[argyllcms] Re: Fakeread with cmyk profile...
- From: "Auke Nauta" <auke@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 19 Feb 2009 14:35:57 +0100 (CET)
Hi Martin,
I am building/designing 3D-LUT's (and/or device link ICC profiles) for
photographic printers (mostly minilabs).
IMO, superior results can be obtained by assigning sRGBv4 to the (usually
sRGB) source images and transforming, perceptually, to the device's output
colour space. Using 3D-LUTs or device link profiles, there are lots of
opportunities to enhance the transform. (I mostly use abstract profiles to
adjust the result towards taste).
Images generally look more pleasing this way, although technically less
correct. (important memory colours are OK though). A big advantage is that
nearly the whole gamut of the output device is used. This would not happen
when assigning standard sRGB v2.1.
The reason that I 'need' a blackpoint of 0 is that the transform from RGB
0,0,0 will yield approximately 0,0,0 after the transform.
When the device blackpoint is too high, the result is way off, forcing me
to fix things later on.
I have tried applying bpc, but this does not necessarily results in good
output...
So, now you know!
Best regards,
Auke
> Hi Auke,
>
> I have no answer for you, sorry. However, I'm curious on how you plan
> to use this black point mapped profile.
>
>> The actual blackpoint of the device is approximately 10.5,1.5,1.0 (Lab).
>> * fakeread -l -b0,0,0 My_CMYK_Profile.icc target (.ti3).
>
> Martin Weberg
>
--
In a world without walls and fences, who needs Windows and Gates?
Other related posts:
