[argyllcms] Re: Perceptual intent

  • From: <robert@xxxxxxxxxxxxxxxxxx>
  • To: <argyllcms@xxxxxxxxxxxxx>
  • Date: Sat, 20 Dec 2014 09:23:21 -0000

Graeme Gill said:

>Right, it helps to have the details. You are applying a viewing environment
>transform. This will affect the luminance scale reproduction. You are
>selecting a destination viewing environment of Print evaluation
>environment, which is quite bright. The source viewing environment you have
>selected is for a dark environment, so the assumption is that the image
>looks right on the display with a high gamma, or a darkened appearance. So
>the overall adaptation from vey dark to a quite bright environment will
>involve significantly reducing the gamma, or brightening of the image,
>raising the shadows.

Well, that explains it, at least as far as the perceptual mapping.  Why does
it not do the same thing for the relative mapping?

OK, OK, I'm beginning to get the idea.  Perceptual changes the image to make
it, as Florian said, well, perceptually pleasing, whereas Relative maps to
the nearest point on the destination gamut, more or less, depending on who
wrote the code.

>I would draw a different lesson - one you will find that I tend to preach a
>lot - if you're not sure why or what the parameter is for, leave it out.
>The tute shows -cmt and -dpp, which should have a much less pronounced
>effect than the ones you have chosen.

Well, the lesson that needs to be learnt is that in general
non-color-management-expert people like me don't know what they are doing
when using a program suite like Argyll; but they eventually learn
(hopefully) by bitter trial and error until they either get it right ... or
give up and use a GUI-based system like i1Profiler (which has some
limitations, but is much less likely to result in booboos like this one,
requiring assistance from the development team).

It would be useful if you stated in your documentation - in big red letters
on every page - IF YOU DON'T KNOW WHAT A PARAMETER DOES, LEAVE IT OUT!

 >>Any chance of an ArgyllWiki for people like me??

>It's not clear to me how it would add value, above the existing
>documentation.

>If the existing documentation has mistakes or is lacking, I think it's
>going to be far better to fix it or improve it, rather than wonder if
>something, somewhere in a Wiki that's not kept up to date with software
>releases, could help.

A Wiki should never be a substitute for the documentation, which has to be
the authority; and of course fixing and improving the documentation and
ensuring that it is up-to-date and covers all previous releases is
essential.

What a wiki can do is to add topics that explain color management concepts
and how Argyll implements these, use-cases demonstrating things like the -d
-c parameters in colprof and why they affect perceptual mappings and not the
other mappings etc.  The potential power (if your users buy in to it and
contribute) is that it can bring the knowledge of many people together in a
way that can benefit all of us ... and make ArgyllCMS more understandable
and approachable to new users.  It would mean that your input would be
supplemented by the input of others, not replaced or reduced.

If we, as users of Argyll, take this seriously, then we will keep the wiki
up to date; of course there would be nothing to stop you from adding to it
(just like anyone else), and correcting errors.

Of course, if you don't try it you'll never know if it's a good idea or a
bad one.  But what you could do is to ask your users if they would be in
favour and if they would be prepared to contribute.  If the response is
positive then you could set it up in a restricted way, not open to everyone
until you saw how it worked out.  And then if it proved to be good, you
could open it out ... or close it down if it turned out to be bad.

I think it would be an addition to Argyll that would help to popularise it
and make it more accessible to people like me.  As an example, I came to
Argyll through Florian's DispCalGui, but even there I had no idea if I
should be using a matrix profile or a table profile, what the (dis)advantage
of using a huge target is compared to a small one etc.  I didn't know what
hardware calibration of a display was or what it meant; how the video card
is used; what is loaded into the video card and what is not and how to avoid
the wrong calibration from being loaded .... etc., etc., etc.  But, had it
not been for DispCalGui and for some posts I found explaining how the user
had managed to produce a print target and a profile from it using Argyll, I
would not have gone near Argyll.

AND ... I happen to be an engineer with a post-graduate in computer science,
AND I like to find out how things work, AND I find the whole vision/color
management thing very interesting (but confusing), AND I do my own printing,
AND so on.  What about the guy who's been a photographer all his life and
has zero technical background?  Don't you think he/she could do with a bit
more help than your (excellent) documentation?  As I'm sure you know, trying
to get the information from the color management books out there and from
the ICC is a painful experience!

Anyway, look, I've made my case ... but it really doesn't matter to me.  At
this stage I can mostly find my way around (although I'm still waiting for
answers I posted months ago on freelists) ... and as I've invested in
i1Profiler, GamutVision etc., I can go my own way.  I just think that
there's some very good stuff in Argyll and I really do think that a wiki
would be of benefit.

Robert



Other related posts: