[argyllcms] Re: "Washed" / low contrast colors on calibrated display
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Mon, 07 Jan 2008 22:46:27 +1100
Frédéric Crozat wrote:
All images I'm testing right now are designed to be display in sRGB
colorspace. Which is why
I'm currently focusing on calibration result and not on profiling result (yet).
Right, but what sRGB calibrated device are you using as a reference ?
I understand that but she has been designing backgrounds for more than
8 years and we know
we can't get exactly the same result on most monitors (since 99%
aren't calibrated and moreover we don't use the same OS for design and
display) but we are kind of used to this and for my test, the result
is really different.
There's no substitute for a known color reference device.
Looking at some of the images on my display (which is calibrated and is
set to interpret the RGB as sRGB using Photoshop), I'd have to say
that the images (subjectively) look a little dark. Now I haven't double
checked my display response, so this might just be my setup.
They certainly seem contrasty and saturated as you'd expect for
"demo" type images.
So, my guess would be to use the ambiam light detector to check what
is the current value of ambiant light and
tune brightness accordingly.
If you want to be scientific about it, yes. It's hard to locate
a well calibrated ambient light instrument though. The absolute
calibration of the devices supported by the Argyll drivers is
not something I'd have a whole lot of confidence in (they
might be accurate within +/- 50%).
Simply adjusting the brightness manually until your happy
with it, isn't such a bad way of going about it if you know
what to look for.
BTW, I'm wondering if it wouldn't be better to tune backlight and not
brightness in that case ?
For an LCD, "brightness" should be controlling the backlight intensity.
As I wrote above, I'm mostly interested in getting correct sRGB output
(but I know I'll be much more careful when choosing my next LCD
monitor ;)
I'd imagine most manufacturers will be targeting sRGB as well,
so something close should be achievable.
Moreover, my monitor is kind of old (2002), so at that time, LCD for
"average" consumer weren't
that precise, I guess (and most consumers were not searching that information).
It's not that noticeable, except when compared side by side, since
your eyes adapt to any given white point.
Ok, I guess I'll stop playing with trying to change whitepoint and use
one of those "predefined", even if their
value look crazy.
Pick the one closest to D65 if you are targeting sRGB.
In that case, you should drop it from dispcal manual too ;)
I can't see it in the current documentation:
<http://www.argyllcms.com/doc7/dispcal.html>
I also tried adding "-v" and I must confess I found the output
extremely interesting. I'm wondering if it wouldn't be better to push
the "iteration x on y in progress" and "patch x on y" in the default
message (when -v is not set), it would give some progress output to
user.
It's hard to judge, but I thought the default behaviour is "busy" enough
as it is, and if you want more it's simple to use "-v" or "-v2".
cheers,
Graeme Gill.
Other related posts:
