High-performance robust speech recognition using stereo training data

We describe a novel technique of SPLICE for highperformance robust speech recognition. It is an efficient noise reduction and channel distortion compensation technique that makes effective use of stereo training data. In this paper, we present a version of SPLICE using the minimum-meansquareerror decision, and describe an extension by training clusters of HMMs with SPLICE processing. Comprehensive results using a Wall Street Journal large vocabulary recognition task and with a wide range of noise types demonstrate superior performance of the SPLICE technique over that under noisy matched conditions (13% word error rate reduction). The new technique is also shown to consistently outperform the spectralsubtraction and the fixed CDCN noise reduction techniques. It is currently being integrated into the Microsoft MiPad, a new generation PDA prototype.