Single Channel Speech Dereverberation based on Spectral Subtraction
Speech signals recorded with a distant microphone usu- ally contain reverberation, which degrades the ¯delity and intelligibility of speech in devices such as hands{free conference telephones, auto- matic speech recognition and hearing aids. One important e ect of re- verberation on speech is overlap{masking, i.e. the energy of the previous phonemes is smeared over time, and overlaps following phonemes. In a single{channel speech dereverberation method based on Spectral Subtraction was introduced to reduce this e ect. The described method estimates the power spectrum of the reverberation based on a statistical model of late reverberation. This model depends on one parameter, the reverberation time. However, the reverberation time is frequency depen- dent due to frequency dependent re°ection coe cients of walls and other objects and the frequency dependent absorption coe cient of air. In this paper, we have taken this dependency into account and studied the e ect on reverberation reduction and distortion. The algorithm is tested using synthetically reverberated signals. The performances for di erent room impulse responses with reverberation times ranging from approximately 200 to 1200 ms show signi¯cant reverberation reduction with little signal distortion. Click here for free
download this paper
Related
- Two-Stage Algorithm for One-Microphone Reverberant Speech Enhancement
- SPECTRAL VOICE CONVERSION FOR TEXT-TO-SPEECH SYNTHESIS
- Antenna subtraction at NNLO with hadronic initial states: initial-final configurations
- Speech Recognition System
- Calibration of a Digital Camera
Robust Voice Activity Detection for Interview Speech in NIST Speaker Recognition Evaluation
Two-Stage Algorithm for One-Microphone Reverberant Speech Enhancement CSE PROJECTS