International workshop on audio-visual affective prosody in social interaction & second language learning

(AVAP 2015)

Bordeaux, 5 and 6 March 2015


Yoshinori Sagisaka

Sentiment information for communicative prosody and multi-modal speech description

Waseda University, Tokyo, Japan


In this talk, we would like to introduce two of our recent studies. The first one concerns on communicative prosody generation using sentiment impressions extracted from constituent lexicons. Based on a series of our studies for a decade, we employ sentiment information consisting of three perceptual impressions extracted using MDS analysis to show the posibility of communicative prosody computation. The second one shows the possibility of sentiment information description embedded in speech using color. A psychological experimet on color selection is introduced to show the correlations between speech features and collor attributes selected by listening single vowels. Through the introduction of these two research topics, we would like to comment on a new possibility for comuting prosody based on sentiment information and its description using other modality.