Cluster Analysis of Differential Spectral Envelopes on Emotional Speech
2010 (English)In: 11TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2010 (INTERSPEECH 2010), VOLS 1-4, 2010, 322-325 p.Conference paper (Refereed)
This paper reports on the analysis of the spectral variation of emotional speech. Spectral envelopes oftime aligned speech frames are compared between emotionally neutral and active utterances. Statisticsare computed over the resulting differential spectral envelopes for each phoneme. Finally, thesestatistics are classified using agglomerative hierarchical clustering and a measure of dissimilaritybetween statistical distributions and the resulting clusters are analysed. The results show that thereare systematic changes in spectral envelopes when going from neutral to sad or happy speech, andthose changes depend on the valence of the emotional content (negative, positive) as well as on thephonetic properties of the sounds such as voicing and place of articulation.
Place, publisher, year, edition, pages
2010. 322-325 p.
emotional speech, hierarchical clustering, spectral envelopes
Computer Science Language Technology (Computational Linguistics)
IdentifiersURN: urn:nbn:se:kth:diva-52172ISI: 000294382400077ScopusID: 2-s2.0-79959855615OAI: oai:DiVA.org:kth-52172DiVA: diva2:465467
11th Annual Conference of the International-Speech-Communication-Association 2010. Makuhari, JAPAN. SEP 26-30, 2010
tmh_import_11_12_14. QC 201112162011-12-142011-12-142011-12-16Bibliographically approved