01:32 am - ARGH! I don't give a damn about the pretty-print equation!
Just tell me what the actual formula is for the case in point, don't give me the 'universal' equation and tell me to plug in the variables for a software algorithm!
*sighs* Sorry... just pissy as all hell. There are NO straight-forward documents anywhere on the entire internet that explain JPEG compression without showing silly calculus equations all over the place.
And the only 'source code' I've found is even harder to understand, because it's NOT a straight-forward implementation of the algorithm. It's a hugely convoluted library (and all that that implies and entails) that supports every possible option in the file format.
Hasn't anyone written a good reference on how progressive JPEG encoding and decoding works? *sighs* Some things are straight-forward, like colour conversion, and the blocking, but some are really akward to try to understand from the documentation. Like the actual math behind the transformation each macroblock goes through. And the actual output file format (JFIF) isn't very nice to try to understand either.
I'm beginning to see why most JPEG software still costs around three times what equivilant GIF or PNG optimization software costs. GIF and PNG are trivial to work with and manipulate because there are straight-forward, simple documents that show how each step works in sequence, not worrying about the theory behind every design decision but simply stating what gets done when, and how you shuffle which bits around.
And yes, this is all because I'm trying to write a tool I plan to release for free that will allow importing of PNG or TIFF files, and 'optimized' adjustment of JPEG compression on a macroblock-by-macroblock basis. So yes, you'll be able to throw away 75% more bits on some sub-block of the image, including doing so seperately to the colour blocks from the grayscale blocks, if I can work up a feasable UI for that adjustment and implement it.2 comments