In practice with context 0.1 we have some files which are almost monomorphic. But as soon as they have a single field with a different value then we have to use one bit per field for all of the fields.
If we used entropy codes like ANS we could get more efficient encodings than Huffman in these cases.
In practice with context 0.1 we have some files which are almost monomorphic. But as soon as they have a single field with a different value then we have to use one bit per field for all of the fields.
If we used entropy codes like ANS we could get more efficient encodings than Huffman in these cases.