Actually, if I remember now, oblique lighting shows much less density change
than direct lighting does because it's reflecting at an angle. Direct
lighting if I am not mistaken, has a component that shines through the
paper, hits a metallic trackerbar or whatever, and glows directly back into
the sensor with both reflected and direct light. But you may be using the
CCD light, in which case you couldn't change that.

If you were to use direct light through the paper, then I think the best
color to start with might be green or blue, since that would be the color I
suspect that would be absorbed best by yellowish or orangish paper. (But you
never know for sure until you try it).

Craig