> For what you're suggesting, you'd have to view your images in both > modes, then go back and tweak the high-color images, and repeat this > process until _both_ sets looked good. The graphics people usually > found it was a lot less work to create their images in 256-color or > 3:3:2 color mode to start with. My usual technique is to draw an icon image in 4-bit RGBI or 3:3:2 true-color to begin with, using dithering and pointillism to generate the subtle shading in that mode, then copy it up to a higher mode and replace the dotty stuff with true shading. That's still sort of double work though.