For those who are curious, ran some comparisons versus jpeg here (http://www.flickr.com/photos/corylinden/3330168391/).
Short version is that, as expected, jpeg block artifacts completely overwhelm image at very low quality levels. Minimal jpeg file size ends up ~3k.
What I want to know is how their algorithm looks without the space constraints.
For any compression scheme to take hold, it'll need to handle both large and small files well.
Before looking at your comparison, I thought their pic looked like crap. Now it looks absolutely stunning given the comparison shots. But how would it look at 200kb? Does it compare (dis)favorably against JPEG then?
