Abstracts Submitted: 250
Number of Users: 297
View Abstracts Submitted
Back to home Page
Based on the relationship between the length of the uniform sectional area acoustic tube and its resonance frequency, we proposed a method of estimating the vocal tract length ratio between two speakers from the formant trajectory, and a highly reliable result was obtained. However, since the conventional method requires utterances of the same word group by each speaker or a group of words having different utterance speeds, it is desirable to exclude restrictions on such utterance and to estimate the vocal tract length from a spoken language or spontaneous speech. In this paper, we propose a simplified method using the accumulated mean of formant frequencies of many words or long sentences. In the estimation by the proposed method, it was shown that estimation accuracy comparable to that of the conventional method can be obtained by using speech material with sufficiently long voiced speech section. In addition, it was found that the variation in the estimation result due to the difference in the material can be reduced by normalization using the occurrence frequency of the vowel.
© Copyright 2017 All Rights Reserved