Re: Can DeviceLink conversions be better?
Re: Can DeviceLink conversions be better?
- Subject: Re: Can DeviceLink conversions be better?
- From: Rolf Gierling <email@hidden>
- Date: Fri, 29 Aug 2008 21:26:37 +0200
Hello Todd,
thanks for your feedback.
Hi Rolf
Thanks for setting up the materials for this test! Based on what I
am seeing on screen, the DeviceLink does indeed produce a more
faithful conversion. I don't know if I'd say "much superior", but
there is less color shift and the tonality is smoother using the
DeviceLink, especially for the out of gamut colors. I don't know if
I would see as much a difference with a hardcopy test, but I assume
there would be some.
I would like to add one of my observations that is not obvious from
this test:
The differences between "normal" profile conversions and DeviceLink
conversions seem to rise
if the lightness range of the destination decreases, i.e:
I can hardly observe a difference if both profiles are matrix
profiles, for example Adobe RGB -> sRGB.
It becomes obvious if the destination is an (RGB) inkjet (bright
papers and a dark black point), as in this test.
More obvious on press (CMYK) destinations (greyish papers, lighter
black point).
Of course the real question now is how did you create this
devicelink profile? As Terry mentioned earlier, "...chances are
it's doing the gamut compression the same way had you done just a
simple ICC profile conversion. There's considerable "secret sauce"
in the various device link applications..." I have a couple
applications which can create device links and they do not do as
good a job as whatever you are using. While I am inclined to
believe that devicelinks are CAPABLE of doing better conversions, I
just haven't been able to make one myself! What software did you use?
If you build a DeviceLink using the relative colorimetric intent,
there is no room for doing a new gamut mapping.
If you choose a perceptual or saturation intent, the profiler can use
"secret souce", for example doing a smooth
transition between compression of the saturated colors to
colorimetric of the unsaturated colors.
I used basICColor DeviL to create the DeviceLink, which has four
additional custom "secret souce" presets
that completely recalculate the gamut mapping and the lookup table.
Note that the demo version does not create
DeviceLinks usable in other applications or RIPs, but you can use
them with the built-in image converter.
A secondary question related to your procedure: What do you mean
"the printer can match the fogra mediawedge with tight tolerances"?
Tight tolerances compared to what, FOGRA 39L? If you are using an
RGB profile and RGB images, how can you use a CMYK control strip
(the fogra mediawedge)? Forgive my ignorance if this is obvious!
I admit, this would have needed an explanation. This has nothing to
do with the DeviceLink
conversion itself, but was just a statement that the printer profile
is of good quality.
It's easy to make a competition between a worse printer profile and a
good DeviceLink
and proove that the DeviceLink is better. Contrary, if you compare a
good printer profile
with a worse DeviceLink, the results could run into the opposite
direction.
Both would not be objective, I think.
If I assign ISOcoated v2 (which represents fogra 39L) to the fogra
mediawedge image,
then go to the print dialog and let Photoshop convert the image at
print time to the
Epson RGB profile using the absolute colorimetric intent, I get a
poor mans proof.
Then I can compare the mediawedge patches against it's reference
values and evaluate the tolerances.
If I am within the tolerances (and with the supplied profile, I am),
I have at least one criteria for the
smaller ISOcoated v2 color gamut, that the profile has a good quality.
Regards
Rolf Gierling
-----------------------------------------------------
Multitools
Dipl.-Ing. Rolf Gierling
Colormanagement
Development - Consulting - Distribution
Mariabrunnstr. 16
52064 Aachen
Germany
Tel. +49 (0) 2 41 / 4 01 25 79
Fax. +49 (0) 2 41 / 4 01 25 83
email@hidden
www.multitools-online.de
Author of "Farbmanagement"
3rd Edition July 2006
mitp Verlag, Bonn
ISBN 3-8266-1626-X
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden