I've been using Image Compressor over the last few months as a final step to improve colorization of images. Could you comment on some of the finer points of using the various parameters? I find that sometimes it creates unwanted color gradients in larger flat-colored areas, and occasionally it doesn't do at all what I expected. It would be nice to have a better feel for how to control the effect beyond the description in the user manual.
If you get the unwanted color gradients, it's because the image compressor is trying to pull detail out of the flat colored area. You can get rid of this by cutting back the Local Adaption control. It defaults to 8, but if your image has a lot of flat areas you might need to cut it back to 4 or less. On the other hand, if your image is really busy, you might be able to increase it even more to provide more punch to the effected image.
So the Local Adaption is the first thing to play with to figure out an optimal boost setting.
Gain and Bias add a non-linear tone curve to the processing. At 50 they are both non-active (50 = linear tone curve). So usually you start there, and then adjust them in increments if you want to play with the tone curve to get to an optimal setting. i usually just leave them set to 50. The more the settings vary from 50, the larger the non-linear tone curve adjust effect.
The Mode changes the overall kind of boost effect.