Which of the many algorithms for converting color input into black and white are you after? Talking about that, could we have yet another super resolution mode, by separating the color components into grayscale pixels? (knowing the geometry of the pixels in the sensor that shouldn't be a big challenge (unless the arrangement is not in a rectangular grid; though if it is, the result might not have correct aspect ratio, but for some purposes that would be acceptable with the additional detail achieved)