2000.03.30 09:36 "jpeg2000 question", by Ivo Penzar

2000.03.30 16:46 "RE: jpeg2000 question", by Bill Radcliffe

I worked on the definition of the JP2 file format. The situation with JPEG2000 is not at all like JPEG.

You should know is that JP2 is a very simple wrapper around the formal JPEG2000 "codestream". It was based on the Quicktime idea of "atoms". This simple idea is that you have a tag and a length followed by the "stuff". You know what the "stuff" is based on the tag type. You can also nest these atoms to build more complex structures.

JP2 is a very simple set of these atoms, most of which are optional. You could write a program to implement this in about 15 minutes. The "hard" part of JP2 is the "codestream". This is the encoded JPEG2000 data. Good luck on that. We don't have a freely available reference implementation at this point because nobody has signed up to do it. There was an effort underway by the same group that did the JPEG-LS implementation. This was given to the public domain..

http://spmg.ece.ubc.ca/research/jpeg/jpeg_ls/jpegls.html

They actually had an implementation of JPEG2000 on their website but pulled it off:

http://www.ece.ubc.ca/~mdadams/jasper/

Apparently, they decided to sell out and try to charge for it instead of providing it to the public domain. We all know how well that is going to work. When are people going to learn that you can't make money selling compression software? It's not a business - its just an enabler.

My opinion is that without a "Tom Lane" type signing up to make a free implementation available, JPEG2000 is doomed. You just don't get real interoperability from a paper spec. We know this.

The good news is that the Digital Imaging Group (DIG) has a project underway to do this. The bad news is that it has not yet started. And so we wait :-)