Van Heerden, CJBarnard, E2008-01-242008-01-242007-11Van Heerden, CJ and Barnard, E. 2007. Speaker-specific variability of phoneme durations. 18th Annual Symposium of the Pattern Recognition Association of South Africa (PRASA), Pietermaritzburg, Kwazulu-Natal, South Africa, 28-30 November 2007, pp 6978-1-86840-656-2http://hdl.handle.net/10204/1974http://search.sabinet.co.za/WebZ/images/ejour/comp/comp_v40_a9.pdf:sessionid=0:bad=http://search.sabinet.co.za/ejour/ejour_badsearch.html:portal=ejournal:2007: PRASAThis paper is published in South African Computer Journal, Vol 40, pp 44-50The durations of phonemes varies for different speakers. To this end, the correlations between phonemes across different speakers are studied and a novel approach to predict unknown phoneme durations from the values of known phoneme durations for a particular speaker are presented, based on the maximum likelihood criterion. Several interesting patterns are observed. Phonemes from the same broad phonetic class tend to covey most strongly (and therefore intra-class predictions of unknown phoneme durations are most accurate), but significant cross-class correlations are also present. Consequently, knowledge of only a few highly-correlated phonemes’ durations is necessary to make a good duration predictionenPhonemes durationSpeech recognitionEigen vectorsConference PresentationVan Heerden, C., & Barnard, E. (2007). Speaker-specific variability of phoneme durations. 18th Annual Symposium of the Pattern Recognition Association of South Africa (PRASA). http://hdl.handle.net/10204/1974Van Heerden, CJ, and E Barnard. "Speaker-specific variability of phoneme durations." (2007): http://hdl.handle.net/10204/1974Van Heerden C, Barnard E, Speaker-specific variability of phoneme durations; 18th Annual Symposium of the Pattern Recognition Association of South Africa (PRASA); 2007. http://hdl.handle.net/10204/1974 .TY - Conference Presentation AU - Van Heerden, CJ AU - Barnard, E AB - The durations of phonemes varies for different speakers. To this end, the correlations between phonemes across different speakers are studied and a novel approach to predict unknown phoneme durations from the values of known phoneme durations for a particular speaker are presented, based on the maximum likelihood criterion. Several interesting patterns are observed. Phonemes from the same broad phonetic class tend to covey most strongly (and therefore intra-class predictions of unknown phoneme durations are most accurate), but significant cross-class correlations are also present. Consequently, knowledge of only a few highly-correlated phonemes’ durations is necessary to make a good duration prediction DA - 2007-11 DB - ResearchSpace DP - CSIR KW - Phonemes duration KW - Speech recognition KW - Eigen vectors LK - https://researchspace.csir.co.za PY - 2007 SM - 978-1-86840-656-2 T1 - Speaker-specific variability of phoneme durations TI - Speaker-specific variability of phoneme durations UR - http://hdl.handle.net/10204/1974 ER -