DTW-Based Phonetic Labeling Using Explicit Phoneme Duration Constraints
Identifieur interne : 00DC57 ( Main/Merge ); précédent : 00DC56; suivant : 00DC58DTW-Based Phonetic Labeling Using Explicit Phoneme Duration Constraints
Auteurs : Y. Gong ; Jean-Paul Haton [France]Source :
English descriptors
Abstract
We address the problem of automatic phonemic machine labeling of speech database. We are particularly interested in using one hand-labeled utterance of a speaker to automatically label utterances of the same text for a large variety of speakers. Conventional DTW algorithms used in speech recognition may produce one-to-many or many-to-one alignment between reference and test. Since the difference in speaking rate is globally much lower, the two types of alignment results are not realistic. We describe a new DTW algorithm which explicitly incorporates in the best path search procedure constraints of phoneme duration. During the alignment of two utterances, the pathes goes across successive phoneme boundaries, both in reference axis and in the test axis. The basic idea is, for each phoneme to-phoneme transition, weighting the cumulated distance by the duration difference between reference and test phonemes such that its durational deviation from the reference will penalize the path cost. Conventional DTW algorithm is just a special case of the new algorithm. The results show that\, : \begin{itemize} \item[-]without phoneme duration constraint, about 11========percnt; of machine labeled segments are useless because of abnormal time duration. With phoneme duration constraint this percent drops to 0.05========percnt; ; \item[-]at the meantime, due to the improved alignment accuracy we observed phrase recognition error reduction of 5========percnt;, on a 33.6 thousand vocabulary task. \end{itemize}
Links toward previous steps (curation, corpus...)
- to stream Crin, to step Corpus: 000E89
- to stream Crin, to step Curation: 000E89
- to stream Crin, to step Checkpoint: 003637
Links to Exploration step
CRIN:gong92eLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="513">DTW-Based Phonetic Labeling Using Explicit Phoneme Duration Constraints</title></titleStmt><publicationStmt><idno type="RBID">CRIN:gong92e</idno><date when="1992" year="1992">1992</date><idno type="wicri:Area/Crin/Corpus">000E89</idno><idno type="wicri:Area/Crin/Curation">000E89</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Curation">000E89</idno><idno type="wicri:Area/Crin/Checkpoint">003637</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Checkpoint">003637</idno><idno type="wicri:Area/Main/Merge">00DC57</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">DTW-Based Phonetic Labeling Using Explicit Phoneme Duration Constraints</title><author><name sortKey="Gong, Y" sort="Gong, Y" uniqKey="Gong Y" first="Y." last="Gong">Y. Gong</name></author><author><name sortKey="Haton, J P" sort="Haton, J P" uniqKey="Haton J" first="J.-P." last="Haton">Jean-Paul Haton</name><affiliation><country>France</country><placeName><settlement type="city">Nancy</settlement><region type="region" nuts="2">Grand Est</region><region type="region" nuts="2">Lorraine (région)</region></placeName><orgName type="laboratoire" n="5">Laboratoire lorrain de recherche en informatique et ses applications</orgName><orgName type="university">Université de Lorraine</orgName><orgName type="institution">Centre national de la recherche scientifique</orgName><orgName type="institution">Institut national de recherche en informatique et en automatique</orgName></affiliation></author></analytic></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>duration constraints</term><term>dynamic time warping</term><term>speech database labeling</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en" wicri:score="4395">We address the problem of automatic phonemic machine labeling of speech database. We are particularly interested in using one hand-labeled utterance of a speaker to automatically label utterances of the same text for a large variety of speakers. Conventional DTW algorithms used in speech recognition may produce one-to-many or many-to-one alignment between reference and test. Since the difference in speaking rate is globally much lower, the two types of alignment results are not realistic. We describe a new DTW algorithm which explicitly incorporates in the best path search procedure constraints of phoneme duration. During the alignment of two utterances, the pathes goes across successive phoneme boundaries, both in reference axis and in the test axis. The basic idea is, for each phoneme to-phoneme transition, weighting the cumulated distance by the duration difference between reference and test phonemes such that its durational deviation from the reference will penalize the path cost. Conventional DTW algorithm is just a special case of the new algorithm. The results show that\, : \begin{itemize} \item[-]without phoneme duration constraint, about 11========percnt; of machine labeled segments are useless because of abnormal time duration. With phoneme duration constraint this percent drops to 0.05========percnt; ; \item[-]at the meantime, due to the improved alignment accuracy we observed phrase recognition error reduction of 5========percnt;, on a 33.6 thousand vocabulary task. \end{itemize}</div></front></TEI></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Main/Merge
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 00DC57 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Merge/biblio.hfd -nk 00DC57 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Lorraine
|area= InforLorV4
|flux= Main
|étape= Merge
|type= RBID
|clé= CRIN:gong92e
|texte= DTW-Based Phonetic Labeling Using Explicit Phoneme Duration Constraints
}}
| This area was generated with Dilib version V0.6.33. |  |