Wavelet-based voice activity detection algorithm in variable-level noise environment

Authors:
Kun-Ching Wang
Affiliations:
Department of Information Technology & Communication, Shin Chien University, Neimen Shiang, Kaohsiung, Taiwan
Venue:
WSEAS Transactions on Computers
Year:
2009

Citing 5
Cited 0

Fundamentals of speech recognition

Fundamentals of speech recognition
Assessment for automatic speech recognition II: NOISEX-92: a database and an experiment to study the effect of additive noise on speech recognition systems

Speech Communication - Special issue on speech processing in adverse conditions
Psychoacoustics: Facts and Models

Psychoacoustics: Facts and Models
High-quality audio compression using an adaptive wavelet packetdecomposition and psychoacoustic modeling

IEEE Transactions on Signal Processing
ITU-T Recommendation G.729 Annex B: a silence compression scheme for use with G.729 optimized for V.70 digital simultaneous voice and data applications

IEEE Communications Magazine

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, a novel entropy-based voice activity detection (VAD) algorithm is presented in variable-level noise environment. Since the frequency energy of different types of noise focuses on different frequency subband, the effect of corrupted noise on each frequency subband is different. It is found that the seriously obscured frequency subbands have little word signal information left, and are harmful for detecting voice activity segment (VAS). First, we use bark-scale wavelet decomposition (BSWD) to split the input speech into 24 critical subbands. In order to discard the seriously corrupted frequency subband, a method of adaptive frequency subband extraction (AFSE) is then applied to only use the frequency subband. Next, we propose a measure of entropy defined on the spectrum domain of selected frequency subband to form a robust voice feature parameter. In addition, unvoiced is usually eliminated. An unvoiced detection is also integrated into the system to improve the intelligibility of voice. Experimental results show that the performance of this algorithm is superior to the G729B and other entropy-based VAD especially for variable-level background noise.