Simulation of vector trellis encoding systems
Abstract
Most tree or trellis encoding data compression systems use decoders which form scalar outputs as a (possibily nonlinear) function of the contents of shift register containing received channel symbols. There, design algorithms are developed for trellis encoding systems having decoders not constrained to have such a scalar slidingblock code structure. In particular, use or the decoder of a finitestate vector quantizer together with a vector trellis search is considered. Simulation results are presentved for vector trellis encoding systems for GaussMarkov sources, sampled speed data, and LPC speech data.
 Publication:

IEEE Transactions on Communications
 Pub Date:
 March 1986
 Bibcode:
 1986ITCom..34..214B
 Keywords:

 Signal Encoding;
 Systems Simulation;
 Turing Machines;
 Vectors (Mathematics);
 GaussMarkov Theorem;
 Histograms;
 Searching;
 Speech;
 Trees (Mathematics);
 Waveforms;
 Communications and Radar