On Thu, Jul 17, 2014 at 1:09 PM, Weldon Goree <wel...@langurwallah.org> wrote: > Now that I say that, I wonder if serial storage of the color channels > would compress better, since you'll generally have low-frequency content > for most of the channels most of the time.
Interesting idea. I'd be interested in knowing wether it actually compresses better. Also interesting is that, in order to have file access patterns that don't confuse the OS too much, you'd need to change some of the algorithms. Even simple things like grey-scaling you would need to do colour-wise pixel-wise (i.e., inverting the nesting of the for-loops in the code). One of the disadvantages of the approach would be that you'd need to wait until you've received all of the data before you can actually put something significant on the screen (which is the complete opposite of what jpeg gets you). Cheers, -- ______________ Raphaël Proust
