Hello, I am a computer engineering student and compression hobbyst and have recently developed a new entropy coding algorithm. It can be used to achieve compression proper of arithmetic coders at very high speeds -almost like Huffman codecs-. Since it could be of interest to you, I send it as an attachment -code and technical report-. Please, drop me a line in case you have any doubt or suggestion. Regards, Eduardo Enrique González Rodríguez. -------------- next part -------------- A non-text attachment was scrubbed... Name: rf-coder.tar.gz Type: application/x-tgz Size: 135839 bytes Desc: rf-coder.tar.gz Url : http://lists.xiph.org/pipermail/vorbis-dev/attachments/20031010/ef58c991/rf-coder.tar-0001.bin
hi ! Eduardo Enrique González Rodríguez wrote:>Hello, I am a computer engineering student and compression hobbyst >and have recently developed a new entropy coding algorithm. It can be >used to achieve compression proper of arithmetic coders at very high >speeds -almost like Huffman codecs-. > > Since it could be of interest to you, I send it as an attachment -code >and technical report-. Please, drop me a line in case you have any doubt or >suggestion. > > Regards, Eduardo Enrique González Rodríguez. > >oh, i get it now :-) you just proposed this to comp.compression a few days ago. i haven't looked at it closely yet but it seemed like there is some potential use. i'm still new to Vorbis, so i can't say, but i'm in the same "wagon" as you, that is : playing with entropy coders. i'm writing a few articles about that now. Consider doing this in your local journal as well. YG --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Eduardo Enrique González Rodríguez wrote:> Hello, I am a computer engineering student and compression hobbyst > and have recently developed a new entropy coding algorithm. It can be > used to achieve compression proper of arithmetic coders at very high > speeds -almost like Huffman codecs-. > > Since it could be of interest to you, I send it as an attachment -code > and technical report-. Please, drop me a line in case you have any doubt or > suggestion.Very interesting approach. Yeah, I can see why this achieves compression ratios very near arithmetic coding. I also can see why it is about the same speed as Huffman codes. Very very interesting. I have some suggestions to make, but first I'll hack up rehuff (my Vorbis stream recoding tool; it replaces the standard Huffman codes which the libvorbis encoder uses (which are optimal _statistically_ for the given encoding settings) with Huffman codes which actually are optimally _for the given stream_) to output some statistics on how well your algorithm will perform (and also how well it will perform with an obvious modification, that should prove to help a bit). Now, I wonder whether anybody already thought of this (and patented it). It seems pretty darn obvious if you look at it ;-) <p>Segher p.s. For those wondering, I soon will package up rehuff, and put a user interface on it etc. It will be GPL. I just need to find the time (and incentive) to do it. Fanmail wouldn't hurt... (doh) --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.