Re: CMYK spaces used for document creation
Re: CMYK spaces used for document creation
- Subject: Re: CMYK spaces used for document creation
- From: Chris Murphy <email@hidden>
- Date: Mon, 2 Nov 2009 16:02:59 -0500
On Nov 2, 2009, at 2:20 PM, Martin Orpen wrote:
On 2 Nov 2009, at 18:55, Chris Murphy wrote:
On Nov 2, 2009, at 10:04 AM, Martin Orpen wrote:
The PS conversion is bandy and ugly looking. The Argyll version is
extremely smooth and sexy looking.
Draw your own conclusions...
There isn't enough data to draw any conclusion. You've likely
triggered a bug, or have a problem in some procedure somewhere. The
differences in the various CMMs is visually typically minor. That
you're getting banding and ugliness with a product that has
converted billions of images over the better part of a decade
implies something unusual is occurring. To figure it out requires a
very clear step by step be published so others can try to reproduce
the the issue.
AFAIK there is no problem with ACE conversions, which have been
32bpc float for a few versions now. So there is no lacking in
precision for the conversions. Banding implies massive detail loss,
so again it's a bug or a procedural flaw.
The major difference between the two procedures is that the superior
proof was produced using a device link profile as well as a
different CMM.
The test page would appear to highlight the problems in the
conversion process NOT problems with the profile and or proofer.
There is still no substantive data here or a procedure for anyone to
follow to try and reproduce your results. You are comparing different
profiles and a different CMM, and it's not clear which is the primary
factor in the adverse results you're seeing with ACE.
You could test the devicelink profile using ACE to do the conversion
and see if you still get adverse results or not to narrow down the
problem. Have you done this?
When building the devicelink, are you using Argyll? And if so are you
using the -G flag? What other flags are you using? What is the
provenance of all profiles being used in both sequences?
At this point the assertion that ACE conversions are so vastly
inferior, visually so, to the effect they're noticeably and obviously
banding and ugly, is total conjecture without a presentation of data
and a sequence for reproducing the behavior. Merely saying things does
not make them true.
ACE for some time has been 32bit float precision and there is only one
other CMM I'm aware of that has that level of precision. So again,
you're either triggering a bug, or possibly you have a very, very
unruly device that simply does better with inverted A2B devicelinks.
Chris Murphy
_______________________________________________
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