Probability Distribution of Time-Series of Speech Spectral Components
Summary :
This paper deals with the statistical modeling of a Time-Frequency Series of Speech (TFSS), obtained by Short-Time Fourier Transform (STFT) analysis of the speech signal picked up by a linear microphone array with two elements. We have attempted to find closer match between the distribution of the TFSS and theoretical distributions like Laplacian Distribution (LD), Gaussian Distribution (GD) and Generalized Gaussian Distribution (GGD) with parameters estimated from the TFSS data. It has been found that GGD provides the best models for real part, imaginary part and polar magnitudes of the time-series of the spectral components. The distribution of the polar magnitude is closer to LD than that of the real and imaginary parts. The distributions of the real and imaginary parts of TFSS correspond to strongly LD. The phase of the TFSS has been found uniformly distributed. The use of GGD based model as PDF in the fixed-point Frequency Domain Independent Component Analysis (FDICA) provides better separation performance and improves convergence speed significantly.
- Publication
- IEICE TRANSACTIONS on Fundamentals Vol.E87-A No.3 pp.584-597
- Publication Date
- 2004/03/01
- Publicized
- Online ISSN
- DOI
- Type of Manuscript
- Special Section PAPER (Special Section on Applications and Implementations of Digital Signal Processing)
- Category
- Audio/Speech Coding
Authors
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Rajkishore PRASAD, Hiroshi SARUWATARI, Kiyohiro SHIKANO, "Probability Distribution of Time-Series of Speech Spectral Components" in IEICE TRANSACTIONS on Fundamentals,
vol. E87-A, no. 3, pp. 584-597, March 2004, doi: .
Abstract: This paper deals with the statistical modeling of a Time-Frequency Series of Speech (TFSS), obtained by Short-Time Fourier Transform (STFT) analysis of the speech signal picked up by a linear microphone array with two elements. We have attempted to find closer match between the distribution of the TFSS and theoretical distributions like Laplacian Distribution (LD), Gaussian Distribution (GD) and Generalized Gaussian Distribution (GGD) with parameters estimated from the TFSS data. It has been found that GGD provides the best models for real part, imaginary part and polar magnitudes of the time-series of the spectral components. The distribution of the polar magnitude is closer to LD than that of the real and imaginary parts. The distributions of the real and imaginary parts of TFSS correspond to strongly LD. The phase of the TFSS has been found uniformly distributed. The use of GGD based model as PDF in the fixed-point Frequency Domain Independent Component Analysis (FDICA) provides better separation performance and improves convergence speed significantly.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e87-a_3_584/_p
Copy
@ARTICLE{e87-a_3_584,
author={Rajkishore PRASAD, Hiroshi SARUWATARI, Kiyohiro SHIKANO, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Probability Distribution of Time-Series of Speech Spectral Components},
year={2004},
volume={E87-A},
number={3},
pages={584-597},
abstract={This paper deals with the statistical modeling of a Time-Frequency Series of Speech (TFSS), obtained by Short-Time Fourier Transform (STFT) analysis of the speech signal picked up by a linear microphone array with two elements. We have attempted to find closer match between the distribution of the TFSS and theoretical distributions like Laplacian Distribution (LD), Gaussian Distribution (GD) and Generalized Gaussian Distribution (GGD) with parameters estimated from the TFSS data. It has been found that GGD provides the best models for real part, imaginary part and polar magnitudes of the time-series of the spectral components. The distribution of the polar magnitude is closer to LD than that of the real and imaginary parts. The distributions of the real and imaginary parts of TFSS correspond to strongly LD. The phase of the TFSS has been found uniformly distributed. The use of GGD based model as PDF in the fixed-point Frequency Domain Independent Component Analysis (FDICA) provides better separation performance and improves convergence speed significantly.},
keywords={},
doi={},
ISSN={},
month={March},}
Copy
TY - JOUR
TI - Probability Distribution of Time-Series of Speech Spectral Components
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 584
EP - 597
AU - Rajkishore PRASAD
AU - Hiroshi SARUWATARI
AU - Kiyohiro SHIKANO
PY - 2004
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E87-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2004
AB - This paper deals with the statistical modeling of a Time-Frequency Series of Speech (TFSS), obtained by Short-Time Fourier Transform (STFT) analysis of the speech signal picked up by a linear microphone array with two elements. We have attempted to find closer match between the distribution of the TFSS and theoretical distributions like Laplacian Distribution (LD), Gaussian Distribution (GD) and Generalized Gaussian Distribution (GGD) with parameters estimated from the TFSS data. It has been found that GGD provides the best models for real part, imaginary part and polar magnitudes of the time-series of the spectral components. The distribution of the polar magnitude is closer to LD than that of the real and imaginary parts. The distributions of the real and imaginary parts of TFSS correspond to strongly LD. The phase of the TFSS has been found uniformly distributed. The use of GGD based model as PDF in the fixed-point Frequency Domain Independent Component Analysis (FDICA) provides better separation performance and improves convergence speed significantly.
ER -
English
