Predictive Trellis Coded Quantization of Speech

Trellis coded quantization (TCQ) is incorporated into a predictive coding structure for encoding sampled speech. The modest complexity of the resulting structure is seen to be a direct consequence of the TCQ formulation. Simulation results are presented for systems using fixed prediction/fixed residual encoding, fixed prediction/adaptive residual encoding, and adaptive prediction/adaptive residual encoding. The performance of predictive trellis coded quantization (PTCQ) is compared to that of other waveform coders, and the effects of channel errors on PTCQ performance are discussed. For a fully adaptive 16 kbit/s speech coding system, segmental signal-to-noise ratios in the range of 19.1–21.9 dB are obtained for a variety of speakers and test sentences. Reconstructed speech obtained from this system can be described as being of excellent communications quality.