Re: Compressive images test

Hello, may I suggest to produce two sets of images:
- one set made of progressive images, JPEGrescan/ImageOptim does a pretty good job producing optimized versions of those.
- the other set made of the same images but recorded in sequential mode (aka "baseline", top to bottom), "jpegtran -optimize" can be used to turn progressive images into sequential ones losslessly.

The idea is to check if the higher memory consumption associated to progressive image decoding could lead to faster memory starvation on entry devices, simultaneous decoding of many progressive JPEGs will accentuate the problem.

Regards
-- 
Frédéric Kayser

Ben Callahan wrote:

> Andy, that would be great. I have a repo up here: https://github.com/bencallahan/compressive-test with some tests for us to play with. @pornel requested that we better compress the examples, so I will be refactoring the imagesthemselves at some point. However, you could use these tests to get your experiment setup.