Re: On smart CMMs and dumb profiles
Re: On smart CMMs and dumb profiles
- Subject: Re: On smart CMMs and dumb profiles
- From: Chris Murphy <email@hidden>
- Date: Mon, 8 Jan 2001 00:34:12 -0700
>
Chris Murphy and some of the other US folks have written in favour of
>
the smart CMM and dumb profile workflow model. I've written in favour
>
of the smart profile and dumb CMM workflow model.
That's from a perspective of whether you want this stuff to have better
performance than it does now. It's not going to get much better than it
is now unless we have a way to do dynamic gamut mapping and this is
impossible to do with a bland CMM system. Profiles can only be made so
smart. They are just text files. They contain no code and no algorithms
so the "smart" profile model really doesn't exist either. What we have
now are mostly stupid CMMs (with one exception) and mostly stupid ICC
profiles. The smarts are in the profiling packages.
>
>
The spec will be updated sometime in November, I think.
I thought it was supposed to be updated last November. I was under the
impression that some of the black point issues were going to be addressed
at that time (not totally resolved of course, but at least addressed to
avoid the black hole black point problem with a lot of profiles that are
floating out there.) When was the last update?
>
Now if the smart CMM approach wins out, then users will vote with
>
their feet, shifting the definition of device independence from CIE
>
color models to a workflow definition that says you have device
>
independence when you're using the same CMM.
Not true. A smart CMM approach says that you have device independence
with profiles, but you only get the same RESULTS with the same CMM.
Right now you don't ever get device independence because every profiling
package produces completely different results for the same device; and at
the same time uses the same gamut compression regardless of source (with
the exception of one profiling package which gets around this at the
profile building stage but not the profile usage stage).
Right now we do NOT have outstanding inter profile operability - i.e.
using profiles of different vendors ensures you don't get ideal results.
>
This again will mean that as distributed color production grows, the
>
only sane strategy left to users is to base workflows on the OS-level
>
default CMM.
That's simplistic. It depends on how much smart CMM's are going to cost.
>
First, manufacturers who hope to make money by selling proprietary
>
CMMs will pull the plug on their own markets (as some already have by
>
relying on proprietary tags).
I'm not sure what this means.
>
Second, manufacturers will have no knowledge of each other's
>
workflows, turning back the clock to pre-ICC days.
This implies they know something about each other's workflows now.
Illustrate how going to a smart CMM model changes things.
>
Third, users will have a hard time negotiating workflows.
They already have a hard time negotiating workflows.
>
This way proprietary technologies spell survival over there and
>
extinction over here.
The current model *is* proprietary right now. Each profiling application
is proprietary. Each gamut mapping method is proprietary as well as
static. You get different results with every CMM right now, and you get
even MORE different results when you mix profiles from different vendors.
I think the idea of going to a smart CMM is more about locking down the
variability on at least one aspect of ICC based color management than
anything else. Right now profiles and CMMs are a moving target.
Chris Murphy