Title :
A frequency-weighted Itakura spectral distortion measure and its application to speech recognition in noise
Author :
Soong, Frank K. ; Sondhi, Man Mohan
Author_Institution :
AT&T Bell Labs., Murray Hill, NJ, USA
fDate :
1/1/1988 12:00:00 AM
Abstract :
The authors propose an adaptively weighted Itakura distortion measure. They studied its effects on the performance of a conventional dynamic time-warping (DTW)-based speech recognizer in a series of speaker-independent, isolated-digit-recognition experiments. The equivalent SNR improvement achieved by using the proposed weighted Itakura distortion at low SNRs is about 5-7 dB
Keywords :
electric distortion measurement; noise; spectral analysis; speech recognition; dynamic time-warping-based speech recognition; frequency-weighted Itakura spectral distortion measure; isolated-digit-recognition experiments; noise; speaker independent speech recognition; speech recognition; Acoustic distortion; Distortion measurement; Frequency measurement; Gain measurement; Linear predictive coding; Noise measurement; Predictive models; Signal to noise ratio; Speech recognition; Testing;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
