As a pet project I have been developing a Lossless video codec designed for CGI or cartoonish video, i.e. anime / cartoons etc.
Once i have the lossless codec working i am planning to extend it to include lossy techniques.
I am nearing the stage of completing the keyframe image compressor/decompressor. It is currently at a usable aplha stage. (0.7a) It is not optimized and as such does not compress/decompress as fast as it should.
As i have been designing this codec for CGI and similar images, The image compressor (Avid) performs poorly for real images (i.e. photos) or images with out large areas with gradients or similar colours.
The image compressor is available here: http://www.geocities.com/budgetanime/avid.html
If the primary site is swamped (geocities
) then try the mirror: http://members.lycos.co.uk/budgetanime/avidm.html
Currently the compressor only accepts 24 bit BMP images.
Also at the website there are some images which compress very well using my compressor.
If you have any comments or questions or found some bugs (no doubt that there aren't any ) please do contact me.
hmm, it looks like i still need a better exponent estimator for the Inner differences it gives consitantly a highe value than the optiumum value.... but it is still quite good.
to Tom: thanks for your help with the flood fill, after all that i scrapped that part from the compressor
but it was a good learning exersise anyway