I've been wondering if it is possible for a large aperture lens, or mirror, to receive parallel rays of light, converge them, then diverge them into a smaller beam of parallel rays as they exited the optical device? If so, directing this greater flux density into a telescope matching the beam's diameter should increase the apparent surface brightness of an extended celestial object, a non-point source (e.g. star). This would improve imaging exposure times and allow for the design of a telescope which one could see celestial objects in color - a colorscope. Improving surface brightness of extended objects is supposedly a violation of the 2nd law, but I don't see it. Any ideas?