Přístupnostní navigace
E-application
Search Search Close
Publication detail
STEJSKAL, V. ESPOSITO, A.
Original Title
Spectral flatness method based empty pause detection in noisy and clean speech conditions
Type
conference paper
Language
English
Original Abstract
Successful pause detection becomes an important part in the process of speech recognition and speech coding as well as in the biometrical field (stress detection) and human-machine interaction. Nowadays, only a few of proposed algorithms are able to reflect various noise conditions. This is considered in presented paper which results from research that has been made at the International Institute for Advanced Scientific Studies (IIASS) and proposes novel method for non-speech activity pause detection in spontaneous speech recordings made in noisy environments. The input signal is transformed into log spectral energy and is divided into specific frequency bands. Each band is smoothed and tracked by dynamically adjusted thresholds based on noise energy estimation. Thresholds are adapted taking into account the dynamic changes of the speech signal under environmental noise. The proposed method run in real time and does not require a priori knowledge of the SNR and a priori threshold values. Experimental results show that its efficacy is comparable with standard VADs.
Keywords
speech, pause detection, noise
Key words in English
Authors
STEJSKAL, V.; ESPOSITO, A.
RIV year
2006
Released
1. 1. 2006
Publisher
Institute of Radio Engineering and Electronics AS CR
Location
Praha
ISBN
8086269159
Book
16th Czech-German Workshop on Speech Processing
Pages from
125
Pages to
254
Pages count
130
BibTex
@inproceedings{BUT24827, author="Vojtěch {Stejskal} and Anna {Esposito}", title="Spectral flatness method based empty pause detection in noisy and clean speech conditions", booktitle="16th Czech-German Workshop on Speech Processing", year="2006", pages="130", publisher="Institute of Radio Engineering and Electronics AS CR", address="Praha", isbn="8086269159" }