Re: Null Profile?
Re: Null Profile?
- Subject: Re: Null Profile?
- From: Steve Upton <email@hidden>
- Date: Tue, 13 Nov 2001 10:24:53 -0800
At 2:48 PM -0800 11/12/01, JWL wrote:
Steve Upton <email@hidden> wrote:
>
The target capture values are in RGB and the data file values are in
Lab. There would be no way to edit the target capture values to be
"the same".
-- I would convert the target capture to Lab, then edit, then back to RGB to
feed to the profiler.
How you converted the file back to RGB would define the colorspace of
the RGB numbers. If you had Adobe RGB as your working space then the
RGB values would be Adobe RGB, not the mystical "null"
>
This also brings me to a point I've been meaning to post.
>
There is, by the definition of how this stuff works, no such thing as
a "null" profile when dealing with device profiles. Device profiles
convert from Lab (or XYZ) to device space or device space to Lab.
There is no way to create a profile that will have no effect (null)
as it will have to convert from a colorimetric value to a device
value or vice versa.
-- ahh! Thank you, Steve. I was hoping someone would respond directly to
the logic of this. I have been struggling to wrap my poor little head around
it. Really what I imagined was a way to achieve a "null" transform - so the
conversion from source input profile to destination working space gives the
same result as just grabbing raw data from the source.
Yes! A "null transform" is often obtainable in just the way you
described. Again, you may be the victim of rounding errors depending
on the CMM doing the job.
Sounds like there is no way to build an input profile to have this effect.
However... any thoughts about that Flextight Input profile from Imacon, if
it was supposed to function that way (for scanning negs)?
Andrew sent that profile to me and I took a look. It seems to be a
slight variant of the Apple RGB space with a slightly different blue.
I suspect color coordinates in the profile match the CCDs in the
Imacon scanners. by selecting this profile you are effectively
importing the raw RGB.
The only profiles that you could create as "null" are abstract
profiles - like Lab->Lab or certain flavors of device link profiles -
RGB->RGB or CMYK->CMYK
-- does this mean it could be possible to create a "null" RGB working space
profile? If so, how would one go about constructing it?
No, it is not possible to construct any profile that converts from
device to lab. The profiles I am referring to would convert from one
space to the same space and perform the changes internally. As you
mentioned above this would be a " null transform".
I guess, in light of the Imacon profile example, it is possible to
have a profile that is specifically matched to the raw output of the
device. In this specific use it could be seen as a null profile but
would still need to be applied a second time in order to restore
something close to the original RGB numbers.
Regards,
Steve Upton
________________________________________________________________________
o Steve Upton CHROMiX www.chromix.com
o (hueman) 866.CHROMiX
o email@hidden 206.985.6837
________________________________________________________________________
--