Robust Voice Activity Detection Algorithm Based on Feature of Frequency Modulation of Harmonics and Its DSP Implementation

Chung-Chien HSU; Kah-Meng CHEONG; Tai-Shih CHI; Yu TSAO

doi:10.1587/transinf.2015EDP7138

Robust Voice Activity Detection Algorithm Based on Feature of Frequency Modulation of Harmonics and Its DSP Implementation

Chung-Chien HSU, Kah-Meng CHEONG, Tai-Shih CHI, Yu TSAO

Full Text Views

0

Cite this

Summary :

This paper proposes a voice activity detection (VAD) algorithm based on an energy related feature of the frequency modulation of harmonics. A multi-resolution spectro-temporal analysis framework, which was developed to extract texture features of the audio signal from its Fourier spectrogram, is used to extract frequency modulation features of the speech signal. The proposed algorithm labels the voice active segments of the speech signal by comparing the energy related feature of the frequency modulation of harmonics with a threshold. Then, the proposed VAD is implemented on one of Texas Instruments (TI) digital signal processor (DSP) platforms for real-time operation. Simulations conducted on the DSP platform demonstrate the proposed VAD performs significantly better than three standard VADs, ITU-T G.729B, ETSI AMR1 and AMR2, in non-stationary noise in terms of the receiver operating characteristic (ROC) curves and the recognition rates from a practical distributed speech recognition (DSR) system.

Publication: IEICE TRANSACTIONS on Information Vol.E98-D No.10 pp.1808-1817

Publication Date: 2015/10/01

Publicized: 2015/07/10

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2015EDP7138

Type of Manuscript: PAPER

Category: Speech and Hearing

Authors

Chung-Chien HSU
  National Chiao Tung University
Kah-Meng CHEONG
  National Chiao Tung University
Tai-Shih CHI
  National Chiao Tung University
Yu TSAO
  Academia Sinica

Keyword

digital signal processor, frequency modulation, spectro-temporal analysis, voice activity detection

Cite this

Copy

Chung-Chien HSU, Kah-Meng CHEONG, Tai-Shih CHI, Yu TSAO, "Robust Voice Activity Detection Algorithm Based on Feature of Frequency Modulation of Harmonics and Its DSP Implementation" in IEICE TRANSACTIONS on Information, vol. E98-D, no. 10, pp. 1808-1817, October 2015, doi: 10.1587/transinf.2015EDP7138.
Abstract: This paper proposes a voice activity detection (VAD) algorithm based on an energy related feature of the frequency modulation of harmonics. A multi-resolution spectro-temporal analysis framework, which was developed to extract texture features of the audio signal from its Fourier spectrogram, is used to extract frequency modulation features of the speech signal. The proposed algorithm labels the voice active segments of the speech signal by comparing the energy related feature of the frequency modulation of harmonics with a threshold. Then, the proposed VAD is implemented on one of Texas Instruments (TI) digital signal processor (DSP) platforms for real-time operation. Simulations conducted on the DSP platform demonstrate the proposed VAD performs significantly better than three standard VADs, ITU-T G.729B, ETSI AMR1 and AMR2, in non-stationary noise in terms of the receiver operating characteristic (ROC) curves and the recognition rates from a practical distributed speech recognition (DSR) system.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2015EDP7138/_p

Copy

@ARTICLE{e98-d_10_1808,
author={Chung-Chien HSU, Kah-Meng CHEONG, Tai-Shih CHI, Yu TSAO, },
journal={IEICE TRANSACTIONS on Information},
title={Robust Voice Activity Detection Algorithm Based on Feature of Frequency Modulation of Harmonics and Its DSP Implementation},
year={2015},
volume={E98-D},
number={10},
pages={1808-1817},
abstract={This paper proposes a voice activity detection (VAD) algorithm based on an energy related feature of the frequency modulation of harmonics. A multi-resolution spectro-temporal analysis framework, which was developed to extract texture features of the audio signal from its Fourier spectrogram, is used to extract frequency modulation features of the speech signal. The proposed algorithm labels the voice active segments of the speech signal by comparing the energy related feature of the frequency modulation of harmonics with a threshold. Then, the proposed VAD is implemented on one of Texas Instruments (TI) digital signal processor (DSP) platforms for real-time operation. Simulations conducted on the DSP platform demonstrate the proposed VAD performs significantly better than three standard VADs, ITU-T G.729B, ETSI AMR1 and AMR2, in non-stationary noise in terms of the receiver operating characteristic (ROC) curves and the recognition rates from a practical distributed speech recognition (DSR) system.},
keywords={},
doi={10.1587/transinf.2015EDP7138},
ISSN={1745-1361},
month={October},}

Copy

TY - JOUR
TI - Robust Voice Activity Detection Algorithm Based on Feature of Frequency Modulation of Harmonics and Its DSP Implementation
T2 - IEICE TRANSACTIONS on Information
SP - 1808
EP - 1817
AU - Chung-Chien HSU
AU - Kah-Meng CHEONG
AU - Tai-Shih CHI
AU - Yu TSAO
PY - 2015
DO - 10.1587/transinf.2015EDP7138
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E98-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2015
AB - This paper proposes a voice activity detection (VAD) algorithm based on an energy related feature of the frequency modulation of harmonics. A multi-resolution spectro-temporal analysis framework, which was developed to extract texture features of the audio signal from its Fourier spectrogram, is used to extract frequency modulation features of the speech signal. The proposed algorithm labels the voice active segments of the speech signal by comparing the energy related feature of the frequency modulation of harmonics with a threshold. Then, the proposed VAD is implemented on one of Texas Instruments (TI) digital signal processor (DSP) platforms for real-time operation. Simulations conducted on the DSP platform demonstrate the proposed VAD performs significantly better than three standard VADs, ITU-T G.729B, ETSI AMR1 and AMR2, in non-stationary noise in terms of the receiver operating characteristic (ROC) curves and the recognition rates from a practical distributed speech recognition (DSR) system.
ER -