METHODS FOR SPEECH QUANTIZATION AND ERROR CORRECTION

摘要

Method of enhancing speech of a speech signal broken into segments. Frequency domain representations of a segment of the speech signal are determined to provide a spectral envelope of the segment. Speech is synthesized from an enhanced spectral envelope. A smoothed spectral envelope of the segment is generated by smoothing the spectral envelope. An enhanced spectral envelope is generated by increasing some frequency regions of the spectral envelope for which the spectral envelope has a greater amplitude than the smoothed envelope and decreasing some frequency regions in which the spectral envelope has lesser amplitude than the smoothed envelope.