[Openicc] [Printing-architecture] [Gimp-print-devel] Colour

[Chris Murphy]

On Nov 16, 2009, at 8:23 PM, Robert Krawitz wrote:

> That's why I think the intermediate color space should be high
> precision with extremely wide gamut.  I know memory and CPU isn't
> free, but I think arguing over 8 vs. 16 bits of precision is
> distinctly penny-wise and pound-foolish.

It's fairly well established that printers do not need more than 8bpc. The challenge is getting the correct 8bpc. And this is most easily achieved by an intermediate space with 16bpc or higher. (e.g. ACE is 32bpc float for some time now, previously it was 20bpc integer regardless of source bit depth). So yeah, any conversions shouldn't be considering 8bpc at all. 16bpc works in most instances, and 32bpc float is just going to totally clobber the issue altogether so I'd work toward a 32bpc float model. And just totally eliminate any question about whatever serious torquing of color spaces may be instituted (by the pipeline or the user or the driver or LUTs or profiles, or whatever).

So *if* there is a debate it's between 16bpc integer and 32bpc float, not 8bpc or 16bpc.


Chris Murphy