EP1526508A1 - Method for the selection of synthesis units - Google Patents
Method for the selection of synthesis units Download PDFInfo
- Publication number
- EP1526508A1 EP1526508A1 EP04105204A EP04105204A EP1526508A1 EP 1526508 A1 EP1526508 A1 EP 1526508A1 EP 04105204 A EP04105204 A EP 04105204A EP 04105204 A EP04105204 A EP 04105204A EP 1526508 A1 EP1526508 A1 EP 1526508A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- pitch
- segment
- similarity
- synthesis
- units
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000015572 biosynthetic process Effects 0.000 title claims abstract description 66
- 238000003786 synthesis reaction Methods 0.000 title claims abstract description 66
- 238000000034 method Methods 0.000 title claims abstract description 52
- 230000003595 spectral effect Effects 0.000 claims abstract description 37
- 238000012937 correction Methods 0.000 claims description 10
- 238000011524 similarity measure Methods 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 101100422780 Caenorhabditis elegans sur-5 gene Proteins 0.000 description 4
- 238000013518 transcription Methods 0.000 description 3
- 230000035897 transcription Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 101100161175 Caenorhabditis elegans sur-6 gene Proteins 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002844 melting Methods 0.000 description 1
- 230000008018 melting Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 230000001755 vocal effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/0018—Speech coding using phonetic or linguistical decoding of the source; Reconstruction using text-to-speech synthesis
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/06—Elementary speech units used in speech synthesisers; Concatenation rules
Definitions
- the invention relates to a method for selecting units of synthesis.
- the indexing techniques of natural speech units have recently allowed the development of synthesis systems from the particularly powerful text. These techniques are now studied in the context of very low speech rate coding, together with algorithms borrowed from the field of speech recognition, Ref [1-5].
- the main idea is to identify in the speech signal to be coded a almost optimal segmentation in elementary units. These units can be units obtained from a phonetic transcription, which has the disadvantage of having to be corrected manually for an optimal result, or automatically according to spectral stability criteria. From this type of segmentation, and for each segment, we look for unity synthesis in a dictionary obtained during a phase prior learning, and containing reference synthesis units.
- the coding scheme used is to model the space acoustic of the speaker (or speakers) by Markov models hidden (HMM or Hidden Markov Models). These dependent models or independent of the speaker are obtained during a learning phase prior to using algorithms identical to those used in the speech recognition systems.
- the essential difference lies in the fact that the models are learned on vectors grouped by classes automatically and not in a supervised way from a phonetic transcription.
- the learning procedure then consists of automatically obtain the segmentation of learning signals (for example using the so-called temporal decomposition method), group segments obtained in a finite number of classes corresponding to the number of HMM models that we wish to build.
- the number of models is directly related to the desired resolution for represent the acoustic space of the speaker (s).
- these models make it possible to segment the signal to be coded using an algorithm from Viterbi. Segmentation allows to associate to each segment, the index of class and its length. This information is not sufficient to model spectral information, for each of the classes a realization of spectral trajectory is selected from among several so-called synthesis. These units are extracted from the learning base at its segmentation using HMM models. It is possible to take into account context for example by using multiple subclasses allowing for take into account transitions from one class to another. A first index indicates the class to which the segment belongs, a second index specifies the subclass to which it belongs as the index of class of the previous segment. The subclass index is therefore not transmit, and the class index must be stored for the next segment. The subclasses thus defined make it possible to take into account the different transitions to the class associated with the segment in question. To the information spectral one adds the information of prosody, that is to say the value of pitch and energy parameters and their evolutions.
- a classic method is initially to select the unit closest to a spectral point of view then, once the unit selected, to code the prosody information, either independently of the selected unit.
- the method according to the present invention proposes a novel method of selecting the closest synthesis unit together with modeling and quantification of additional information necessary at the level of the decoder for the reproduction of the speech signal.
- information is a segment of speech to code and one uses as proximity criteria the fundamental frequency or pitch, or the spectral distortion, and / or the energy profile and a step of merge of the criteria used to determine the synthesis unit representative.
- the method comprises for example a coding step and / or a pitch correction step by modifying the synthesis profile.
- the step of encoding and / or correcting the pitch may be a Linear transformation of the pitch profile.
- the method is for example used for the selection and / or the coding of synthesis units for a very low bit rate speech coder.
- the speech signal is analyzed frame to frame to extract characteristic parameters (spectral parameters, pitch, energy).
- This analysis is classically using a sliding window set on the horizon of the frame. This frame has a duration of the order of 20 ms, and the update is done with a shift of the analysis window of the order of 10 ms to 20 ms.
- HMM Hidden Markov Model
- model speech segments set of successive frames
- phonemes if the learning phase is supervised (phonetic segmentation and transcription available) or to spectrally stable sounds in the case of a segmentation obtained from Automatic way.
- 64 HMM models are used here, which make it possible recognition phase to associate with each segment the model index HMM identified, and therefore the class to which it belongs.
- HMM models are also used with the help of a Viterbi type algorithm to be coding phase the segmentation and classification of each of the segments (belonging to a class). Each segment is therefore identified by an index between 1 and 64 which is transmitted to the decoder.
- the decoder uses this index to find the synthesis unit in the dictionary built during the learning phase.
- Units of synthesis that make up the dictionary are simply the sequences of parameters associated with the segments obtained on the learning corpus.
- a dictionary class contains all units associated with the same model HMM. Each synthesis unit is therefore characterized by a sequence of spectral parameters, a sequence of pitch value (pitch profile), a sequence of gains (energy profile).
- each class (from 1 to 64) of the dictionary is subdivided into 64 subclasses, where each subclass contains synthesis units that are temporally preceded by a segment belonging to the same class. This approach allows take into account the past context, and thus improve the recovery of the areas transients from one unit to another.
- the present invention relates in particular to a method of selection of a multicriteria synthesis unit.
- the process allows example to take into account simultaneously the pitch, the spectral distortion, and pitch and energy evolution profiles.
- the method may comprise in a variant embodiment a pitch coding step by correction of the synthesis pitch profile detailed below.
- the criterion relating to the evolution profile of the pitch makes it possible in part to consider the voicing information. It is however possible to disable when the segment is totally unvoiced, or the subclass selected is also unvoiced. Indeed, we can notice mainly three types of subclasses: the subclasses containing majority of voiced units, those containing predominantly unvoiced units, and subclasses containing predominantly units mixed.
- the method according to the invention is not limited to optimizing the flow rate allocated to prosody information but also allows to keep for the coding phase the entirety of the synthesis units obtained during the learning phase with a constant number of bits to encode the unit of synthesis.
- the synthesis unit is characterized by both the value pitch and its index. This approach allows in a scheme of speaker-independent coding to cover all pitch values possible and to select the synthesis unit taking into account in part characteristics of the speaker, there is indeed for the same speaker a correlation between the pitch variation range and the characteristics of the vocal tract (especially the length).
- the similarity measure may be a spectral distance.
- Step A9) comprises for example a step where the average the set of spectra of the same segment and the measure of similarity is an intercorrelation measure.
- the spectral distortion criterion is for example calculated on harmonic structures resampled to constant pitch or resampled the pitch of the segment to be coded, after interpolation of initial harmonic structures.
- the similarity criterion will depend on the spectral parameters used (for example the type of parameters used for the representation of the envelope).
- spectral parameters for example the type of parameters used for the representation of the envelope.
- LSP Line Spectral Pair, LSF, Line Spectral Frequencies
- cepstral parameters are usually used, and they can either be derived from a linear prediction analysis (LPCC, Linear Prediction Cepstrum Coefficients) or estimated from a filter bank often on a Mel or Bark perceptual scale (MFCC, Mel Frequency Cepstrum Coefficients).
- Pre-treatment then consists in estimating a spectral envelope from the amplitudes harmonics (linear or polynomial spline interpolation) and to resample the envelope thus obtained, either by using the frequency of the segment to be coded, either by using a frequency fundamental constant (100 Hz for example).
- a fundamental frequency constant allows to pre-calculate all the harmonic structures of synthesis units during the learning phase.
- Re-sampling is then only on the segment to be coded.
- the similarity measure can then be estimated simply from the average harmonic structure of the segment to be coded, and that of the synthesis unit considered.
- This measure of similarity can also be a standardized cross-correlation measurement. It can also be noted that resampling procedure can be done on a scale perceptual frequencies (Mel or Bark).
- the method comprises a step of pitch coding by modifying the synthesis profile. This consists of resynthesizing a pitch profile from that of the synthesis unit selected and a linearly variable gain over the duration of the segment code. It is then sufficient to transmit an additional value for characterize the correction gain over the entire segment.
- the flow associated with the prosody is then between 225 and 300 bits / sec, which leads to an overall bit rate between 450 and 600 bits / sec.
Abstract
Description
L'invention concerne un procédé de sélection d'unités de synthèse.The invention relates to a method for selecting units of synthesis.
Elle concerne par exemple un procédé de sélection et de codage d'unités de synthèse pour un codeur de parole très bas débit, par exemple inférieur à 600 bits/sec.It concerns for example a selection and coding method of synthesis units for a very low speed speech coder, for example less than 600 bits / sec.
Les techniques d'indexation d'unités de parole naturelle ont récemment permis le développement de systèmes de synthèse à partir du texte particulièrement performants. Ces techniques sont dorénavant étudiées dans le cadre du codage à très bas débit de la parole, conjointement avec des algorithmes empruntés au domaine de la reconnaissance vocale, Ref [1-5]. L'idée principale consiste à identifier dans le signal de parole à coder, une segmentation quasi optimale en unités élémentaires. Ces unités peuvent être des unités obtenues à partir d'une transcription phonétique, qui a l'inconvénient de devoir être corrigée manuellement pour un résultat optimal, ou de façon automatique selon des critères de stabilité spectrale. A partir de ce type de segmentation, et pour chacun des segments, on cherche l'unité de synthèse la plus proche dans un dictionnaire obtenu lors d'une phase d'apprentissage préalable, et contenant des unités de synthèse de référence.The indexing techniques of natural speech units have recently allowed the development of synthesis systems from the particularly powerful text. These techniques are now studied in the context of very low speech rate coding, together with algorithms borrowed from the field of speech recognition, Ref [1-5]. The main idea is to identify in the speech signal to be coded a almost optimal segmentation in elementary units. These units can be units obtained from a phonetic transcription, which has the disadvantage of having to be corrected manually for an optimal result, or automatically according to spectral stability criteria. From this type of segmentation, and for each segment, we look for unity synthesis in a dictionary obtained during a phase prior learning, and containing reference synthesis units.
Le schéma de codage utilisé consiste à modéliser l'espace acoustique du locuteur (ou des locuteurs) par des modèles de Markov cachés (HMM ou Hidden Markov Models). Ces modèles dépendants ou indépendants du locuteur sont obtenus lors d'une phase d'apprentissage préalable à partir d'algorithmes identiques à ceux mis en oeuvre dans les systèmes de reconnaissance de la parole. La différence essentielle réside dans le fait que les modèles sont appris sur des vecteurs regroupés par classes de façon automatique et non de manière supervisée à partir d'une transcription phonétique. La procédure d'apprentissage consiste alors à obtenir de façon automatique la segmentation des signaux d'apprentissage (par exemple en utilisant la méthode dite de décomposition temporelle), à regrouper les segments obtenus dans un nombre fini de classes correspondant au nombre de modèles HMM que l'on souhaite construire. Le nombre de modèles est directement lié à la résolution recherchée pour représenter l'espace acoustique du ou des locuteurs. Une fois obtenus, ces modèles permettent de segmenter le signal à coder en utilisant un algorithme de Viterbi. La segmentation permet d'associer à chaque segment, l'indice de classe et sa longueur. Cette information n'étant pas suffisante pour modéliser l'information spectrale, pour chacune des classes une réalisation de trajectoire spectrale est sélectionnée parmi plusieurs unités dites de synthèse. Ces unités sont extraites de la base d'apprentissage lors de sa segmentation utilisant les modèles HMM. Il est possible de tenir compte du contexte par exemple en utilisant plusieurs sous-classes permettant de prendre en compte les transitions d'une classe vers l'autre. Un premier indice indique la classe à laquelle appartient le segment considéré, un deuxième indice précise la sous-classe à laquelle il appartient comme étant l'indice de classe du segment précédent. L'indice de sous-classe n'est donc pas à transmettre, et l'indice de classe doit être mémorisé pour le segment suivant. Les sous-classes ainsi définies permettent de tenir compte des différentes transitions vers la classe associée au segment considéré. A l'information spectrale on ajoute l'information de prosodie, c'est-à-dire la valeur des paramètres de pitch et d'énergie et leurs évolutions.The coding scheme used is to model the space acoustic of the speaker (or speakers) by Markov models hidden (HMM or Hidden Markov Models). These dependent models or independent of the speaker are obtained during a learning phase prior to using algorithms identical to those used in the speech recognition systems. The essential difference lies in the fact that the models are learned on vectors grouped by classes automatically and not in a supervised way from a phonetic transcription. The learning procedure then consists of automatically obtain the segmentation of learning signals (for example using the so-called temporal decomposition method), group segments obtained in a finite number of classes corresponding to the number of HMM models that we wish to build. The number of models is directly related to the desired resolution for represent the acoustic space of the speaker (s). Once obtained, these models make it possible to segment the signal to be coded using an algorithm from Viterbi. Segmentation allows to associate to each segment, the index of class and its length. This information is not sufficient to model spectral information, for each of the classes a realization of spectral trajectory is selected from among several so-called synthesis. These units are extracted from the learning base at its segmentation using HMM models. It is possible to take into account context for example by using multiple subclasses allowing for take into account transitions from one class to another. A first index indicates the class to which the segment belongs, a second index specifies the subclass to which it belongs as the index of class of the previous segment. The subclass index is therefore not transmit, and the class index must be stored for the next segment. The subclasses thus defined make it possible to take into account the different transitions to the class associated with the segment in question. To the information spectral one adds the information of prosody, that is to say the value of pitch and energy parameters and their evolutions.
Dans l'optique de réaliser un codeur très bas débit, il est nécessaire d'optimiser l'allocation des bits et donc du débit entre les paramètres associés à l'enveloppe spectrale et à l'information de prosodie . Une méthode classique consiste dans un premier temps à sélectionner l'unité la plus proche d'un point de vue spectral puis, une fois l'unité sélectionnée, à coder l'information de prosodie, soit de façon indépendante de l'unité sélectionnée.In order to achieve a very low bitrate encoder, it is necessary to optimize the bit allocation and thus the bit rate between parameters associated with the spectral envelope and the prosody information. A classic method is initially to select the unit closest to a spectral point of view then, once the unit selected, to code the prosody information, either independently of the selected unit.
Le procédé selon la présente invention propose une nouvelle méthode de sélection de l'unité de synthèse la plus proche conjointement à la modélisation et à la quantification des informations supplémentaires nécessaires au niveau du décodeur pour la restitution du signal de parole.The method according to the present invention proposes a novel method of selecting the closest synthesis unit together with modeling and quantification of additional information necessary at the level of the decoder for the reproduction of the speech signal.
L'invention concerne un procédé de sélection d'unités de synthèse
d'une information pouvant être décomposée en unités de synthèse. Il
comporte au moins les étapes suivantes :
- déterminer la valeur F0 de la fréquence fondamentale moyenne pour le segment d'information considéré,
- sélectionner un sous-ensemble d'unités de synthèse défini comme étant celui dont les valeurs moyennes de pitch sont les plus proches de la valeur de pitch F0,
- appliquer un ou plusieurs critères de proximité aux unités de synthèse sélectionnées pour déterminer une unité de synthèse représentative du segment d'information.
has at least the following steps:
- determine the value F0 of the average fundamental frequency for the segment of information considered,
- selecting a subset of synthesis units defined as being the one whose average pitch values are closest to the pitch value F0,
- applying one or more proximity criteria to the selected synthesis units to determine a summary unit representative of the information segment.
L'information est par exemple un segment de parole à coder et l'on utilise comme critères de proximité la fréquence fondamentale ou pitch, ou la distorsion spectrale, et/ou le profil d'énergie et on exécute une étape de fusion des critères utilisés afin de déterminer l'unité de synthèse représentative.For example, information is a segment of speech to code and one uses as proximity criteria the fundamental frequency or pitch, or the spectral distortion, and / or the energy profile and a step of merge of the criteria used to determine the synthesis unit representative.
Le procédé comporte par exemple une étape de codage et/ou une étape de correction du pitch par modification du profil de synthèse.The method comprises for example a coding step and / or a pitch correction step by modifying the synthesis profile.
L'étape de codage et/ou correction du pitch peut être une transformation linéaire du profil du pitch d'origine.The step of encoding and / or correcting the pitch may be a Linear transformation of the pitch profile.
Le procédé est par exemple utilisé pour la sélection et/ou le codage d'unités de synthèse pour un codeur de parole très bas débit.The method is for example used for the selection and / or the coding of synthesis units for a very low bit rate speech coder.
L'invention présente notamment les avantages suivants :
- le procédé permet d'optimiser le débit alloué à l'information de prosodie dans le domaine de la parole.
- il permet de conserver, lors de la phase de codage, l'intégralité des unités de synthèse déterminées lors de la phase d'apprentissage avec cependant un nombre de bits constant pour coder l'unité de synthèse.
- Dans un schéma de codage indépendant du locuteur, ce procédé offre la possibilité de couvrir l'ensemble des valeurs de pitch possibles (ou fréquences fondamentales) et de sélectionner l'unité de synthèse en tenant compte en partie des caractéristiques du locuteur.
- La sélection peut s'appliquer à tout système basé sur une sélection d'unités et donc aussi à un système de synthèse à partir du texte.
- the method optimizes the bit rate allocated to prosody information in the speech domain.
- it makes it possible to retain, during the coding phase, the entirety of the synthesis units determined during the learning phase, with however a constant number of bits for coding the synthesis unit.
- In a speaker-independent coding scheme, this method offers the possibility of covering all of the possible pitch values (or fundamental frequencies) and of selecting the synthesis unit taking into account in part the characteristics of the speaker.
- The selection can be applied to any system based on a selection of units and therefore also to a system of synthesis from the text.
D'autres caractéristiques et avantages de l'invention apparaítront mieux à la lecture de la description qui suit d'un exemple de réalisation non limitatif annexé des figures qui représentent :
- La figure 1 un schéma de principe de sélection de l'unité de synthèse associée au segment d'information à coder,
- La figure 2 un schéma de principe d'estimation des critères de similarité pour le profil du pitch,
- La figure 3 un schéma de principe d'estimation des critères de similarité pour le profil énergétique,
- La figure 4 un schéma de principe d'estimation des critères de similarité pour l'enveloppe spectrale,
- La figure 5 un schéma de principe du codage du pitch par correction du profil de pitch de synthèse.
- FIG. 1 is a block diagram for selecting the synthesis unit associated with the information segment to be coded,
- FIG. 2 is a schematic diagram of estimation of the similarity criteria for the pitch profile,
- Figure 3 a schematic diagram of estimation of the similarity criteria for the energy profile,
- FIG. 4 is a schematic diagram of estimation of the similarity criteria for the spectral envelope,
- FIG. 5 is a schematic diagram of pitch coding by correction of the synthesis pitch profile.
Afin de mieux faire comprendre l'idée mise en oeuvre dans la présente l'invention, l'exemple qui suit est donné à titre illustratif et nullement limitatif pour un procédé mis en oeuvre dans un vocodeur, en particulier la sélection et le codage d'unités de synthèse pour un codeur de parole très bas débit. In order to better understand the idea implemented in the presents the invention, the following example is given for illustrative purposes and in no way for a method implemented in a vocoder, in particular the selection and coding of synthesis units for a very speech coder low flow.
Pour rappel, au niveau d'un vocodeur, le signal de parole est analysé trame à trame afin d'extraire les paramètres caractéristiques (paramètres spectraux, pitch, énergie). Cette analyse se fait classiquement à l'aide d'une fenêtre glissante définie sur l'horizon de la trame. Cette trame a une durée de l'ordre de 20 ms, et la mise à jour se fait avec un décalage de la fenêtre d'analyse de l'ordre de 10ms à 20 ms.As a reminder, at the level of a vocoder, the speech signal is analyzed frame to frame to extract characteristic parameters (spectral parameters, pitch, energy). This analysis is classically using a sliding window set on the horizon of the frame. This frame has a duration of the order of 20 ms, and the update is done with a shift of the analysis window of the order of 10 ms to 20 ms.
Lors d'une phase d'apprentissage, un ensemble de modèles de Markov cachés (HMM, Hidden Markov Model) sont appris. Ils permettent de modéliser des segments de parole (ensemble de trames successives) pouvant être associés à des phonèmes si la phase d'apprentissage est supervisée (segmentation et transcription phonétique disponibles) ou à des sons spectralement stables dans le cas d'une segmentation obtenue de façon automatique. On utilise ici 64 modèles HMM, qui permettent lors de la phase de reconnaissance d'associer à chaque segment l'indice du modèle HMM identifié, et donc la classe à laquelle il appartient. Les modèles HMM servent aussi à l'aide d'un algorithme de type Viterbi à réaliser lors de la phase de codage la segmentation et la classification de chacun des segments (appartenance à une classe). Chaque segment est donc identifié par un indice compris entre 1 et 64 qui est transmis au décodeur.During a learning phase, a set of models of Hidden Markov (HMM, Hidden Markov Model) are learned. They allow model speech segments (set of successive frames) can be associated with phonemes if the learning phase is supervised (phonetic segmentation and transcription available) or to spectrally stable sounds in the case of a segmentation obtained from Automatic way. 64 HMM models are used here, which make it possible recognition phase to associate with each segment the model index HMM identified, and therefore the class to which it belongs. HMM models are also used with the help of a Viterbi type algorithm to be coding phase the segmentation and classification of each of the segments (belonging to a class). Each segment is therefore identified by an index between 1 and 64 which is transmitted to the decoder.
Le décodeur utilise cet indice pour retrouver l'unité de synthèse dans le dictionnaire construit lors de la phase d'apprentissage. Les unités de synthèse qui constituent le dictionnaire sont simplement les séquences de paramètres associés aux segments obtenus sur le corpus d'apprentissage.The decoder uses this index to find the synthesis unit in the dictionary built during the learning phase. Units of synthesis that make up the dictionary are simply the sequences of parameters associated with the segments obtained on the learning corpus.
Une classe du dictionnaire contient l'ensemble des unités associées à un même modèle HMM. Chaque unité de synthèse est donc caractérisée par une séquence de paramètres spectraux, une séquence de valeur de pitch (profil de pitch), une séquence de gains (profil énergétique).A dictionary class contains all units associated with the same model HMM. Each synthesis unit is therefore characterized by a sequence of spectral parameters, a sequence of pitch value (pitch profile), a sequence of gains (energy profile).
Afin d'améliorer la qualité de la synthèse, chaque classe (de 1 à 64) du dictionnaire est subdivisée en 64 sous-classes, où chaque sous-classe contient les unités de synthèse qui sont précédées temporellement par un segment appartenant à une même classe. Cette approche permet de tenir compte du contexte passé, et donc d'améliorer la restitution des zones transitoires d'une unité vers l'autre.In order to improve the quality of synthesis, each class (from 1 to 64) of the dictionary is subdivided into 64 subclasses, where each subclass contains synthesis units that are temporally preceded by a segment belonging to the same class. This approach allows take into account the past context, and thus improve the recovery of the areas transients from one unit to another.
La présente invention concerne notamment un procédé de sélection d'une unité de synthèse multicritères. Le procédé permet par exemple de tenir compte simultanément du pitch, de la distorsion spectrale, et des profils d'évolution du pitch et de l'énergie.The present invention relates in particular to a method of selection of a multicriteria synthesis unit. The process allows example to take into account simultaneously the pitch, the spectral distortion, and pitch and energy evolution profiles.
Le procédé de sélection pour un segment de parole à coder
comporte par exemple les étapes de sélection schématisées à la figure 1 :
Il est possible de représenter le pitch sur 5 bits, en utilisant par exemple un quantificateur non uniforme (compression logarithmique) appliqué à la période de pitch.
La valeur du pitch de référence est par exemple obtenue à partir d'un générateur de prosodie dans le cas d'une application en synthèse.
Dans la configuration précédente cela conduit à retenir de façon systématique les 32 unités les plus proches selon le critère du pitch moyen. Il est donc possible de retrouver ces unités au niveau du décodeur à partir du pitch moyen transmis.
It is possible to represent the pitch on 5 bits, using for example a non-uniform quantizer (logarithmic compression) applied to the pitch period.
The value of the reference pitch is for example obtained from a prosody generator in the case of an application in synthesis.
In the previous configuration this leads to systematically retain the 32 closest units according to the average pitch criterion. It is therefore possible to find these units at the decoder from the average pitch transmitted.
Le procédé peut comporter dans une variante de réalisation une étape de codage de pitch par correction du profil de pitch de synthèse exposée en détail ci-après.The method may comprise in a variant embodiment a pitch coding step by correction of the synthesis pitch profile detailed below.
Le critère relatif au profil d'évolution du pitch permet en partie de tenir compte de l'information de voisement. Il est cependant possible de le désactiver lorsque le segment est totalement non voisé, ou que la sous-classe sélectionnée est aussi non voisée. En effet, on peut remarquer principalement trois types de sous-classes : les sous-classes contenant majoritairement des unités voisées, celles contenant majoritairement des unités non voisées, et les sous-classes contenant majoritairement des unités mixtes.The criterion relating to the evolution profile of the pitch makes it possible in part to consider the voicing information. It is however possible to disable when the segment is totally unvoiced, or the subclass selected is also unvoiced. Indeed, we can notice mainly three types of subclasses: the subclasses containing majority of voiced units, those containing predominantly unvoiced units, and subclasses containing predominantly units mixed.
Le procédé selon l'invention ne se limite pas à optimiser le débit alloué à l'information de prosodie mais permet aussi de conserver pour la phase de codage l'intégralité des unités de synthèse obtenues lors de la phase d'apprentissage avec un nombre de bits constant pour coder l'unité de synthèse. En effet l'unité de synthèse est caractérisée à la fois par la valeur de pitch et par son indice. Cette approche permet dans un schéma de codage indépendant du locuteur de couvrir l'ensemble des valeurs de pitch possibles et de sélectionner l'unité de synthèse en tenant compte en partie des caractéristiques du locuteur, il existe en effet pour un même locuteur une corrélation entre la plage de variation du pitch et les caractéristiques du conduit vocal (en particulier la longueur).The method according to the invention is not limited to optimizing the flow rate allocated to prosody information but also allows to keep for the coding phase the entirety of the synthesis units obtained during the learning phase with a constant number of bits to encode the unit of synthesis. Indeed the synthesis unit is characterized by both the value pitch and its index. This approach allows in a scheme of speaker-independent coding to cover all pitch values possible and to select the synthesis unit taking into account in part characteristics of the speaker, there is indeed for the same speaker a correlation between the pitch variation range and the characteristics of the vocal tract (especially the length).
On peut remarquer que le principe de sélection d'unités décrit peut s'appliquer à tout système dont le fonctionnement est basé sur une sélection d'unités et donc aussi à un système de synthèse à partir du texte.It can be noticed that the principle of selection of units described can apply to any system whose operation is based on a selection of units and therefore also to a system of synthesis from the text.
La figure 2 schématise un principe d'estimation des critères de
similarité pour le profil du pitch.
Le procédé comporte par exemple les étapes suivantes :
Le procédé comporte par exemple les étapes suivantes :
Le procédé comporte les étapes suivantes :
The method comprises for example the following steps:
The method comprises for example the following steps:
The method comprises the following steps:
La mesure de similarité peut être une distance spectrale.The similarity measure may be a spectral distance.
L'étape A9) comprend par exemple une étape où l'on moyenne l'ensemble des spectres d'un même segment et la mesure de similarité est une mesure d'intercorrélation.Step A9) comprises for example a step where the average the set of spectra of the same segment and the measure of similarity is an intercorrelation measure.
Le critère de distorsion spectrale est par exemple calculé sur des structures harmoniques ré-échantillonnées à pitch constant ou ré-échantillonnées au pitch du segment à coder, après interpolation des structures harmoniques initiales.The spectral distortion criterion is for example calculated on harmonic structures resampled to constant pitch or resampled the pitch of the segment to be coded, after interpolation of initial harmonic structures.
Le critère de similarité va dépendre des paramètres spectraux utilisés (par exemple du type de paramètres utilisés pour la représentation de l'enveloppe). Plusieurs types de paramètres spectraux peuvent être utilisés, dans la mesure où ils permettent de définir une mesure de distorsion spectrale. Dans le domaine du codage de la parole, il est courant d'utiliser les paramètres LSP ou LSF (LSP, Line Spectral Pair, LSF, Line Spectral Frequencies) dérivés d'une analyse par prédiction linéaire. Dans le domaine de la reconnaissance vocale, les paramètres cepstraux sont généralement utilisés, et ils peuvent soit être dérivés d'une analyse par prédiction linéaire (LPCC, Linear Prediction Cepstrum Coefficients) ou estimés à partir d'un banc de filtres souvent sur une échelle perceptuelle de type Mel ou Bark (MFCC, Mel Frequency Cepstrum Coefficients). Il est aussi possible dans la mesure où on utilise une modélisation sinusoïdale de la composante harmonique du signal de parole, d'utiliser directement les amplitudes des fréquences harmoniques. Ces derniers paramètres étant estimés en fonction du pitch ne peuvent être utilisés directement pour calculer une distance. Le nombre de coefficients obtenus est en effet variable en fonction du pitch, contrairement aux paramètres LPCC, MFCC ou LSF. Un pré-traitement consiste alors à estimer une enveloppe spectrale à partir des amplitudes harmoniques (interpolation linéaire ou polynomiale de type spline) et à réechantillonner l'enveloppe ainsi obtenue, soit en utilisant la fréquence fondamentale du segment à coder, soit en utilisant une fréquence fondamentale constante (100 Hz par exemple). Une fréquence fondamentale constante permet de pré-calculer l'ensemble des structures harmoniques des unités de synthèse lors de la phase d'apprentissage. Le re-échantillonnage se fait alors uniquement sur le segment à coder. D'autre part, si on se limite à un alignement temporel par interpolation linéaire, il est possible de moyenner les structures harmoniques sur l'ensemble des segments considérés. La mesure de similarité peut alors être estimée simplement à partir de la structure harmonique moyenne du segment à coder, et celle de l'unité de synthèse considérée. Cette mesure de similarité peut aussi être une mesure d'intercorrélation normalisée. On peut aussi noter que la procédure de ré-échantillonnage peut s'effectuer sur une échelle perceptuelle des fréquences (Mel ou Bark). The similarity criterion will depend on the spectral parameters used (for example the type of parameters used for the representation of the envelope). Several types of spectral parameters can be used, insofar as they make it possible to define a measure of distortion spectral. In the field of speech coding, it is common to use LSP or LSF parameters (LSP, Line Spectral Pair, LSF, Line Spectral Frequencies) derived from a linear prediction analysis. In the field speech recognition, cepstral parameters are usually used, and they can either be derived from a linear prediction analysis (LPCC, Linear Prediction Cepstrum Coefficients) or estimated from a filter bank often on a Mel or Bark perceptual scale (MFCC, Mel Frequency Cepstrum Coefficients). It is also possible in the as we use a sinusoidal modeling of the component harmonic of the speech signal, to directly use the amplitudes of the harmonic frequencies. These last parameters being estimated according to pitch can not be used directly to calculate a distance. The number of coefficients obtained is indeed variable depending on the pitch, unlike LPCC, MFCC or LSF parameters. Pre-treatment then consists in estimating a spectral envelope from the amplitudes harmonics (linear or polynomial spline interpolation) and to resample the envelope thus obtained, either by using the frequency of the segment to be coded, either by using a frequency fundamental constant (100 Hz for example). A fundamental frequency constant allows to pre-calculate all the harmonic structures of synthesis units during the learning phase. Re-sampling is then only on the segment to be coded. On the other hand, if we limit ourselves to linear alignment by linear interpolation, it is possible to average harmonic structures on all segments considered. The similarity measure can then be estimated simply from the average harmonic structure of the segment to be coded, and that of the synthesis unit considered. This measure of similarity can also be a standardized cross-correlation measurement. It can also be noted that resampling procedure can be done on a scale perceptual frequencies (Mel or Bark).
Pour la procédure d'alignement temporel il est possible d'utiliser
soit un algorithme de programmation dynamique (DTW, Dynamic Time
Warping), soit d'effectuer une interpolation linéaire simple (ajustement
linéaire des longueurs). Dans l'hypothèse où l'on ne souhaite pas transmettre
d'information supplémentaire relative au chemin d'alignement, il est
préférable d'utiliser une simple interpolation linéaire des paramètres. La prise
en compte du meilleur alignement est alors en partie réaliser par la
procédure de sélection.
Codage du pitch par modification du profil de synthèseFor the time alignment procedure it is possible to use either a Dynamic Time Warping (DTW) algorithm or to perform a simple linear interpolation (linear length adjustment). In the event that it is not desired to transmit additional information relating to the alignment path, it is preferable to use a simple linear interpolation of the parameters. Taking into account the best alignment is then partly achieved by the selection procedure.
Pitch coding by modifying the synthesis profile
Selon un mode de réalisation, le procédé comporte une étape de codage du pitch par modification du profil de synthèse. Cela consiste à resynthétiser un profil de pitch à partir de celui de l'unité de synthèse sélectionnée et un gain linéairement variable sur la durée du segment à coder. Il suffit alors de transmettre une valeur supplémentaire pour caractériser le gain correcteur sur l'ensemble du segment.According to one embodiment, the method comprises a step of pitch coding by modifying the synthesis profile. This consists of resynthesizing a pitch profile from that of the synthesis unit selected and a linearly variable gain over the duration of the segment code. It is then sufficient to transmit an additional value for characterize the correction gain over the entire segment.
Le pitch reconstruit au niveau du décodeur est donné par
l'équation suivante :
Cela correspond à une transformation linéaire du profil du pitch.
Les valeurs optimales de a et b sont estimées au niveau du codeur en
minimisant l'erreur quadratique moyenne :
ce qui conduit aux relations suivantes :
Remarque : cette méthode de correction peut bien entendu s'appliquer au
profil énergétique. The pitch reconstructed at the level of the decoder is given by the following equation:
This corresponds to a linear transformation of the pitch profile.
The optimal values of a and b are estimated at the encoder level by minimizing the mean squared error: which leads to the following relationships:
Note: This correction method can of course be applied to the energy profile.
Les informations relatives au débit associé au schéma de codage décrit précédemment sont les suivantes :
- Indice de classe sur 6 bits (64 classes)
- Indice de l'unité sélectionnée sur 5 bits (32 unités par sous-classe)
- Longueur du segment sur 4 bits (
de 3 à 18 trames)
- 6-bit class index (64 classes)
- Index of the unit selected on 5 bits (32 units per subclass)
- Segment length on 4 bits (from 3 to 18 frames)
Le nombre moyen de segments par seconde se situe entre 15 et 20; ce qui conduit à un débit de base situé entre 225 et 300 bits/sec pour la configuration précédente. A ce débit de base vient s'ajouter le débit nécessaire pour représenter l'information de pitch et d'énergie.
- FO moyen sur 5 bits
- Coefficient correcteur du profil de pitch sur 5bits
- Gain correcteur sur 5 bits
- Mean FO on 5 bits
- Corrective coefficient of the 5bits pitch profile
- 5-bit correction gain
Le débit associé à la prosodie se situe alors entre 225 et 300 bits/sec, ce qui conduit à un débit global entre 450 et 600 bits/sec.The flow associated with the prosody is then between 225 and 300 bits / sec, which leads to an overall bit rate between 450 and 600 bits / sec.
Claims (14)
moins les étapes suivantes :
pour un segment d'information considéré :
less the following steps:
for a segment of information considered:
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
FR0312494 | 2003-10-24 | ||
FR0312494A FR2861491B1 (en) | 2003-10-24 | 2003-10-24 | METHOD FOR SELECTING SYNTHESIS UNITS |
Publications (2)
Publication Number | Publication Date |
---|---|
EP1526508A1 true EP1526508A1 (en) | 2005-04-27 |
EP1526508B1 EP1526508B1 (en) | 2009-05-27 |
Family
ID=34385390
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP04105204A Active EP1526508B1 (en) | 2003-10-24 | 2004-10-21 | Method for the selection of synthesis units |
Country Status (6)
Country | Link |
---|---|
US (1) | US8195463B2 (en) |
EP (1) | EP1526508B1 (en) |
AT (1) | ATE432525T1 (en) |
DE (1) | DE602004021221D1 (en) |
ES (1) | ES2326646T3 (en) |
FR (1) | FR2861491B1 (en) |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4265501B2 (en) * | 2004-07-15 | 2009-05-20 | ヤマハ株式会社 | Speech synthesis apparatus and program |
WO2006040908A1 (en) * | 2004-10-13 | 2006-04-20 | Matsushita Electric Industrial Co., Ltd. | Speech synthesizer and speech synthesizing method |
US7126324B1 (en) * | 2005-11-23 | 2006-10-24 | Innalabs Technologies, Inc. | Precision digital phase meter |
EP2058803B1 (en) * | 2007-10-29 | 2010-01-20 | Harman/Becker Automotive Systems GmbH | Partial speech reconstruction |
US8401849B2 (en) * | 2008-12-18 | 2013-03-19 | Lessac Technologies, Inc. | Methods employing phase state analysis for use in speech synthesis and recognition |
US8731931B2 (en) | 2010-06-18 | 2014-05-20 | At&T Intellectual Property I, L.P. | System and method for unit selection text-to-speech using a modified Viterbi approach |
US9664518B2 (en) * | 2010-08-27 | 2017-05-30 | Strava, Inc. | Method and system for comparing performance statistics with respect to location |
CN102651217A (en) * | 2011-02-25 | 2012-08-29 | 株式会社东芝 | Method and equipment for voice synthesis and method for training acoustic model used in voice synthesis |
US9116922B2 (en) | 2011-03-31 | 2015-08-25 | Strava, Inc. | Defining and matching segments |
US9291713B2 (en) | 2011-03-31 | 2016-03-22 | Strava, Inc. | Providing real-time segment performance information |
US8620646B2 (en) * | 2011-08-08 | 2013-12-31 | The Intellisis Corporation | System and method for tracking sound pitch across an audio signal using harmonic envelope |
US10453479B2 (en) | 2011-09-23 | 2019-10-22 | Lessac Technologies, Inc. | Methods for aligning expressive speech utterances with text and systems therefor |
US8718927B2 (en) | 2012-03-12 | 2014-05-06 | Strava, Inc. | GPS data repair |
US8886539B2 (en) * | 2012-12-03 | 2014-11-11 | Chengjun Julian Chen | Prosody generation using syllable-centered polynomial representation of pitch contours |
CN113412512A (en) * | 2019-02-20 | 2021-09-17 | 雅马哈株式会社 | Sound signal synthesis method, training method for generating model, sound signal synthesis system, and program |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020065655A1 (en) * | 2000-10-18 | 2002-05-30 | Thales | Method for the encoding of prosody for a speech encoder working at very low bit rates |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10260692A (en) * | 1997-03-18 | 1998-09-29 | Toshiba Corp | Method and system for recognition synthesis encoding and decoding of speech |
JP2000056789A (en) * | 1998-06-02 | 2000-02-25 | Sanyo Electric Co Ltd | Speech synthesis device and telephone set |
JP2000075878A (en) * | 1998-08-31 | 2000-03-14 | Canon Inc | Device and method for voice synthesis and storage medium |
US6574593B1 (en) * | 1999-09-22 | 2003-06-03 | Conexant Systems, Inc. | Codebook tables for encoding and decoding |
US6581032B1 (en) * | 1999-09-22 | 2003-06-17 | Conexant Systems, Inc. | Bitstream protocol for transmission of encoded voice signals |
JP3515039B2 (en) * | 2000-03-03 | 2004-04-05 | 沖電気工業株式会社 | Pitch pattern control method in text-to-speech converter |
JP3728172B2 (en) * | 2000-03-31 | 2005-12-21 | キヤノン株式会社 | Speech synthesis method and apparatus |
SE521600C2 (en) * | 2001-12-04 | 2003-11-18 | Global Ip Sound Ab | Lågbittaktskodek |
CA2388352A1 (en) * | 2002-05-31 | 2003-11-30 | Voiceage Corporation | A method and device for frequency-selective pitch enhancement of synthesized speed |
-
2003
- 2003-10-24 FR FR0312494A patent/FR2861491B1/en not_active Expired - Fee Related
-
2004
- 2004-10-21 EP EP04105204A patent/EP1526508B1/en active Active
- 2004-10-21 DE DE602004021221T patent/DE602004021221D1/en active Active
- 2004-10-21 ES ES04105204T patent/ES2326646T3/en active Active
- 2004-10-21 AT AT04105204T patent/ATE432525T1/en not_active IP Right Cessation
- 2004-10-22 US US10/970,731 patent/US8195463B2/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020065655A1 (en) * | 2000-10-18 | 2002-05-30 | Thales | Method for the encoding of prosody for a speech encoder working at very low bit rates |
Non-Patent Citations (4)
Title |
---|
BAUDOIN G ; EL CHAMI F: "Corpus based very low bit rate speech coding", 2003 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, vol. 1, 6 April 2003 (2003-04-06) - 10 April 2003 (2003-04-10), HONG KONG, CHINA, pages 792 - 795, XP002285304, ISBN: 0-7803-7663-3 * |
C.BAVEREL ET AL., CODAGE DE LA PAROLE À TRÈS BAS DÉBIT PAR INDEXATION D'UNITÉS DE TAILLE VARIABLE, 2001 |
M. PADELLINI, F. CAPMAN, G. BAUDOIN: "Dynamic Unit Selection for Very Low Bit Rate Coding at 500 bits/sec", TSD 2004, 8 September 2004 (2004-09-08) - 11 September 2004 (2004-09-11), BRNO, CZ, XP002312723 * |
M. PADELLINI, G. BAUDOIN AND F. CAPMAN: "Codage de la parole a très bas débit par indexation d'unités de taille variable", RJC'2003, 5ÈMES RENCONTRES JEUNES CHERCHEURS EN PAROLE, 23 September 2003 (2003-09-23) - 26 September 2003 (2003-09-26), GRENOBLE, FRANCE, XP002285303 * |
Also Published As
Publication number | Publication date |
---|---|
FR2861491B1 (en) | 2006-01-06 |
ATE432525T1 (en) | 2009-06-15 |
US20050137871A1 (en) | 2005-06-23 |
US8195463B2 (en) | 2012-06-05 |
DE602004021221D1 (en) | 2009-07-09 |
ES2326646T3 (en) | 2009-10-16 |
EP1526508B1 (en) | 2009-05-27 |
FR2861491A1 (en) | 2005-04-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP1316087B1 (en) | Transmission error concealment in an audio signal | |
US7996222B2 (en) | Prosody conversion | |
JP4824167B2 (en) | Periodic speech coding | |
Acero | Formant analysis and synthesis using hidden Markov models | |
EP1526508B1 (en) | Method for the selection of synthesis units | |
JP2002533772A (en) | Variable rate speech coding | |
WO2005106853A1 (en) | Method and system for the quick conversion of a voice signal | |
JPH075892A (en) | Voice recognition method | |
WO2002097798A1 (en) | Method and apparatus for improved voicing determination in speech signals containing high levels of jitter | |
EP1606792B1 (en) | Method for analyzing fundamental frequency information and voice conversion method and system implementing said analysis method | |
JPH08123484A (en) | Method and device for signal synthesis | |
Bhatt | Simulation and overall comparative evaluation of performance between different techniques for high band feature extraction based on artificial bandwidth extension of speech over proposed global system for mobile full rate narrow band coder | |
Miyamoto et al. | Non-linear harmonic generation based blind bandwidth extension considering aliasing artifacts | |
Khonglah et al. | Speech enhancement using source information for phoneme recognition of speech with background music | |
EP1846918B1 (en) | Method of estimating a voice conversion function | |
Bachhav et al. | Exploiting explicit memory inclusion for artificial bandwidth extension | |
Ramasubramanian et al. | Ultra low bit-rate speech coding | |
JP2000514207A (en) | Speech synthesis system | |
Min et al. | Deep vocoder: Low bit rate compression of speech with deep autoencoder | |
Albahri et al. | Artificial bandwidth extension to improve automatic emotion recognition from narrow-band coded speech | |
Ramasubramanian | Ultra low bit-rate speech coding: An overview and recent results | |
Sharma et al. | Non-intrusive bit-rate detection of coded speech | |
Do et al. | Objective evaluation of HMM-based speech synthesis system using Kullback-Leibler divergence | |
Vuppala et al. | Effect of low bit rate speech coding on epoch extraction | |
En-Najjary et al. | Fast GMM-based voice conversion for text-to-speech synthesis systems |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LI LU MC NL PL PT RO SE SI SK TR |
|
AX | Request for extension of the european patent |
Extension state: AL HR LT LV MK |
|
17P | Request for examination filed |
Effective date: 20051007 |
|
AKX | Designation fees paid |
Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LI LU MC NL PL PT RO SE SI SK TR |
|
17Q | First examination report despatched |
Effective date: 20070612 |
|
GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
GRAS | Grant fee paid |
Free format text: ORIGINAL CODE: EPIDOSNIGR3 |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Kind code of ref document: B1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LI LU MC NL PL PT RO SE SI SK TR |
|
REG | Reference to a national code |
Ref country code: GB Ref legal event code: FG4D Free format text: NOT ENGLISH |
|
REG | Reference to a national code |
Ref country code: CH Ref legal event code: EP |
|
REG | Reference to a national code |
Ref country code: IE Ref legal event code: FG4D Free format text: LANGUAGE OF EP DOCUMENT: FRENCH |
|
REF | Corresponds to: |
Ref document number: 602004021221 Country of ref document: DE Date of ref document: 20090709 Kind code of ref document: P |
|
REG | Reference to a national code |
Ref country code: ES Ref legal event code: FG2A Ref document number: 2326646 Country of ref document: ES Kind code of ref document: T3 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: PT Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090927 Ref country code: FI Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: AT Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 |
|
NLV1 | Nl: lapsed or annulled due to failure to fulfill the requirements of art. 29p and 29m of the patents act | ||
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: SI Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: NL Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: PL Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: SE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090827 |
|
REG | Reference to a national code |
Ref country code: IE Ref legal event code: FD4D |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: EE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: DK Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: RO Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: CZ Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 Ref country code: IE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: SK Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: BG Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090827 |
|
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
BERE | Be: lapsed |
Owner name: THALES Effective date: 20091031 |
|
26N | No opposition filed |
Effective date: 20100302 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: MC Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20091031 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: GR Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090828 Ref country code: BE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20091031 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: LU Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20091021 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: HU Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20091128 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: TR Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: CY Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20090527 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: PLFP Year of fee payment: 13 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: PLFP Year of fee payment: 14 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: DE Payment date: 20171018 Year of fee payment: 14 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: CH Payment date: 20171013 Year of fee payment: 14 Ref country code: GB Payment date: 20171013 Year of fee payment: 14 Ref country code: ES Payment date: 20171102 Year of fee payment: 14 Ref country code: IT Payment date: 20171024 Year of fee payment: 14 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: PLFP Year of fee payment: 15 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R119 Ref document number: 602004021221 Country of ref document: DE |
|
REG | Reference to a national code |
Ref country code: CH Ref legal event code: PL |
|
GBPC | Gb: european patent ceased through non-payment of renewal fee |
Effective date: 20181021 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20190501 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: CH Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20181031 Ref country code: LI Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20181031 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: GB Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20181021 Ref country code: IT Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20181021 |
|
REG | Reference to a national code |
Ref country code: ES Ref legal event code: FD2A Effective date: 20191203 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: ES Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20181022 |
|
P01 | Opt-out of the competence of the unified patent court (upc) registered |
Effective date: 20230517 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: FR Payment date: 20230921 Year of fee payment: 20 |