ImprovingReverberantVTSforHands-freeRobustSpeechRecognitionY.-Q.Wang,M - PPT Presentation
writtenasytClogexpC1xthexpC1ntfxthnt2wherentisthenoisecoefcienththeconvolutionalnoiseandCthetruncatedDCTmatrixNotethatthismismatchfunctionassumesthenoiseandspeecharelinearl ID: 158645