FAST BINARY SEARCH ALGORITHM FOR TRELLIS-CODED QUANTIZATION
Original Publication Date: 2000-Jan-01
Included in the Prior Art Database: 2002-Sep-12
A new codebook search algorithm was devel- Recall that for encoding a memoryless source at oped for trellis-coded quantization that enables R bits/per sample, trellis-coded quantization (TCQ) either fixed-rate codebooks or entropy-constrained [l] uses a codebook of size 2R+1 that is partitioned codebooks to be searched proportional to lograN, into four subsets, each containing 2R-1 codewords. where N is the number of codewords. These subsets are labeled Da, D,, D,, and D,, and are used as labels for the branches of a suitably cho- This is in contrast to traditional trellis-coded sen trellis. quantization, where the search is proportional to N.