Interesting newsgroup or forum posts, mostly to comp.compression and comp.compression.research. Note that nearly all of these newsgroup posts are retrieved via the groups.google (former Deja News) service.
Alternative to arithmetic coding - instead of dividing a range into two subranges, we distribute them uniformally over the range. See http://cs.fit.edu/~mmahoney/compression/#fpaq0 for implementations.
Created: 20/11/2007
More...
This forum thread describes an experimental high quality wavelet videocodec named "Snow". It is being developed by FFmpeg team. There is a 3rd party encoding GUI for the codec.
Created: 26/02/2006
by Maxim SmirnovMore...
There is a news feed and a useful list of links on lossless audio compression, processing and playback on this site.
Created: 11/02/2006
by Maxim SmirnovMore...
It is a very good site to know about audio technology and ask your questions. The site is organized as a forum, it has comprehensive FAQ.
Created: 11/02/2006
by Maxim SmirnovMore...
ACB is poorly understood. The author, George Buynovsky, has only made a few public comments on the algorithm. This is one of them.
Created: 06/01/2000
by Mark NelsonMore...
A careful analysis of degenerate Huffman trees created when input counts follow a Fibonacci sequence, by Alex Vinokur. This post is from sci.math.
Created: 14/02/2002
by Mark NelsonMore...
The first time I ever heard the phrase Canonical Huffman Coder was in reference to the technique used in PKZip to store Huffman tables. I don't know where the technique originated, but it is basically a way to construct a Huffman table so that the actual codes don't have to be stored when storing the table. This makes for some nice space savings when compared to a first-pass naive implementation. (Like the ones I've done in the past.)
It turns out that somebody named Gareth was attempting to implement this code but was having a bit of trouble. His post to comp.compression brought out some useful help from some of the newsgroup regulars, and did a lot to shed light on the topic, and includes a reference to the paper that actually gave birth to the concept.
Created: 13/03/2003
by Mark NelsonMore...
Somebody posting as thewhizkid complained to the comp.compression newsgroup that he just couldn't figure out how to do a CRC calculation. He got a couple of good pointers here, including one to the zlib code, and another including a bit of Java that could do the job. I'll add a pointer to my ancient-but-still-cogent article from DDJ,
File Verification Using CRC to the mix. Between the three choices, I hope the poster was able to get a handle on the mysterious CRC.
Created: 12/03/2003
by Mark NelsonMore...
Alex Ng posed the question on comp.compression: "In general, you cannot have one image coder can compress well for ALL
images. Why?"
This set off a lively and somewhat useful discussion, with lots of thoughtful posts from Thomas Richter and others. Always good to find some intelligent thought on the compression newsgroup.
Created: 07/03/2003
by Mark NelsonMore...
According to Pete Fraser, you can download three free specifications from the ITU per year, after free registration. This is a great deal - these specs can run into the hundreds of dollars per copy, and if you aren't near a university library, you may have no other way to get them.
Created: 15/09/2002
by Mark NelsonMore...
This thread reveals the actual dates for the LZW patent expiration, including US, European, and Japanese patents.
Created: 15/09/2002
by Mark NelsonMore...
A forum dedicated to the discussion of Wavelets. This link takes you to the French language version of the board - click on the "English" link on the left side of the page for English language discussion.
Created: 05/06/2002
by Mark NelsonMore...
A post by Edgar Binder discussing his distance coding algorithm. Distance coding is used in place of the MTF stage in BWT compression.
Created: 03/06/2002
by Mark NelsonMore...
C++ code posted to comp.compression that describes extraction from PKWare's Data Compression Library.
Update: The author posted this correction to comp.compression:
There's a bug in the code posted 2001-10-07 19:36:38 PST. To fix:
In the
void tcDecoder::Decode(char *apBuffer, unsigned int *apSize, unsigned int anBufferSize)
function after both
if (lnIndex == mnCurrentPos) lnIndex = lnStartIndex;
add
if (lnIndex == mnDictionarySize) lnIndex = 0;
Created: 22/04/2002
by Mark NelsonMore...
This site is mostly oriented towards capturing and storing DVD - good advice on how to use various tools, codecs, hardware, etc. Very active site.
Created: 19/04/2002
by Mark NelsonMore...
The folks at Slashdot are having a chat trying to decide what the best non-MP3 codec is for home music recording. Ogg Vorbis gets a lot of votes the last time I checked the thread, but WMA and MP3 have a fair number of adherents.
Created: 22/01/2002
by Mark NelsonMore...
Jules Gilbert has made a number of interesting and provocative posts to comp.compression. His claims have been met with vocal objections, but lack of public resolution will hamper the interested researcher.
Created: 05/12/2001
by Mark NelsonMore...
Teddy Turner, son of CNN mogul Ted Turner, was apparently named in a lawsuit in 1999 over an alleged misrepresentation of compression claims. A Florida company named Vision Tek claimed it could pump video over normal phone lines at high speeds, but was not able to satisy its investors, leading to a court case.
Created: 26/10/2001
by Mark NelsonMore...
The comp.compression newsgroup regularly sees posts from folks with incredible compression claims. This one, started by Itai Bar-Haim, is remarkable not so much for its claims, but rather for the incredible length of the thread it spawned. All Atai said was that mathematical combinations have nothing to do with compression. So there.
Created: 27/06/2000
by Mark NelsonMore...
Charles Bloom puts the noggin to work on this question. It's not as foolish as the title might make it sound.
Created: 06/01/2000
by Mark NelsonMore...
Charles Bloom works up a quick coder that will compress random bytes down to about 6 bits per byte. Quite a feat, but it doesn't stand up to close scrutiny.
Created: 06/01/2000
by Mark NelsonMore...
This newsgroup doesn't generate as much compression as its sibling, comp.compression. I know of nothing else that differentiates the two.
Created: 07/11/1999
by Mark NelsonMore...
This newgroup carries a good deal of interesting traffic on data compression. The lack of moderation means the noise level is still fairly high, but the interested reader can find much to enjoy here.
Created: 07/11/1999
by Mark NelsonMore...
This article contains the answers to some Frequently Asked Questions (FAQ) often seen in the USENET newsgroup comp.dcom.fax relating to facsimile standards, software, and hardware. Most importantly, contains an explanation of FAX encoding.
Created: 30/01/1999
by Mark NelsonMore...
"I do believe that my product is superior to all other compression algorithms to date. It is lossless, works on any file type on any platform, is recursive (ie, the compressed file can be compressed again, and then again, and so on) so that ANY size file can be compressed to 16 bytes or less. I know this sounds like "perpetual motion" but it does work and I can prove it."
Created: 11/01/1999
by Mark NelsonMore...
