# Compression of Memoryless Sources

Are there any compression algorithms outside of entropy encodings that yield a significant compression ratio for memoryless sources (not the Markov kind, but this kind)?

More concretely, if a million ASCII characters, each generated from the same probability distribution that is not necessarily even but where the probability of each ASCII character being chosen is non-zero, what compression algorithms outside of entropy encoders (such as arithmetic, huffman) would yield decent compression ratios?

• Why have you rejected entropy encoders? – D.W. Jun 16 '17 at 21:28