Kevin Jenkins wrote: >I'm currently using the huffman encoding algorithm for compression in my >networking engine. However, I'd like something that compresses a bit >better. I read about a lossless compression algorithm that used floating >point that was supposed to be even better but can't remember. > >Anyone knows what this is? > You're probably thinking about arithmetic coding (but don't expect too much from it when compared to huffman). If you want to improve compresison significantly, you might also want to use a first pass to encode redundancies (like RLE or lempel-ziv or ...). Stephane --------------------- To unsubscribe go to http://gameprogrammer.com/mailinglist.html