Model-based dereverberation of speech in the mel-spectral domain

Sehr, Armin; Kellermann, Walter

doi:10.1109/acssc.2008.5074516

Cited by 1 publication

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a last step, we adapted the ASR system to the dereverberated elderly speech using the R [1,2,3]-o data dereverberated with the method described in this work, resulting in R [1,2,3]…”

Section: Combined Acoustics and Age Adaptationmentioning

confidence: 99%

“…In a second step, we adapted the ASR system to (de)reverberated and elderly speech at the same time using the original room microphone data of the TEST data (labeled R [1,2,3]-o) to see if only adaptation without dereverberation could do the job. Note that we used data of R1 to test on R1, R2 to test on R2 and so on.…”

Section: Combined Acoustics and Age Adaptationmentioning

confidence: 99%

“…Beginning top-down, reverberation can be integrated into the actual decoding of the feature sequence. In [1], the authors use spectral features in combination with hidden Markov models (HMM) for continuous digit recognition. By modeling the reverberation as an additive component to the mean values of the output distributions, the authors could show a significant improvement in terms of recognition performance.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Clap your hands! Calibrating spectral subtraction for dereverberation

Zah

Riedhammer

Nöth

2010

2010 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

Reverberation effects as observed by room microphones severely degrade the performance of automatic speech recognition systems. We investigate the use of dereverberation by spectral subtraction as proposed by Lebart and Boucher and introduce a simple approach to estimate the required decay parameter by clapping hands. Experiments on small vocabulary continuous speech recognition task on read speech show that using the calibrated dereverberation improves WER from 73.2 to 54.7 for the best microphone. In combination with system adaptation, the WER could be reduced to 28.2, which is only a 16% relative loss of performance comparison to using a headset instead of a room microphone.

show abstract

“…In a last step, we adapted the ASR system to the dereverberated elderly speech using the R [1,2,3]-o data dereverberated with the method described in this work, resulting in R [1,2,3]…”

Section: Combined Acoustics and Age Adaptationmentioning

confidence: 99%

Section: Combined Acoustics and Age Adaptationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation