Originally Posted by
Marko
A 16-bit image provides 65536 discrete levels of information that can be altered by processing. That's what you would be able to use if you scan as grayscale. If you scan as RGB, you are getting 16 bits per channel, or a 48-bit image with billions of discrete levels of information. And in order to handle all this information, you need to assign it a colorspace so it can be managed properly.
Even though you will use only a fraction of all that information for printing, you lose some of it with every operation you apply, and the more information you start with, the more you will be left with after all is done. So it makes sense to work in the widest available colorspace and let the color management system handle the translation. That's the purpose of color management.
Bookmarks