Reverberation model-based decoding in the logmelspec domain for robust distant-talking speech recognition

Journal article
(Original article)


Publication Details

Author(s): Sehr A, Maas R, Kellermann W
Journal: IEEE Transactions on Audio Speech and Language Processing
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Publication year: 2010
Volume: 18
Journal issue: 7
Pages range: 1676-1691
ISSN: 1558-7916


Abstract


The REMOS (REverberation MOdeling for Speech recognition) concept for reverberation-robust distant-talking speech recognition, introduced in Distant-talking continuous speech recognition based on a novel reverberation model in the feature domain (A. Sehr , in Proc. Interspeech, 2006, pp. 769772) for melspectral features, is extended to logarithmic melspectral (logmelspec) features in this contribution. Thus, the favorable properties of REMOS, including its high flexibility with respect to changing reverberation conditions, become available in the more competitive logmelspec domain. Based on a combined acoustic model consisting of a hidden Markov model (HMM) network and a reverberation model (RM), REMOS determines clean-speech and reverberation estimates during recognition. Therefore, in each iteration of a modified Viterbi algorithm, an inner optimization operation maximizes the joint density of the current HMM output and the RM output subject to the constraint that their combination is equal to the current reverberant observation. Since the combination operation in the logmelspec domain is nonlinear, numerical methods appear necessary for solving the constrained inner optimization problem. A novel reformulation of the constraint, which allows for an efficient solution by nonlinear optimization algorithms, is derived in this paper so that a practicable implementation of REMOS for logmelspec features becomes possible. An in-depth analysis of this REMOS implementation investigates the statistical properties of its reverberation estimates and thus derives possibilities for further improving the performance of REMOS. Connected digit recognition experiments show that the proposed REMOS version in the logmelspec domain significantly outperforms the melspec version. While the proposed RMs with parameters estimated by straightforward training for a given room are robust to a mismatch of the speakermicrophone distance, their performance significantly decreases if they are used in a room with substantially different conditions. However, by training multi-style RMs with data from several rooms, good performance can be achieved across different rooms. © 2010 IEEE.



FAU Authors / FAU Editors

Kellermann, Walter Prof. Dr.-Ing.
Professur für Nachrichtentechnik
Maas, Roland Dr.-Ing.
Lehrstuhl für Multimediakommunikation und Signalverarbeitung
Sehr, Armin Dr.-Ing.
Professur für Nachrichtentechnik


How to cite

APA:
Sehr, A., Maas, R., & Kellermann, W. (2010). Reverberation model-based decoding in the logmelspec domain for robust distant-talking speech recognition. IEEE Transactions on Audio Speech and Language Processing, 18(7), 1676-1691. https://dx.doi.org/10.1109/TASL.2010.2050511

MLA:
Sehr, Armin, Roland Maas, and Walter Kellermann. "Reverberation model-based decoding in the logmelspec domain for robust distant-talking speech recognition." IEEE Transactions on Audio Speech and Language Processing 18.7 (2010): 1676-1691.

BibTeX: 

Last updated on 2018-10-08 at 09:08