Data Compression - Systematisation
---- Precoding ----
by T.Strutz
Figure 4: 
Precoding comprises all techniques, which exploit existing statistical dependencies between symbols. These algorithms map symbols from one alphabet to symbols of another alphabet or support this process, respectively. The number of symbols N may change in doing so (contrary to decorrelation methods, which also dissolve correlations between signal values). Very different algorithms belong to this sub-category: run length coding, phrase coding (better known as dictionary-based coding like LZ77, LZ78 or LZW), block sorting (also known as Burrows-Wheeler Transformation), quad-tree coding and others.
Strutz / 31.08.2007
