[Openicc] [ANNOUNCE] Compicc 0.8.3

[Kai-Uwe Behrmann]

Am 30.11.10, 10:19 +1100 schrieb Graeme Gill:
> Richard Hughes wrote:
>> One thing that confuses me is how you should be dealing with VCGT
>> data. For profiles that have VCGT data, do you create a tonecurve for
>> each channel to remove the VCGT effect before you apply the GPU
>> transform or just push the VCGT data to the video card like normal?
>
> The VCGT should be sent the the hardware - that's the idea of it.
> Some hardware has higher resolution lookup table output than
> the frame buffer, and therefore it is an advantage to linearise
> and make neutral the hardware response with the extra resolution.
> (My current graphics card has 8 bit/component frame buffer, and
> 10 bit lookup table output to the D/A).

Exactly. But is not this kind of hardware going nearly completely away? I 
mean most computer to monitor paths are D/D.

>> I must admit, I'm not completely sure of the role of VCGT in a
>> late-bound color workflow.
>
> Same as always. It's hardware calibration state being carried along
> by the profile. The profile assumes the hardware is in the state
> described by the VCGT, so it needs to be put in that state to be valid..

kind regards
Kai-Uwe Behrmann
-- 
developing for colour management 
www.behrmann.name + www.oyranos.org