Talk:Adaptive coding

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

This article needs serious work. It is improperly restricted to arithmetic coding when it is a far more general technique, and it is (in my judgement) almost incomprehensible to anyone who does not already understand what adaptive coding is. -- Antaeus Feldspar 05:30, 22 Sep 2004 (UTC)

Adaptive variants[edit]

There seems to be an impression that some entropy coding methods are adaptive, and some are static. This is not the case. Any method has static variants and adaptive variants. For example, Huffman coding can use a data model based on English text that is agreed upon by encoder and decoder, so not transmitted. Or it can make a first pass over the data to create a model and transmit that. These are both static variants. Huffman coding can also be adaptive, with many variations on how to adapt the data model.

The difficulty with the term "adaptive coding" is that it encompasses so many particular methods, so the description should be general, especially as a contrast with static coding. But describing a particular method might also be illustrative. For instance, a technique for creating an adaptive Huffman tree would be useful. The danger is that the impression might be left that it is the only way to implement adaptive Huffman coding.

Scottcraig 20:13, 15 Nov 2004 (UTC)