http://www.brucelindbloom.com/index.html?DanMargulis.html
The above linked page sort of refutes the validity of the "challenge", but essentially the assertion of the challenge is editing in 16 bits affords you no advantage.
- Must be actual photographs, not computer-generated art. May be scans or digital captures.
- Must be in color (this wasn't part of the original challenge, but I've seen enough B/W to understand what goes on there). May be in CMYK, LAB, or RGB.
- Must be free from excessive retouching.
- The corrections can be as extreme as you like but they must be real-world types of corrections. Therefore, nothing designed specifically to defeat the test, such as an extreme darken followed by an extreme lighten followed by an extreme darken etc.
- In doing the test, it is OK for me to convert immediately to 8-bit and then back to 16-bit. This is to ensure that only the *data* in the original 16-bit file is being tested and not merely the method of calculation.
- Must be willing to have me publish the result.
(emphaisis mine)
Essentially what this does is dump all of the extra data that 16 bits gives you to work with, now you're editing an 8 bit image in 16 bit mode. There's no advantage there.
There is no way to beat this challenge. It's rigged -- much like a carnival game, you cannot win.
