:

The energies per sub-frame R0 _st are also calculated:

où R0 est l'énergie totale de la trame (R0 = R0₀+ R0₁+...+ R0_nst-1), et X_st(K_st)=C_st ²(K_st)/G_st(K_st) désigne le maximum déterminé à l'étape 90 relativement à la sous-trame st. Comme l'indique la figure 6, la comparaison 92 peut être effectuée sans avoir à calculer le logarithme.In step 90, the module 36 also determines, for each sub-frame st, the entire delay K _st which maximizes the open-loop estimation P _st (k) of the long-term prediction gain on the sub-frame st, excluding the delays k for which the autocorrelation c _st (k) is negative or smaller than a small fraction E of the energy R0 _st of the subframe. The estimate P _st (k) expressed in decibels is written: P st (k) = 20.log 10 [RO st / (RO st -VS st 2 (k) / G st (k))] Maximizing P _st (k) therefore amounts to maximizing the expression X _st (k) = C _st ² (k) / G _st (k) as shown in Figure 6. The entire delay K _st is the basic delay in resolution integer for the subframe st. Step 90 is followed by a comparison 92 between a first open loop estimate of the overall prediction gain over the current frame and a predetermined threshold S0 typically between 1 and 2 decibels (for example S0 = 1.5 dB). The first estimate of the overall prediction gain is equal to:

where R0 is the total energy of the frame (R0 = R0 ₀ + R0 ₁ + ... + R0 _nst-1 ), and X _st (K _st ) = C _st ² (K _st ) / G _st (K _st ) designates the maximum determined in step 90 relative to the subframe st. As shown in Figure 6, comparison 92 can be performed without having to calculate the logarithm.

If the comparison 92 shows a first estimate of the prediction gain below the threshold S0, it is considered that the speech signal contains too few long-term correlations to be seen, and the degree of voicing MV of the current frame is taken equal to 0 in step 94, which in this case ends the operations performed by the module 36 on this frame. If on the contrary the threshold S0 is exceeded in step 92, the current frame is detected as voiced and the degree MV will be equal to 1, 2 or 3. The module 36 then calculates, for each subframe st ^, a list I _st containing candidate delays to constitute the ZP center of the search interval for long-term prediction delays.

The operations performed by the module 36 for each subframe st (st initialized to 0 at step 96) of a voiced frame begin with the determination 98 of a selection threshold SE _st in decibels equal to a determined fraction β of the estimate P _st (K _st ) of the prediction gain in decibels on the subframe, maximized in step 90 (β = 0.75 typically). For each sub-frame st of a voiced frame, the module 36 determines the basic delay rbf in full resolution for the rest of the processing. This basic delay could be taken equal to the integer K _st obtained in step 90. The fact of finding the basic delay in fractional resolution around K _st however makes it possible to gain in precision. Step 100 thus consists in finding, around the integer delay K _st obtained in step 90, the fractional delay which maximizes the expression C _st ² / G _st. This search can be carried out at the maximum resolution of the fractional delays (1/6 in the example described here) even if the entire delay K _st is not in the domain where this maximum resolution applies. We determine for example the number Δ _st which maximizes C _st ² (K _st + δ / 6) / G _st (K _st + δ / 6) for -6 <δ <+6, then the basic delay rbf in maximum resolution is taken equal to K _st + Δ _st / 6. For the fractional values T of the delay, the autocorrelations C _st (T) and the delayed energies G _st (T) are obtained by interpolation from the values stored in step 90 for the whole delays. Of course, the basic delay relating to a sub-frame could also be determined in fractional resolution from step 90 and taken into account in the first estimation of the overall prediction gain on the frame.

Once the basic delay rbf has been determined for a subframe, an examination 101 of the submultiples of this delay is carried out in order to select those for which the prediction gain is relatively large (FIG. 4), then multiples of the smallest sub-multiple selected (Figure 5). In step 102, the address j in the list I _st and the index m of the submultiple are initialized to 0 and 1, respectively. A comparison 104 is made between the submultiple rbf / m and the minimum delay rmin. The submultiple rbf / m is to be examined if it is greater than rmin. We then take for the integer i the value of the index of the quantized delay r _i closest to rbf / m (step 106), then we compare, in 108, the estimated value of the prediction gain P _st (r _i ) associated with the quantized delay r _i for the sub-frame considered at the selection threshold SE _st calculated in step 98: P st (r i ) = 20.log 10 [RO st / [RO st - VS st 2 (r i ) / G st (r i )]] with, for the fractional delays an interpolation of the values C _st and G _st calculated in step 90 for the whole delays. If P _st (r _i) <SE _st , the delay r _i is not taken into account, and we go directly to step 110 of incrementing the index m before performing the comparison 104 again the next submultiple. If test 108 shows that P _st (r _i ) ≥ SE _st , the delay r _i is retained and step 112 is executed before incrementing the index m in step 110. In step 112, we stores the index at address j in the list I _St , we give the value m to the integer m0 intended to be equal to the index of the smallest submultiple retained, then we increment by one the unit address j.

The examination of the sub-multiples of the basic delay is finished when the comparison 104 shows rbf / m <rmin. We then examine the multiple delays of the smallest rbf / m0 of the submultiples previously selected according to the process illustrated in FIG. 5. This examination begins with an initialization 114 of the index n of the multiple: n = 2. A comparison 116 is made between the multiple n.rbf / m0 and the maximum delay rmax. If n.rbf / m0> rmax, test 118 is carried out to determine whether the index m0 of the smallest sub-multiple is an integer multiple of n. If so, the delay n.rbf / m0 has already been examined when examining the sub-multiples of rbf, and we go directly to step 120 of incrementing the index n before carrying out again comparison 116 for the next multiple. If test 118 shows that m0 is not an integer multiple of n, the multiple n.rbf / m0 is to be examined. We then take for the integer i the value of the index of the quantized delay r _i closest to n.rbf / m0 (step 122), then we compare, at 124, the estimated value of the prediction gain P _st ( r _i ) at the selection threshold SE _st . If P _st (r _i ) <SE _st , the delay r _i is not taken into account, and we go directly to step 120 of incrementing the index n. If test 124 shows that p _st (r _i ) ≥ SE _st , the delay r _i is retained and step 126 is executed before incrementing the index n in step 120. In step 126, we stores the index i at address j in the list I _st , then the address j is incremented by one.

The examination of the multiples of the smallest sub-multiple is finished when the comparison 116 shows that n.rbf / m0> rmax. At this time, the list I _st contains j candidate delay index. If we wish to limit the maximum length of the list I _st to jmax for the following steps, we can take the length j _st of this list equal to min (j, jmax) (step 128) and then, in step 130, order the list I _st in the order of gains C _st ² (r _{Ist (j)} ) / G _st ² (r _{Ist (j)} ⁾ decreasing for 0≤j <j _st so as to keep only the j _st delays providing the largest gain values. The value of jmax is chosen according to the compromise sought between the efficiency of the search for LTP delays and the complexity of this search. Typical values of jmax range from 3 to 5.

puis comparée à Ymax, où Ymax représente la valeur à maximiser. Cette valeur Ymax est par exemple initialisée à 0 en même temps que l'index st à l'étape 96. Si Y≤Ymax, on passe directement à l'étape 140 d'incrémentation de l'index j. Si la comparaison 150 montre que Y > Ymax, on exécute l'étape 152 avant d'incrémenter l'adresse j à l'étape 140. A cette étape 152, l'index ZP est pris égal à I_st(j) et les index ZP0 et ZP1 sont respectivement pris égaux au plus petit et au plus grand des index i_st' déterminés à l'étape 148.Once the sub-multiples and the multiples have been examined and the list I _st has thus been obtained (FIG. 3), the analysis module 36 calculates a quantity Ymax determining a second open-loop estimate of the prediction gain at long term over the entire frame, as well as indexes ZP, ZP0 and ZP1 in a phase 132, the progress of which is detailed in FIG. 6. This phase 132 consists in testing search intervals of length N1 to determine which one maximizes a second estimate of the overall prediction gain on the frame. The intervals tested are those whose centers are the candidate delays contained in the list I _st calculated during phase 101. Phase 132 begins with a step 136 where the address j in the list I _st is initialized to 0. At l 'step 138, it is checked whether the index I _st (j) has already been encountered by testing a previous interval centered on I _st , (j') with st '<st and 0≤j'<j _{st '} , in order d '' Avoid testing the same interval twice. If test 138 reveals that I _st (j) already appeared in a list I _{st '} with st'<st, we directly increment the address j in step 140, then we compare it to the length j _st of the list I _st . If the comparison 142 shows that j <j _st , we return to step 138 for the new value of the address j. When the comparison 142 shows that j = j _st , all the intervals relating to the list I _st have been tested, and phase 132 is terminated. When the test 138 is negative, the interval centered on I _st (j) is tested, starting with step 148 where we determine, for each sub-frame st ', the index i _st , of the optimal delay which maximizes on this interval the open loop estimation P _st (r _i ) of the long-term prediction gain, that is to say which maximizes the quantity Y _st , (i) = C _{st '} ² (r _i ) / G _st , (r _i ) where r _i denotes the quantized delay of index i for I _st (j) -N1 / 2 ≤i <I _st (j) + N1 / 2 and 0≤i <N. During the maximization 148 relating to a sub-frame st ', we a priori discard the indexes i for which the autocorrelation C _st' (r _i ) is negative, to avoid degrading the coding. If it turns out that all the values of i included in the tested interval [I (j) -N1 / 2, I (j) + N1 / 2 [give rise to negative autocorrelations C _{st '} (r _i ), selecting the index i _{st '} for which this autocorrelation is the smallest in absolute value. Then, in 150, the quantity Y determining the second estimate of the overall prediction gain for the interval centered on I _st (j) is calculated according to:

then compared to Ymax, where Ymax represents the value to be maximized. This value Ymax is for example initialized to 0 at the same time as the index st in step 96. If Y≤Ymax, we go directly to step 140 for incrementing the index j. If the comparison 150 shows that Y> Ymax, step 152 is executed before incrementing the address j in step 140. At this step 152, the index ZP is taken equal to I _st (j) and the indices ZP0 and ZP1 are respectively taken equal to the smallest and the largest of the indices i _{st '} determined in step 148.

At the end of phase 132 relating to a subframe st, the index st is incremented by one (step 154) then compared, in step 156, to the number nst of subframes per frame. If st <nst, we return to step 98 to perform the operations relating to the following sub-frame. When the comparison 156 shows that st = nst, the index ZP denotes the center of the search interval that will be provided to the module 38 closed loop LTP analysis, and ZP0 and ZP1 are index whose difference is representative of the dispersion of optimal delays per subframe in the interval centered on ZP.

In step 158, the module 36 determines the degree of voicing MV, on the basis of the second open-loop estimate of the gain expressed in decibels: Gp = 20.log ₁₀ (RO / RO-Ymax). Two other thresholds S1 and S2 are used. If Gp≤S1, the degree of voicing MV is taken equal to 1 for the current frame. The threshold S1 is typically between 3 and 5 dB; for example S1 = 4 dB. If S1 <Gp <S2, the degree of voicing MV is taken equal to 2 for the current frame. The threshold S2 is typically between 5 and 8 dB; for example S2 = 7 dB. If Gp> S2, the dispersion of the optimal delays for the different sub-frames of the current frame is examined. If ZP1-ZP <N3 / 2 and ZP-ZP0≤N3 / 2, an interval of length N3 centered on ZP is sufficient to take into account all the optimal delays and the degree of voicing is taken equal to 3 (if Gp> S2) . Otherwise, if ZP1-ZP≥N3 / 2 or ZP-ZPO> N3 / 2, the degree of voicing is taken equal to 2 (if Gp> S2).

The ZP index of the center of the search interval of prediction delay for a voiced frame can be understood between 0 and N-1 = 255, and the differential index DP determined for module 38 can range from -16 to +15 if MV = 1 or 2, and from -8 to +7 if MV = 3 (case N1 = 32, N3 = 16). ZP + DP index of TP delay ultimately determined may therefore in some cases be more small than 0 or larger than 255. This allows analysis LTP in closed loop to also carry on some delays TP smaller than rmin or larger than rmax. We improve thus the subjective quality of the so-called voice reproduction pathological and non-vocal signals (vocal frequencies DTMF or signaling frequencies used by the network dial-up). Another possibility is to take for the search interval the first 32 or last delay quantification index if ZP <16 or ZP> 240 with MV = 1 or 2, and the first 16 or last index if ZP <8 or ZP> 248 with MV = 3.

Reducing the delay search interval for very closely spaced frames (typically 16 values for MV = 3 instead of 32 for MV = 1 or 2) reduces the complexity of the closed loop LTP analysis performed by the module 38 by reducing the number of convolutions y _T (i) to be calculated according to formula (1). Another advantage is that a coding bit of the differential DP index is saved. Since the output rate is constant, this bit can be reallocated for coding other parameters. We can in particular allocate this additional bit to the quantification of the long-term prediction gain g _P calculated by the module 40. Indeed, better precision on the gain g _P thanks to an additional quantization bit is appreciable because this parameter is perceptually important for very voiced subframes (MV = 3). Another possibility is to provide a parity bit for the delay TP and / or the gain g _P , making it possible to detect possible errors affecting these parameters.

It is possible to make some modifications to the open loop LTP analysis process described above with reference to Figures 3 to 6.

According to a first variant of this process, the first optimizations carried out in step 90 relative to the different sub-frames are replaced by a single optimization relating to the entire frame. In addition to the parameters C _st (k) and G _st (k) calculated for each sub-frame st, the autocorrelations C (k) and the delayed energies G (k) for the entire frame are also calculated:

We then determine the basic delay in whole resolution K which maximizes X (k) = C ² (k) / G (k) for rmin ≤ k ≤ rmax. The first gain estimate compared to S0 in step 92 is then P (K) = 20.log ₁₀ [R0 / [R0-X (K)]]. One then determines, around K, a single basic delay in fractional resolution rbf and the examination 101 of the sub-multiples and of the multiples is carried out only once and produces a single list I instead of nst lists I _st . Phase 132 is then performed only once for this list I, distinguishing the subframes only in steps 148, 150 and 152. This variant embodiment has the advantage of reducing the complexity of the analysis in open loop.

According to a second variant of the open loop LTP analysis process, the domain [rmin, rmax] of possible delays is subdivided into nz sub-intervals having for example the same length (nz = 3 typically), and the first optimizations carried out at step 90 relative to the different subframes are replaced by nz optimizations in the different subintervals each relating to the entire frame. We thus obtain nz basic delays K ₁ ', ..., K _nz ' in full resolution. The voiced / unvoiced decision (step 92) is taken on the basis of that of the basic delays K _i 'which provides the greatest value for the first open-loop estimate of the long-term prediction gain. Then, if the frame is voiced, the basic delays in fractional resolution are determined by the same process as in step 100, but only allowing the quantized delay values. Examination 101 of the sub-multiples and multiples is not performed. For the phase 132 of calculating the second estimate of the prediction gain, the nz basic delays previously determined are taken as candidate delays. This second variant makes it possible to dispense with the systematic examination of the submultiples and of the multiples which are generally taken into account by virtue of the subdivision of the domain of possible delays.

According to a third variant of the LTP open-loop analysis process, phase 132 is modified in that, in the optimization steps 148, the index i _st , which maximizes C _{st '} ² (r _i ) / G _{st '} (r _i ) for I _st (j) -N1 / 2≤i <I _st (j) + N1 / 2 and 0≤i <N, and on the other hand, during the same maximization loop, the index k _st , which maximizes this same quantity over a reduced interval I _st (j) -N3 / 2≤ i <I _st (j) + N3 / 2 and 0≤1 <N. Step 152 is also modified: the indexes ZP0 and ZP1 are no longer stored, but a quantity Ymax 'defined in the same way as Ymax but with reference to the reduced length interval:

In this third variant, the determination 158 of the voicing mode leads to more often selecting the voicing degree MV = 3. We also take into account, in addition to the gain Gp previously described, a third open loop estimate of the gain LTP, corresponding to Ymax ': Gp' = 20.log ₁₀ [R0 / (R0-Ymax ')]. The degree of voicing is MV = 1 if Gp≤S1, MV = 3 if Gp '> S2 and MV = 2 if neither of these two conditions is verified. By thus increasing the proportion of frames of degree MV = 3, the average complexity of the closed loop analysis is reduced and the robustness to transmission errors is improved.

A fourth variant of the open loop LTP analysis process mainly concerns weakly voiced frames (MV = 1). These frames often correspond to a start or an end of a voicing area. Frequently, these frames can comprise from one to three sub-frames for which the gain coefficient of the long-term synthesis filter is zero or even negative. It is proposed not to perform LTP analysis in closed loop for the sub-frames in question, in order to reduce the average complexity of the coding. This can be achieved by storing in step 152 of FIG. 6 nst pointers indicating for each sub-frame st 'whether the autocorrelation C _st , corresponding to the index delay i _st' is negative or even very small. Once all the intervals referenced in the lists I _{st ′} the sub-frames for which the prediction gain is negative or negligible can be identified by consulting the nst pointers. If necessary, the module 38 is deactivated for the corresponding sub-frames. This does not affect the quality of the LTP analysis since the prediction gain corresponding to these subframes will be almost zero anyway.

Another aspect of the invention relates to the module 42 for calculating the impulse response of the weighted synthesis filter. The closed loop LTP analysis module 38 needs this impulse response h over the duration of a subframe to calculate the convolutions y _T (i) according to formula (1). The stochastic analysis module 40 also needs it to calculate convolutions as will be seen below. The fact of having to calculate convolutions with a response h extending over the duration of a subframe (lst = 40 typically) implies a relative complexity of the coding, which it would be desirable to reduce in particular to increase the autonomy of the mobile station. In some cases it has been proposed to truncate the impulse response to a length less than the length of a sub-frame (for example to 20 samples), but this can degrade the quality of the coding. It is proposed according to the invention to truncate the impulse response h taking into account on the one hand the energy distribution of this response and on the other hand the degree of voicing MV of the frame considered, determined by the LTP analysis module 36 open loop.

The operations performed by the module 42 are for example in accordance with the flowchart of FIG. 7. The impulse response is first calculated in step 160 over a length pst greater than the length of a subframe and sufficiently large to be sure to take into account all the energy of the impulse response (for example pst = 60 for nst = 4 and lst = 40 if the short-term linear prediction is of order q = 10). In step 160, the truncated energies of the impulse response are also calculated:

Eh(i) = Eh(i-1) + [h(i)]2 pour 0<i<pst, avec f(i)=h(i)=0 pour i<0, δ(0)=f(0)=h(0) =Eh(0)=1, et δ(i)=0 pour i≠0. Dans l'expression (2), les coefficients a_k sont ceux intervenant dans le filtre de pondération perceptuelle, c'est-à-dire les coefficients de prédiction linéaire interpolés mais non quantifiés, tandis que dans l'expression (3), les coefficients a_k sont ceux appliqués au filtre de synthèse, c'est-à-dire les coefficients de prédiction linéaire quantifiés et interpolés.The components h (i) of the impulse response and the truncated energies Eh (i) can be obtained by filtering a unitary pulse by means of a transfer function filter W (z) / A (z) of zero initial states , or by recurrence:

Eh (i) = Eh (i-1) + [h (i)] 2 for 0 <i <pst, with f (i) = h (i) = 0 for i <0, δ (0) = f (0) = h (0) = Eh (0) = 1, and δ (i ) = 0 for i ≠ 0. In expression (2), the coefficients a _k are those involved in the perceptual weighting filter, i.e. the linear prediction coefficients interpolated but not quantified, while in expression (3), the coefficients a _k are those applied to the synthesis filter, i.e. the quantized and interpolated linear prediction coefficients.

Then the module 42 determines the shortest length Lα such that the energy Eh (Lα-1) of the impulse response truncated at Lα samples is at least equal to a proportion α of its total energy Eh (pst-1) estimated over pst samples. A typical value of α is 98%. The number Lα is initialized to pst in step 162 and decremented by unit as 166 as Eh (Lα-2)> α.Eh (pst-1) (test 164). The length Lα sought is obtained when test 164 shows that Eh (Lα-2) ≤α.Eh (pst-1).

To take into account the degree of voicing MV, a term corrector Δ (MV) is added to the value of Lα which has been obtained (step 168). This corrector term is preferably an increasing function of the degree of voicing. We can by example take Δ (0) = - 5, Δ (1) = 0, Δ (2) = + 5 and Δ (3) = + 7. Of this way, the impulse response h will be determined so all the more precise as the voicing of speech is important. The truncation length Lh of the impulse response is taken equal to Lα if Lα≤nst and to nst otherwise. The remaining samples of the impulse response (h (i) = 0 with i≥Lh) can be canceled.

With the truncation of the impulse response, the calculation (1) of the convolutions y _T (i) by the module 38 of LTP analysis in closed loop is modified as follows:

Obtaining these convolutions, which represents a significant part of the calculations performed, therefore requires significantly less multiplication, addition and addressing in the adaptive repertoire when the impulse response is truncated. The dynamic truncation of the impulse response involving the degree of voicing MV makes it possible to obtain such a reduction in complexity without affecting the quality of the coding. The same considerations apply for the convolution calculations performed by the module 40 of stochastic analysis. These advantages are particularly appreciable when the perceptual weighting filter has a transfer function of the form W (z) = A (z / γ ₁ ) / A (z / γ ₂ ) with 0 <γ ₂ <γ ₁ <1 which gives rise to impulse responses generally longer than those of the form W (z) = A (z) / A (z / γ) more commonly used in coders using analysis by synthesis.

A third aspect of the invention relates to the module 40 of stochastic analysis used to model the unpredictable part of the excitement.

The stochastic excitation considered here is of the multi-pulse type. The stochastic excitation relating to a subframe is represented by np pulses of positions p (n) and of amplitudes, or gains, g (n) (1≤n≤np). The gain g _P of long-term prediction can also be calculated during the same process. In general, it can be considered that the excitation sequence relating to a sub-frame comprises nc contributions associated respectively with nc gains. The contributions are lst sample vectors which, weighted by the associated and summed gains correspond to the excitation sequence of the short-term synthesis filter. One of the contributions can be predictable, or several in the case of a long-term synthesis filter with several takes ("multi-tap pitch synthesis filter"). The other contributions are in the present case np vectors comprising only 0 except an impulse of amplitude 1. We therefore have nc = np if MV = 0, and nc = np + 1 if MV = 1, 2 or 3.

les gains étant solution du système linéaire g.B=b.The multi-pulse analysis including the calculation of the gain g _P = g (0) consists, in a known manner, of finding for each sub-frame positions p (n) (1≤n≤np) and gains g (n ) (0≤n≤np) which minimize the perceptually weighted quadratic error E between the speech signal and the synthesized signal, given by:

the gains being solution of the linear system gB = b.

• X désigne un vecteur-cible initial composé des lst échantillons du signal de parole pondéré SW sans mémoire : X=(x(0),x(1), ..., x(lst-1)), les x(i) ayant été calculés comme indiqué précédemment lors de l'analyse LTP en boucle fermée ;
• g désigne le vecteur ligne composé des np+1 gains : g= (g(0)=g_P, g(1), ...,g(np)) ;
• les vecteurs-ligne F_p(n) (0≤n<nc) sont des contributions pondérées ayant pour composantes i(0≤i<lst) les produits de convolution entre la contribution n à la séquence d'excitation et la réponse impulsionnelle h du filtre de synthèse pondéré ;
• b désigne le vecteur ligne composé des nc produits scalaires entre le vecteur X et les vecteurs ligne F_p(n);
• B désigne une matrice symétrique à nc lignes et nc colonnes dont le terme B_i,j=F_p(i).F_p(j).^T (0≤i,j<nc) est égal au produit scalaire entre les vecteurs F_p(i) et F_p(j) précédemment définis ;
• (.)^T désigne la transposition matricielle.

In the above notations:

• X denotes an initial target vector composed of the lst samples of the weighted speech signal SW without memory: X = (x (0), x (1), ..., x (lst-1)), the x (i) having been calculated as indicated previously during the closed loop LTP analysis;
• g denotes the line vector composed of the np + 1 gains: g = (g (0) = g _P , g (1), ..., g (np));
• the line vectors F _p (n) (0≤n <nc) are weighted contributions having as components i (0≤i <lst) the products of convolution between the contribution n to the excitation sequence and the impulse response h weighted synthesis filter;
• b denotes the line vector composed of the nc scalar products between the vector X and the line vectors F _{p (n)} ;
• B denotes a symmetric matrix with nc rows and nc columns whose term B _i , j = F _{p (i)} .F _{p (j)} . ^T (0≤i, j <nc) is equal to the scalar product between the vectors F _{p (i)} and F _{p (j)} previously defined;
• (.) ^T denotes the matrix transposition.

For the stochastic excitation pulses (1≤n ≤np = nc-1) the vectors F _{p (n)} are simply constituted by the vector of the impulse response h shifted by p (n) samples. Truncating the impulse response as described above therefore makes it possible to significantly reduce the number of operations useful for calculating the scalar products involving these vectors F _{p (n)} . For the predictable contribution of the excitation, the vector F _{p (0)} = Y _TP has for components F _{p (0)} (i) (0≤i <lst) the convolutions Y _TP (i) that the module 38 has calculated according to formula (1) or (1 ') for the long-term prediction delay selected TP. If MV = 0, the contribution n = 0 is also of impulse type and the position p (0) is to be calculated.

Minimizing the quadratic error E defined above amounts to finding the set of positions p (n) which maximize the normalized correlation bB ^-1 .b ^T then calculating the gains according to g = bB ^-1 .

But an exhaustive search of the pulse positions would require an excessive volume of calculations. To alleviate this problem, the multi-pulse approach generally applies a sub-optimal procedure consisting of successively calculating the gains and / or the pulse positions for each contribution. For each contribution n (0≤n <nc), we first determine the position p (n) which maximizes the normalized correlation (F _p .e _n-1 ^T ) ² / (F _p .F _p ^T ), we recalculates the gains g _n (0) to g _n (n) according to g _n = b _n .B _n ^-1 , where g _n = (g _n (0) g _n (n)), b _n = (b (0 ), ..., b (n)) and B _n = {B _{i, j} } _{0≤i, j≤n} , then for the next iteration we calculate the target vector e _n equal to the initial target vector X to which we subtract the contributions 0 to n of the weighted synthetic signal multiplied by their respective gains:

At the end of the last iteration nc-1, the gains g _nc-1 (i) are the selected gains and the minimized quadratic error E is equal to the energy of the target vector e _nc-1 .

The above method gives satisfactory results, but it requires the inversion of a matrix B _n at each iteration. In their article "Amplitude Optimization and Pitch Prediction in Multipulse Coders" (IEEE Trans. On Acoustics, Speech, and Signal Processing, Vol.37, N ° 3, March 1989, pages 317-327), S. Singhal and BS Atal have proposed to simplify the problem of inverting the matrices B _n using the Cholesky decomposition: B _n = M _n .M _n ^T where M _n is a lower triangular matrix. This decomposition is possible because B _n is a symmetric matrix with positive eigenvalues. The advantage of this approach is that the inversion of a triangular matrix is relatively uncomplicated, B _n ^-1 can be obtained by B _n ^-1 = (M _n ^-1 ) ^T .M _n ^-1 .

The decomposition of Cholesky and the inversion of the matrix M _n however require to carry out divisions and calculations of square roots which are operations demanding in terms of computation complexity. The invention proposes to considerably simplify the implementation of the optimization by modifying the decomposition of the matrices B _n as follows: B not = L not .R not T = L not . (L not .K not -1 ) T where K _n is a diagonal matrix and L _n is a lower triangular matrix having only 1s on its main diagonal (ie L _n = M _n .K _n ^1/2 with the previous notations). Given the structure of the matrix B _n , the matrices L _n = R _n .K _n , R _n , K _n and L _n ^-1 are each constructed by simply adding a line to the corresponding matrices of the iteration former :

Under these conditions, the decomposition of B _n , the inversion of L _n , the obtaining of B _n ^-1 = K _n . (L _n ^-1 ) ^T .L _n ^-1 and the recalculation of the gains only require only one division per iteration and no square root calculation.

où e=(e(0), ..., e(lst-1)) est un vecteur-cible calculé lors de l'itération précédente. Différentes contraintes peuvent être apportées au domaine de maximisation de la quantité ci-dessus inclus dans l'intervalle [0,lst[. L'invention utilise de préférence une recherche segmentaire dans laquelle la sous-trame d'excitation est subdivisée en ns segments de même longueur (par exemple ns=10 pour lst=40). Pour la première impulsion (n=1), la maximisation de (F_p.e^T)²/(F_p.F_p ^T) est effectuée sur l'ensemble des positions possibles p dans la sous-trame. A l'itération n>1, la maximisation est effectuée à l'étape 182 sur l'ensemble des positions possibles à l'exclusion des segments dans lesquels ont été respectivement trouvées les positions p(1),...,p(n-1) des impulsions lors des itérations précédentes.The stochastic analysis relating to a subframe of a voiced frame (MV = 1,2 or 3) can therefore take place as indicated in FIGS. 8 to 11. To calculate the long-term prediction gain, the contribution index n is initialized to 0 in step 180 and the vector F _{p (0)} is taken equal to the long-term contribution Y _TP provided by the module 38. If n> 0, the iteration n begins with the determination 182 of the position p (n) of the pulse n which maximizes the quantity

where e = (e (0), ..., e (lst-1)) is a target vector calculated during the previous iteration. Different constraints can be brought to the domain of maximization of the quantity above included in the interval [0, lst [. The invention preferably uses a segmental search in which the excitation subframe is subdivided into ns segments of the same length (for example ns = 10 for lst = 40). For the first pulse (n = 1), the maximization of (F _p. E ^T ) ² / (F _p. F _p ^T ) is performed on all the possible positions p in the sub-frame. At iteration n> 1, the maximization is carried out in step 182 on the set of possible positions excluding the segments in which the positions p (1), ..., p (n have been found respectively) -1) pulses during previous iterations.

In the case where the current frame has been detected as unvoiced, the contribution n = 0 also consists of a position pulse p (0). Step 180 then only includes initialization n = 0, and it is followed by a maximization step identical to step 182 to find p (0), with e = e _-1 = X as the initial value of the vector- target.

Note that when the contribution n = 0 is predictable (MV = 1, 2 or 3), the closed-loop LTP analysis module 38 has performed an operation of a nature similar to the maximization 182, since it has determined the contribution in the long term, characterized by the delay TP, by maximizing the quantity (Y _T .e ^T ) ² / (Y _T .Y _T ^T ) in the delay search interval T, with e = e _-1 = X as initial value of the target vector. We can also, when the energy of the LTP contribution is very low, ignore this contribution in the process of recalculation of the gains.

pour la composante située à la ligne n et à la colonne j. On peut donc écrire, pour j croissant de 0 à n-1 :

L(n,j) = R(n,j).K(j) et, pour j=n :

L(n,n) = 1 After step 180 or 182, the module 40 proceeds to the calculation 184 of the line n of the matrices L, R and K involved in the decomposition of the matrix B, which makes it possible to complete the matrices L _n , R _n and K _n defined above. The decomposition of the matrix B makes it possible to write:

for the component located in row n and in column j. We can therefore write, for j increasing from 0 to n-1:

L (n, j) = R (n, j) .K (j) and, for j = n:

L (n, n) = 1

These relationships are used in the calculation 184 detailed in FIG. 9. The column index j is first initialized at 0, in step 186. For the column index j, the variable tmp is first initialized at the value of component B (n, j), that is:

In step 188, the integer k is also initialized to 0. A comparison 190 is then made between the integers k and j. If k <j, we add the term L (n, k). R (j, k) to the variable tmp, then we increment the whole k by one unit (step 192) before re-performing the comparison 190. When comparison 190 shows that k = j, a comparison 194 is carried out between the integers j and n. If j <n, the component R (n, j) is taken equal to tmp and the component L (n, j) to tmp.K (j) in step 196, then the column index j is incremented d 'a unit before returning to step 188 to calculate the following components. When the comparison 194 shows that j = n, the component K (n) of the line n of the matrix K is calculated, which ends the calculation 184 relating to the line n. K (n) is taken equal to 1 / tmp if tmp ≠ 0 (step 198) and to 0 otherwise. We note that the calculation 184 requires at most one division 198, to obtain K (n). In addition, any singularity of the matrix B _n does not cause instabilities since we avoid divisions by 0.

pour 0≤j'<n et L^-1(n,n)=1, c'est-à-dire que l'inversion peut être faite sans avoir à opérer une division. En outre, comme les composantes de la ligne n de L^-1 suffisent à recalculer les gains, l'utilisation de la relation (5) permet de faire l'inversion sans avoir à mémoriser toute la matrice L^-1, mais seulement un vecteur Linv=(Linv(0),...,Linv(n-1)) avec Linv(j')=L^-1(n,j'). L'inversion 200 commence alors par une initialisation 202 de l'index de colonne j' à n-1. A l'étape 204, le terme Linv(j') est initialisé à -L(n,j') et l'entier k' à j'+1. On effectue ensuite une comparaison 206 entre les entiers k' et n. Si k'<n, on retranche le terme L(k',j').Linv(k') à Linv(j'), puis on incrémente d'une unité l'entier k' (étape 208) avant de réexécuter la comparaison 206. Quand la comparaison 206 montre que k'=n, on compare j' à 0 (test 210). Si j'>0, on décrémente l'entier j' d'une unité (étape 212) et on revient à l'étape 204 pour calculer la composante suivante. L'inversion 200 est terminée lorsque le test 210 montre que j'=0.With reference to FIG. 8, the calculation 184 of the lines n of L, R and K is followed by the inversion. 200 of the matrix L _n consisting of rows and columns 0 to n of the matrix L. The fact that L is triangular with 1s on its main diagonal greatly simplifies the inversion as shown in Figure 10. We can indeed to write :

for 0≤j '<n and L ^-1 (n, n) = 1, that is to say that the inversion can be done without having to make a division. In addition, since the components of the line n of L ^-1 are sufficient to recalculate the gains, the use of the relation (5) makes it possible to make the inversion without having to memorize the entire matrix L ^-1 , but only one vector Linv = (Linv (0), ..., Linv (n-1)) with Linv (j ') = L ^-1 (n, j'). The inversion 200 then begins with an initialization 202 of the column index j 'at n-1. In step 204, the term Linv (j ') is initialized to -L (n, j') and the integer k 'to j' + 1. A comparison 206 is then carried out between the integers k ′ and n. If k '<n, we subtract the term L (k', j '). Linv (k') to Linv (j '), then we increment the whole k' by one unit (step 208) before re-executing comparison 206. When comparison 206 shows that k '= n, we compare j' to 0 (test 210). If j '> 0, the whole j' is decremented by one unit (step 212) and we return to step 204 to calculate the next component. The inversion 200 is finished when the test 210 shows that j '= 0.

et g_n(i')=g_n-1(i')+L^-1(n,i').g_n(n) pour 0≤i'<n. Le calcul 214 est détaillé sur la figure 11. On calcule d'abord la composante b(n) du vecteur b :

b(n) sert de valeur d'initialisation pour la variable tmq. A l'étape 216, on initialise également l'index i à 0. On effectue ensuite la comparaison 218 entre les entiers i et n. Si i<n, on ajoute le terme b(i). Linv(i) à la variable tmq et on incrémente i d'une unité (étape 220) avant de revenir à la comparaison 218. Quand la comparaison 218 montre que i=n, on calcule le gain relatif à la contribution n selon g(n)=tmq.K(n), et on initialise la boucle de calcul des autres gains et du vecteur-cible (étape 222) en prenant e=X-g(n).F_p(n) et i'=0. Cette boucle comprend une comparaison 224 entre les entiers i' et n. Si i'<n, le gain g(i') est recalculé à l'étape 226 en ajoutant Linv(i').g(n) à sa valeur calculée lors de l'itération précédente n-1, puis on retranche au vecteur-cible e le vecteur g(i').F_p(i'). L'étape 226 comprend également l'incrémentation de l'index i' avant de revenir à la comparaison 224. Le calcul 214 des gains et du vecteur-cible est terminé lorsque la comparaison 224 montre que i'=n. On voit que les gains ont pu être mis à jour en ne faisant appel qu'à la ligne n de la matrice inverse L_n ^-1 .With reference to FIG. 8, the inversion 200 is followed by the calculation 214 of the reoptimized gains and of the target vector E for the following iteration. The computation of the reoptimized gains is also very simplified by the decomposition retained for the matrix B. One can indeed compute the vector g _n = (g _n (0), ..., g _n (n)) solution of g _n . B _n = b _n according to:

and g _n (i ') = g _n-1 (i') + L ^-1 (n, i '). g _n (n) for 0≤i'<n. The calculation 214 is detailed in FIG. 11. The component b (n) of the vector b is first calculated:

b (n) serves as the initialization value for the variable tmq. In step 216, the index i is also initialized to 0. The comparison 218 is then carried out between the integers i and n. If i <n, we add the term b (i). Linv (i) to the variable tmq and we increment i by one unit (step 220) before returning to the comparison 218. When the comparison 218 shows that i = n, we calculate the gain relative to the contribution n according to g ( n) = tmq.K (n), and the loop for calculating the other gains and the target vector is initialized (step 222) by taking e = Xg (n) .F _{p (n)} and i '= 0. This loop includes a comparison 224 between the integers i 'and n. If i '<n, the gain g (i') is recalculated in step 226 by adding Linv (i '). G (n) to its value calculated during the previous iteration n-1, then we subtract from target vector e the vector g (i '). F _{p (i')} . Step 226 also includes the incrementation of the index i 'before returning to the comparison 224. The calculation 214 of the gains and of the target vector is finished when the comparison 224 shows that i' = n. We see that the gains could be updated by using only the line n of the inverse matrix L _n ^-1 .

The calculation 214 is followed by an increment 228 of the index n of the contribution, then by a comparison 230 between index n and the number of contributions nc. If n <nc, we return to step 182 for the next iteration. Optimization of positions and gains is complete when n = nc in test 230.

Segmental pulse search significantly decreases the number of pulse positions to be evaluated during steps 182 of the search for stochastic excitation. It also allows efficient quantification of the positions found. In the typical case where the subframe of lst = 40 samples is divided into ns = 10 segments of ls = 4 samples, the set of possible pulse positions can take ns! .Ls ^np / [np! (Ns- np)!] = 258,048 values if np = 5 (Mv = 1, 2 or 3) or 860,160 if np = 6 (MV = 0), instead of lst! / [np! (lst-np)!] = 658 008 values if np = 5 or 3 838 380 if np = 6 in the case where it is imposed only that two pulses cannot have the same position. In other words, we can quantify the positions on 18 bits instead of 20 bits if np = 5, and on 20 bits instead of 22 if np = 6.

The particular case where the number of segments per subframe is equal to the number of pulses by stochastic excitation (ns = np) leads to the greatest simplicity of the search for stochastic excitation, as well as to the lowest bit rate (if lst = 40 and np = 5, there are 8 ⁵ = 32 768 sets of possible positions, quantifiable on 15 bits only instead of 18 if ns = 10). But by reducing the number of possible innovation sequences to this point, the quality of the coding can be reduced. For a given number of pulses, the number of segments can be optimized according to a targeted compromise between the quality of the coding and its simplicity of implementation (as well as the required bit rate).

The case where ns> np also has the advantage that good robustness to transmission errors can be obtained with regard to the positions of the pulses, by virtue of a separate quantification of the sequence numbers of the occupied segments and of the relative positions pulses in each occupied segment. For a pulse n, the sequence number s _n of the segment and the relative position pr _n are respectively the quotient and the remainder of the Euclidean division of p (n) by the length ls of a segment: p (n) = s _n .ls + pr _n (0≤s _n <ns, 0≤pr _n <ls). The relative positions are each quantized separately on 2 bits, if ls = 4. In the event of a transmission error affecting one of these bits, the corresponding pulse will be only slightly displaced, and the perceptual impact of the error will be limited. The serial numbers of the occupied segments are identified by a binary word of ns = 10 bits each equal to 1 for the occupied segments and 0 for the segments in which the stochastic excitation has no pulse. Possible binary words are those with a Hamming weight of np; they are ns! / [np! (ns-np)!] = 252 if np = 5, or 210 if np = 6. This word is quantifiable by an index of nb bits with 2 ^nb-1 <ns! / [Np! (Ns-np)!] ≤2 ^nb , that is nb = 8 in the example considered. If, for example, the stochastic analysis provided np = 5 pulses of positions 4, 12, 21, 34, 38, the relative positions quantified scalarly are 0,0,1,2,2 and the binary word representing the occupied segments is 0101010011, or 339 in decimal translation.

At the decoder, the possible binary words are stored in a quantification table in which the reading addresses are the quantization indexes received. The order in this table, determined once for all, can be optimized so that an error of transmission affecting a bit of the index (the error case the more frequent, especially when interlacing is used work in the channel encoder 22) has, on average, minimal consequences according to a neighborhood criterion. The neighborhood criterion is for example that a word of ns bits does not can be replaced only by words "neighbors", distant a Hamming distance at most equal to an np-2δ threshold, so as to keep all the pulses except δ of them at valid positions in case of transmission error the single-bit index. Other criteria would be usable in substitution or in addition, for example that two words are considered neighbors if the replacement of one by the other does not change the order of assignment of gains associated with pulses.

For illustration purposes, we can consider the case simplified where ns = 4 and np = 2, i.e. 6 possible binary words quantifiable on nb = 3 bits. In this case, we can check that the quantification table presented in Table II allows to keep np-1 = 1 pulse well positioned for all error affecting a bit of the transmitted index. There are 4 cases error (out of a total of 18), for which we receive a quantification index that we know to be wrong (6 instead of 2 or 4; 7 instead of 3 or 5), but the decoder can then take measures to limit distortion, for example repeat the innovation sequence relating to the subframe previous or assign acceptable binary words "impossible" indexes (for example 1001 or 1010 for index 6 and 1100 or 0110 for index 7 still drive at np-1 = 1 well-positioned pulse when receiving 6 or 7 with a binary error).

In the general case, the order in the table word quantification can be determined from arithmetic considerations or, if this is insufficient, in simulating error scenarios on a computer (so exhaustive or by statistical sampling of the type Monte-Carlo according to the number of possible error cases).

To secure the transmission of the quantization index of the occupied segments, one can also take advantage of the different protection categories offered by the channel encoder 22, in particular if the neighborhood criterion cannot be satisfactorily verified for all cases d 'possible errors affecting a bit of the index. The scheduling module 46 can thus put in the minimum protection category, or in the unprotected category, a certain number nx of the bits of the index which, if affected by a transmission error, give rise to a wrong word but checking the neighborhood criterion with a probability deemed satisfactory, and putting the other bits of the index in a more protected category. This procedure calls for a different ordering of the words in the quantification table. This scheduling can also be optimized by means of simulations if it is desired to maximize the number nx of the bits of the index assigned to the least protected category. quantification index word of occupation of the segments decimal natural binary natural binary decimal 0 000 0011 3 1 001 0101 5 2 010 1001 9 3 011 1100 12 4 100 1010 10 5 101 0110 6 (6) (110) (1001 or 1010) (9 or 10) (7) (111) (1100 or 0110) (12 or 6)

One possibility is to start by constituting a list of words of ns bits by counting in Gray code from 0 to 2nS-1, and to obtain the ordered quantification table by deleting from this list the words not having a weight of Hamming of NP. The table thus obtained is such that two consecutive words have a Hamming distance of np-2. If the indexes in this table have a binary representation in Gray code, any error on the least significant bit causes the index to vary by ± 1 and therefore causes the replacement of the actual occupancy word by a neighboring word in the sense of the np-2 threshold on the Hamming distance, and an error on the i-th least significant bit also varies the index by ± 1 with a probability of approximately 2 ^1-i . By placing the nx least significant bits of the index in Gray code in an unprotected category, a possible transmission error affecting one of these bits leads to the replacement of the busy word by a neighboring word with a probability at least equal. to (1 + 1/2 + ... + 1/2 ^nx-1 ) / nx. This minimum probability decreases from 1 to (2 / nb) (1-1 / 2 ^nb ) for nx increasing from 1 to nb. The errors affecting the nb-nx most significant bits of the index will most often be corrected thanks to the protection applied to them by the channel coder. The value of nx is in this case chosen according to a compromise between robustness to errors (small values) and a reduced size of the protected categories (large values).

At the encoder level, the possible binary words for represent the occupation of the segments are arranged in order growing in a search table. An indexing table associates with each address the serial number, in the table of quantization stored at the decoder, of the binary word having this address in the lookup table. In the example simplified mentioned above, the content of the table search and index table is given in the table III (in decimal values).

The quantification of the occupation word of the segments deduced from the np positions provided by the stochastic analysis module 40 is carried out in two stages by the quantization module 44. A dichotomous search is first carried out in the search table to determine the address in this table of the word to be quantified. The quantization index is then obtained at the address determined in the indexing table and then supplied to the bit scheduling module 46. Address Search table Indexing table 0 3 0 1 5 1 2 6 5 3 9 2 4 10 4 5 12 3

The module 44 also performs the quantification of the gains calculated by the module 40. The gain g _TP is for example quantified in the interval [0; 1.6], on 5 bits if MV = 1 or 2 and on 6 bits if MV = 3 to take into account the greater perceptual importance of this parameter for very close frames. For the coding of the gains associated with the pulses of the stochastic excitation, the greatest absolute value Gs of the gains g (1), ..., g (np) is quantified over 5 bits, for example by taking 32 quantization values in geometric progression in the interval [0; 32767], and we quantify each of the relative gains g (1) / Gs, ..., g (np) / Gs in the interval [-1; +1], over 4 bits if MV = 1, 2 or 3, or on 5 bits if MV = 0.

The quantization bits of Gs are placed in a category protected by the channel 22 encoder, as well as most significant bits of the gain quantification indexes relative. The relative gain quantization bits are ordered to allow assignment to impulses associated belonging to the segments localized by the word of occupation. Segmental research according to the invention also allows effective protection of positions relative pulses associated with the largest values gain.

In the case where np = 5 and ls = 4, ten bits per sub-frame are needed to quantify the relative positions of pulses in the segments. We consider the case where 5 of these 10 bits are placed in a category with little or no protection (II) and where the 5 others are placed in a more category protected (IB). The most natural distribution is place the most significant bit of each relative position in protected category IB, so that any transmission errors rather affect the weight bits strong and therefore only cause a shift of a sample for the corresponding pulse. It’s wise, however, for the quantification of relative positions, to consider pulses in descending order of absolute values associated gains and to place in the IB category both quantization bits of each of the first two relative positions as well as the most significant bit of the third. In this way, the positions of the pulses are preferentially protected when associated with significant gains, which improves the average quality especially for the most voiced subframes.

To reconstruct impulse contributions of excitation, the decoder 54 first locates the segments by means of the occupation word received; he then assigns the associated earnings; then he assigns the positions relative to pulses based on the order of importance of the gains.

It will be understood that the different aspects of the invention described above provide each of the own improvements, and that it is therefore possible to envisage them implement independently of each other. Their combination allows for a performance encoder particularly interesting.

In the embodiment described in what precedes, the 13 kbit / s speech coder requires order 15 million comma instructions per second (Mips) fixed. So we typically do this by programming a commercial digital signal processor (DSP) as well as the decoder which requires only about 5 Mips.

Claims (5)

1. Analysis-by-synthesis speech coding method for coding a speech signal digitised into successive frames which are subdivided into sub-frames including a defined number of samples, wherein a linear prediction analysis of the speech signal is performed for each frame in order to determine the coefficients of a short-term synthesis filter (60), and an open-loop analysis is performed for each frame in order to determine a degree of voicing of the frame, and at least one closed-loop analysis is performed for each sub-frame in order to determine an excitation sequence which, submitted to the short-term synthesis filter, produces a synthetic signal representative of the speech signal, each closed-loop analysis using the impulse response of a composite filter consisting of the short-term synthesis filter and of a perceptual weighting filter,
      characterised in that, during each closed-loop analysis, said impulse response is used, truncating it to a truncation length (Lh) equal at most to the number (lst) of samples per sub-frame and dependent on the energy distribution of said response and on the degree of voicing of the frame.
2. Method according to Claim 1, characterised in that the impulse response of the composite filter is calculated over a total length (pst) greater than the number (lst) of samples per sub-frame, in that a minimum length Lα is determined such that the energy of the impulse response calculated by truncating said response to Lα samples is at least equal to a defined fraction (α) of the energy of the impulse response calculated over said total length (pst), and in that the truncation length (Lh) is equal to the sum of said minimum length Lα and a corrector term (Δ(MV)) dependent on the degree of voicing of the frame if said sum is less than the number (lst) of samples per sub-frame.
3. Method according to Claim 2, characterised in that said corrector term (Δ(MV)) is an increasing function of the degree of voicing.
4. Method according to any one of Claims 1 to 3, characterised in that the perceptual weighting filter has a transfer function of the form W(z) = A(z/γ₁)/A(z/γ₂) where 1/A(z) designates the transfer function of the short-term synthesis filter and γ₁ and γ₂ are two coefficients such that 0 < γ₂ < γ₁ < 1.
5. Method according to Claim 4, characterised in that the coefficients of the short-term synthesis filter are represented by line spectrum parameters (LSP), in that said line spectrum parameters are quantified, in that, in order to constitute the short-term synthesis filter to which the excitation sequence relating to at least one sub-frame of a frame is submitted, an interpolation is performed between the line spectrum parameters relating to said frame and those relating to the preceding frame, and in that, in order to calculate the impulse response of the composite filter, the short-term synthesis filter is calculated on the basis of the quantified and interpolated line spectrum parameters, whereas the perceptual weighting filter is calculated on the basis of the interpolated but unquantified line spectrum parameters.

Images