I am looking for a good algorithm for bit stream compression (packet payload compression). I would like to avoid algorithms that are based on symbol probability. I have already tried the LZ family algorithms, and found none of them useful, even with BWT.
I am trying to accomplish a minimum compression percentage of 30%, but have only managed 3-5% using RLE.
What is a good algorithm that has a compression above 30%?
If you have no knowledge about your input data, it's hard to achieve good compression (just like a general purpose compressor).
But at least you can try some context-based model. use several prefix bits as context and predict the probability of next bit, then pass the probability to a range coder.
Further compression can be achieved with a context mixing model without byte-alignment. see http://mattmahoney.net/dc/dce.html#Section_43.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With