Skip to Main Content
In the traditional characteristic waveform interpolation speech coding, the high frequency harmonics of the synthesized speech are usually lost, due to the phase information neglect of the characteristic waveform in the encoding section and its whole waveform alignment in the decoding section. In order to obtain a higher quality synthesized speech, the multiband markers are first introduced. The phase spectrums of the slowly and rapidly evolving waveform are estimated depending on them. The part characteristic waveform alignment is developed in the speech synthesis section. Finally, a 2.4 kbps multiband characteristic waveform interpolation speech coding algorithm is proposed. The result shows that the new algorithm can rebuild the speech high frequency harmonics perfectly. In the speech subjective quality test, the new coder is obviously superior to the conventional CWI coder and also slightly exceeds the MELP coder.