On 6-Aug-09, at 12:04 PM, Joel Webber wrote: > On Thu, Aug 6, 2009 at 1:47 PM, Joel Webber <[email protected]> wrote: > Side note: God, I thought I would never have to think about this > problem again after everybody finally dropped their old VGA cards, > and we could at least just deal with 5-5-5 vs. 5-6-5 16-bit color > modes in all our assembly code. Wait, did I just date myself pretty > badly? > > Side side note: Holy crap, I just realized that NeuQuant uses a > Kohonen network to iteratively converge on a palette. No wonder it > works so much better. But does it end up taking a really long time > to converge? I'd hate to make the build take forever because it's > waiting on a neural network to converge :)
NeuQuant was pretty fast in all of my experiements. For our build pipeline I actually ended up using pngquant instead of NeuQuant because of the difficulty getting NeuQuant to process transparency. I haven't tested the JAI version of it, but the NeuQuant version linked on that website completely ignores transparency. On top of that, I was having trouble getting Java to output IE-transparency-compatible PNG files. It also seemed like I had to construct PNG files by hand with a very specific set of chunks to get IE to recognize them, meaning I had to skip the convenience of ImageIO PNG output . The output quality is great, but the hoops I had to jump through to hook it into the pipeline forced me to drop it. Dither like it's 1999! --~--~---------~--~----~------------~-------~--~----~ http://groups.google.com/group/Google-Web-Toolkit-Contributors -~----------~----~----~----~------~----~------~--~---
