"Prominence" is a program representing the results obtained by the author in 5 years of research about the automatic identification of prominent syllables in continuous speech. It has been designed as a plug-in of a very famous open-source software "WaveSurfer" for phonetic-acoustic processing of speech utterances developed at the Royal Institute of Technology - KTH - of Stockholm, (http://www.speech.kth.se/wavesurfer/). Prominence can identify prominent syllables contained into utterances using only information derived by the acoustic component of speech (waveforms), reaching classification performances comparable to expert human annotators. The program can be configured and adapted to different languages modifying the contribution due to the different acoustic parameters considered in the underlying model (overall energy, spectral emphasis, duration and pitch accents). It has been successfully used for identifying prominence in English, Italian and German utterances and the results obtained using this program have been already presented at a number of international conferences and journals in the last years.
Prominence (version 1.0) / Tamburini F.. - ELETTRONICO. - (2007).
Prominence (version 1.0)
TAMBURINI, FABIO
2007
Abstract
"Prominence" is a program representing the results obtained by the author in 5 years of research about the automatic identification of prominent syllables in continuous speech. It has been designed as a plug-in of a very famous open-source software "WaveSurfer" for phonetic-acoustic processing of speech utterances developed at the Royal Institute of Technology - KTH - of Stockholm, (http://www.speech.kth.se/wavesurfer/). Prominence can identify prominent syllables contained into utterances using only information derived by the acoustic component of speech (waveforms), reaching classification performances comparable to expert human annotators. The program can be configured and adapted to different languages modifying the contribution due to the different acoustic parameters considered in the underlying model (overall energy, spectral emphasis, duration and pitch accents). It has been successfully used for identifying prominence in English, Italian and German utterances and the results obtained using this program have been already presented at a number of international conferences and journals in the last years.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.