Integration of noise reduction algorithms for Aurora2 task
A PHP Error was encountered
Message: Undefined index: id
Line Number: 218
Abstract To achieve high recognition performance,for a wide variety of noise and for a wide range of signal-to-noise ratios, this pa- per presents the integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. Recognition results on the Aurora2 task show,that the effectiveness of these algo- rithms and their combinations,strongly depends on noise condi- tions, and excessive noise reduction tends to degrade recogni- tion performance,in multicondition training.