* They usually self synchronize when some data is corrupted (but not guaranteed, does not apply where the Huffman table is dynamic)
* Neither Huffman nor arithmetic codes are easy to parallelize the decoding of, but Huffman is slightly easier.
* They usually self synchronize when some data is corrupted (but not guaranteed, does not apply where the Huffman table is dynamic)
* Neither Huffman nor arithmetic codes are easy to parallelize the decoding of, but Huffman is slightly easier.