Evaluating long-term spectral subtraction for reverberant ASR
Tóm tắt
Even a modest degree of room reverberation can greatly increase the difficulty of automatic speech recognition. We have observed large increases in speech recognition word error rates when using a far-field (3-6 feet) microphone in a conference room, in comparison with recordings from head-mounted microphones. In this paper, we describe experiments with a proposed remedy based on the subtraction of an estimate of the log spectrum from a long-term (e.g., 2 s) analysis window, followed by overlap-add resynthesis. Since the technique is essentially one of enhancement, the processed signal it generates can be used as input for complete speech recognition systems. Here we report results with both the HTK and the SRI Hub-5 recognizer. For simpler recognizer configurations and/or moderate-sized training, the improvements are huge, while moderate improvements are still observed for more complex configurations under a number of conditions.
Từ khóa
#Automatic speech recognition #Spectral analysis #Reverberation #Speech recognition #Microwave integrated circuits #Cepstral analysis #Computer science #Error analysis #Absorption #Fourier transformsTài liệu tham khảo
hirsch, 2000, The AURORA Experimental Framework for the Performance Evaluations of Speech Recognition Systems under Noisy Conditions, ISCA ITRW ASR2000
10.1109/ICASSP.1984.1172716
0, NIST Speech Quality Assurance Package Version 2.3
stolcke, 2000, The SRI March 2000 Hub-5 Conversational Speech Transcription System, NIST Speech Transcription Workshop
0, The ICSI Meeting Recorder Project
1997, CALLHOME American English Speech, Linguistic Data Consortium
10.1109/ICASSP.1992.225858
avendano, 1997, Multiresolution Channel Normalization for ASR in Reverberant Environments, Eurospeech 1997
bernstein, 1994, Macro-phone: An American English telephone speech corpus for the Polyphone project, ICASSP 1994
avendano, 1997, Temporal processing of speech in a time-feature space