Most tree or trellis encoding data compression systems use decoders which form scalar outputs as a (possibly nonlinear) function of the contents of a shift register containing received channel symbols. We here develop design algorithms for trellis encoding systems having decoders not constrained to have such a scalar sliding-block code structure. In particular, we consider using the decoder of a finite-state vector quantizer together with a vector trellis search. Simulation results are presented for vector trellis encoding systems for Gauss-Markov sources, sampled speech data, and LPC speech data.

In this note we present a very simple method for improving the efficiency of minimum distortion encoding for vector quantization. Simulations indicates a reduction of up to 70 percent in the number of multiplications for a full search vector quantizer with a large number of codewords, and about 25-40 percent for a tree search vector quantizer. Similar improvement can be achieved in other vector quantization systems.