A characteristic waveform decomposition algorithm using biorthogonal wavelet transform

This paper presents a characteristic waveform decomposition algorithm based on wavelet transform. The characteristic waveform surface is decomposed into a series of reduced time resolution surfaces. The biorthogonal linear phase filter banks based on B-spline wavelet are performed in the evolutionary waveform domain. This multi-scale characteristic leads to more efficient quantization by exploiting each surface's perceptual importance and inherent transmission rate requirements. The waveform interpolation coder based on wavelet decomposition and reconstruction performs well with efficient quantization and is very suitable for speech storage. Also a 3.3kbps waveform interpolation speech coder is proposed based on the wavelet-based waveform decomposition algorithm. For low delay application, casual, stable IIR filters or low-delay FIR filter designing method can be used.