Search Results for author: Nicolas Audibert

Found 9 papers, 0 papers with code

PATATRA and PATAFreq: two French databases for the documentation of within-speaker variability in speech

no code implementations LREC 2022 Cécile Fougeron, Nicolas Audibert, Cedric Gendrot, Estelle Chardenon, Louise Wohmann

In this paper, we introduce two French databases which contain recordings of 9 to 11 speakers recorded over 9 to 18 sessions, allowing comparisons of speech tasks with a different delay between the repetitions: 3 repetitions within the same session, 6 to 10 repetitions on different days during a two months period, 5 to 9 repetitions on different years.

Dis-moi comment tu varies ton d\'ebit, je te dirai qui tu es (Tell me how you vary your speech flow, I'll tell you who you are Studying inter-speaker variability makes it possible to identify discriminating or even identifying phonetic characteristics)

no code implementations JEPTALNRECITAL 2020 Estelle Chardenon, C{\'e}cile Fougeron, Nicolas Audibert, C{\'e}dric Gendrot

Si l{'}{\'e}tude de la variabilit{\'e} entre locuteurs permet d{'}identifier des caract{\'e}ristiques phon{\'e}tiques potentiellement discriminantes, voire sp{\'e}cifiques, il est essentiel de comprendre, si et comment, ces caract{\'e}ristiques varient chez un m{\^e}me locuteur.

Towards Interactive Annotation for Hesitation in Conversational Speech

no code implementations LREC 2020 Jane Wottawa, Marie Tahon, Apolline Marin, Nicolas Audibert

Furthermore, regression models were trained in order to allow automatic prediction of hesitation for speech chunks that do not have a manual annotation.

Que nous apprennent les gros corpus sur l'harmonie vocalique en fran\ccais ? (What can we learn from big speech corpora about French vowel harmony?)

no code implementations JEPTALNRECITAL 2016 Giuseppina Turco, C{\'e}cile Fougeron, Nicolas Audibert

Afin de mieux identifier le poids relatif des diff{\'e}rents facteurs d{\'e}crits dans la litt{\'e}rature comme influen{\c{c}}ant le ph{\'e}nom{\`e}ne d{'}harmonie vocalique (HV) en fran{\c{c}}ais, 33k mots extraits de deux corpus de parole continue et pr{\'e}sentant un contexte d{'}HV possible V1C(C)V2 (V1∈e, ɛ, o, ɔ) sont analys{\'e}s. Le degr{\'e} d{'}HV est mesur{\'e} en termes d{'}abaissement du F1 de V1 induit par la pr{\'e}sence d{'}une V2 /+haut/ (ferm{\'e}e ou mi-ferm{\'e}e) par rapport {\`a} une V2 /-haut/ (ouverte ou mi-ouverte).

Cannot find the paper you are looking for? You can Submit a new open access paper.