In a 16 bit file, there is no banding in the data. Any banding you see there is in the display system.
With a small qualifier. Selections are always 8 bit, so if you have a mask based on a selection - even if the mask itself is 16 bit - and you then use that mask to make an extreme adjustment, that might result in banding in some cases. But it will never be as pronounced as in a natively 8-bit file.
Got it. So, essentially, if working with 16 bit files, the only benefit of having a full 10 bit per channel workflow is to see the data that's actually there more accurately. Having 8 bits per channel(say, due to graphics card output limitation) only impacts what you see on the display - not what is actually being drawn. Did I understand correctly?
Yes - that is correct
Got it -- thanks Dave!