Generalized predictive TCQ of speech
Trellis coded quantization (TCQ) is incorporated into a noise feedback coding structure for encoding sampled speech. The effects of varying the encoding delay and the number of symbols released per trace-back on system performance and complexity are investigated.