DocumentCode :
2726405
Title :
Significance of Word and Syllable Level Information for Expressive Speech Processing
Author :
Rao, K. Sreenivasa ; Prasanna, S.R.M. ; Sagar, T.V.
Author_Institution :
Sch. of Inf. Technol., IIT Kharagpur, Kharagpur
fYear :
2009
fDate :
4-6 Feb. 2009
Firstpage :
159
Lastpage :
162
Abstract :
In general, human beings make use of expressions (emotions) through speech, facial movements and gestures for conveying the crucial information. Mostly, expressions in speech can be attributed to longer segments, i.e., suprasegmental features also known to be prosodic features. In this paper we analyze the expressions in speech using prosodic features from utterance level, word level and syllable level. The emotions considered for the analysis are anger,compassion, happy and neutral. The prosodic features used in the analysis are duration, intonation (pitch) and energy. The analysis is performed on SUSE (Speech Under Simulated Emotion) database. The results of the analysis are used for synthesizing the expressions in neutral speech. The synthesis experiments using the features from utterance level to syllable level showed that a steady improvement in the quality of speech for the desired expressions.
Keywords :
emotion recognition; speech recognition; speech synthesis; emotion recognition; facial movement; gesture analysis; prosodic feature; speech expression; speech processing; speech synthesis; suprasegmental feature; syllable level information; utterance level; word level information; Emotion recognition; Feature extraction; Humans; Information technology; Pattern recognition; Performance analysis; Speech analysis; Speech processing; Speech recognition; Speech synthesis; Prosodic features; duration; emotions; energy; expression; pitch; syllable level; utterance level; word level;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advances in Pattern Recognition, 2009. ICAPR '09. Seventh International Conference on
Conference_Location :
Kolkata
Print_ISBN :
978-1-4244-3335-3
Type :
conf
DOI :
10.1109/ICAPR.2009.47
Filename :
4782764
Link To Document :
بازگشت