On the Statistical Properties of Reverberant Speech Feature Vector Sequences
Sehr A, Kellermann W (2010)
Publication Language: English
Publication Type: Conference contribution
Publication year: 2010
City/Town: Tel Aviv, Israel
Conference Proceedings Title: Proc. International Workshop on Acoustic Echo and Noise Control (IWAENC)
Event location: Tel Aviv
Abstract
The statistical properties of reverberant logarithmic mel-spectral feature vector sequences that are relevant for acoustic mod-eling in robust speech recognition are analyzed and compared to the corresponding properties of clean-speech feature vector sequences. The investigation focuses on probability densities of feature vector elements and the correlation between feature vectors of different frames. A Monte-Carlo method is used for a quantitative analysis of the density changes and the increase in inter-frame correlation due to reverberation. As an example for the insights that can be obtained from this analysis, the densities and inter-frame correlations of reverberant features are compared to the modeling capabilities of reverberantly-trained and adapted HMMs. Thus, limitations of these approaches can be clearly identified.
Authors with CRIS profile
How to cite
APA:
Sehr, A., & Kellermann, W. (2010). On the Statistical Properties of Reverberant Speech Feature Vector Sequences. In Proc. International Workshop on Acoustic Echo and Noise Control (IWAENC). Tel Aviv, IL: Tel Aviv, Israel.
MLA:
Sehr, Armin, and Walter Kellermann. "On the Statistical Properties of Reverberant Speech Feature Vector Sequences." Proceedings of the Proc. International Workshop on Acoustic Echo and Noise Control (IWAENC), Tel Aviv Tel Aviv, Israel, 2010.
BibTeX: Download