OK - as promised, here are the results (not as good as I had hoped). The image on the left, is using no dither, 1000 solutions, 45 million iterations and an image mask of my entire car. This made the red car above it have too little detail. The image on the right is also no dither, 1000 solutions, 63 million iterations, and no image mask. Before running Rasta Convertor, I trimmed the image in Photoshop, converted it to a 256-colour image using the Altirra.act palette, and applied a 75% diffusion dither to the image. [attachment=291383:45 million with mask.png][attachment=291384:63 million no mask.png]
Here are the source images [attachment=291385:87Z.png][attachment=291386:87Z_mask.png]
You need to let Dither run in the rasta proram, it makes a huge difference I have found. The Knoll method is slower but gives much better results. The car photo you converted is nice though.
Lots of great images here but some are horribly stretched. Like the Tolkien picture. It would be cool if he were not squashed. Philsan's images look good. Maybe he could post a tip for the others on correcting aspect ratio.
That was my fault. I did a resize of the pictur before processing and did not keep the correct aspect ratio.
The wolf is amazing! How long are these images taking to convert?
Using Knoll dithering, 10000 or more solutions, I try different pallets until I get one that looks good on destination screen I let them run for 100 million to 200 million cycles which can take 8 to 12 or more hours depending on the picture. I let it run, minimize it and kind of forget it's there for hours at a time.