Automatyczna ekstrakcja i klasteryzacja głosek w sygnale mowy dla wielojęzykowej analizy porównawczej
Automatic extraction and clusterization of phones from speech signal for multilingual comparative analysis

Summary/Abstract: The paper presents the possibility of automatic speech processing in order to determine the acoustic similarity between phones. Subsequent processing steps of recorded speech signal result in phones’ segmentation, even without prior knowledge of their boundaries. The use of frequency signal parameterization and clustering algorithms facilitates a division of phones into groups (clusters) on the basis of their acoustic similarities. The paper also presents the results of experiments on speech recordings during which the automatically created groups of phones have been compared with the manually annotated Polish phonemes.

• Details
• Contents