[Openicc] colord 0.1.0 released!
Leonard Rosenthol
leonardr at pdfsages.com
Wed Jan 19 17:35:02 PST 2011
On Wed, Jan 19, 2011 at 7:49 PM, Chris Murphy <lists at colorremedies.com>wrote:
> My understanding is that in PDF/X-3 (and 4 and 5), that /DeviceRGB =!
> ICCBased and is effectively "no color management" in the context of
> outputting to the OutputIntent device.
>
There is NO DeviceRGB allowed in a PDF/X-3/4/5 file - all RGB data MUST be
associated with an ICC profile. You can have DeviceCMYK and DeviceGray -
though they are actually only device values for the embedded OutputIntent.
The only "true" device-dependent colors in a PDF/X file are Spot/Separation
colors, since they are resolved independently on each device.
> If there is a mismatch, a redirection for example, of that PDF to some
> other output device: a proofing device, or a display, then the context is
> different and it means all /DeviceRGB objects have an implicit source
> profile = OutputIntent and thus are treated as ICCBased. So there is the
> possibility of "no color management" but it is conditional.
>
>
Retargeting a PDF/X file is "out of scope" of the standard itself. However,
the application notes clearly recommend that when the devices don't match
(for some definition of "don't match") that you need to actually do a
CMYK->CMYK transform using the original OutputIntent as source profile on
your DeviceCMYK data to the new OutputIntent profile.
Tagged/ICCBased colors (be they Lab, RGB, CMYK or Gray) will actually be
left alone in that case.
> Is there a need for an immutable "off" flag for ensuring objects aren't
> color managed no matter what?
>
It's been suggested at least twice to the ISO 32000 committee and both times
it's been rejected, because it's not clear what it means in all contexts.
For example, does it used when rendering to screen? What about if the
screen is simulating a printer (aka "Overprint Preview")?? etc. etc. etc.
Leonard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/openicc/attachments/20110119/25d74558/attachment.html>
More information about the openicc
mailing list