Title :
Enhancement of connected words in an extremely noisy environment
Author :
Cohen, Yuval ; Erell, Adoram ; Bistritz, Yuval
Author_Institution :
Electr. Eng. Group, RND Networks Ltd., Tel Aviv, Israel
fDate :
3/1/1997 12:00:00 AM
Abstract :
A speech enhancement algorithm that is based on a connected-word hidden Markov model (HMM) is developed. Speech is assumed to be highly degraded by statistically independent additive noise. The minimum mean square error estimator is derived for a connected-word HMM. Further, we derive an estimator based on a connected-word HMM with explicit state duration. Listening experiments performed with digit strings have shown an increase of intelligibility. The best results were achieved when subjects who listened to the enhanced speech were given the results of an automatic recognition system
Keywords :
acoustic noise; hidden Markov models; least mean squares methods; speech enhancement; speech intelligibility; speech recognition; automatic recognition system; connected-word HMM; connected-word hidden Markov model; digit strings; enhanced speech; explicit state duration; extremely noisy environment; intelligibility; listening experiments; minimum mean square error estimator; speech enhancement algorithm; statistically independent additive noise; Additive noise; Automatic speech recognition; Degradation; Hidden Markov models; Mean square error methods; Noise level; Signal to noise ratio; Speech enhancement; Speech recognition; Working environment noise;
Journal_Title :
Speech and Audio Processing, IEEE Transactions on
