What is Monitor Gain and How Does it Affect Color Gamut?

As I understand it, when I adjust the individual RGB controls of my monitor during calibration, I'm adjusting the gain of the monitor. Maybe I'm misunderstanding how gain works, but if a gain of 100 on all three channels gives me the brightest, most saturated colors and a gain of 0 gives me muted desaturated colors, aren't I basically reducing the number of colors my monitor can display by calibrating it? This would seem counter-productive to me.

The reason I ask is because I have a computer monitor whose default settings are extremely bright and I can't adjust my screen's brightness to the standard of 120 cd/m2 without first reducing each of the RGB controls to 70, then fine tuning the RGB output during calibration using my Gretag Macbeth i1 Display 2 device, then reducing my monitor's brightness to achieve the appropriate 120 cd/m2.

I've tried to search the 'net for an explanation of just what gain is but haven't been successful. Any insight is greatly appreciated.

Think of "gain" on your monitor as being analogous to "volume" on a sound system.

On a sound system you can adjust your volume from 0 to 10. But that doesn't mean that the optimal way to listen to music is with the volume at 10. In fact, the music will probably sound terrible and distorted with the volume all the way up.

Same with a monitor.

Hi Rick,

Thanks for the quick response. So it sounds like decreasing my RGB values to 70 before calibration didn't really 'hurt' anything as far as the color gamut of my monitor is concerned? I found this was the only way for me to get the brightness down to a reasonable level. With RGB at 100 my screen was at about 240 cd/m2 even with brightness set to 0 in the monitor's OSD.

I should also add my system specs: Currently I'm using a homebuilt Windows system running Windows Vista x64 on an Intel Quad-Core 2.88 GHz processor with 4 GB of RAM. My monitor is a 24" Dell 2408WFP LCD display. I have an nVidia 8600GT 512MB Video Card as well.

Brad

But, yes, ANY form of calibration or colour management reduces the
total number of distinct colours that can be displayed, at least if
you are starting with an 8-bit/channel original. That's the price you
pay.

To see the effect, it's easiest to think of a greyscale image.

Consider: if you have an 8-bit greyscale image on an 8-bit/channel
monitor, and you use the image to control the colours directly then
you get (in theory, and barring other effects) the maximum number of
possible colours: 256. If you have any kind of adjustment to the
colours then either
(a) some two colours (greys) in the original map to the same colour on
the monitor, meaning that there will be some unused shades of grey; or
(b) the adjustment is so minor it has no effect at all.

But, the shades of grey you see would then be a property of your
monitor, and nothing else. That may be fine, but if you want to see
the same as other people on other monitors, you need to do some
adjustment. And this means a loss of colours for, perhaps, all but one
monitor and maybe not even that.

The same applies to all kinds of editing of images, such as
brightening or darkening an image. It reduces the number of colours.
To avoid this effect _while editing_ it's useful to work with a
16-bit/channel image.

Aandi Inston

Aandi Inston wrote:

But, yes, ANY form of calibration or colour management reduces the total
number of distinct colours that can be displayed, at least if
you are starting with an 8-bit/channel original. That's the price you
pay.





Right, but only if the calibration is applied to an 8-bit graphic card internal to the CPU. If, instead, the monitor display has its own 10-bit LUT (or higher, like several units have these days), then the CPU's graphic card is left linear, and no decrease occurs in the maximum possible number of unique colors that can be eventually displayed on the monitor.

Good point. My discussion only covers 8-bit/channel hardware.

Aandi Inston

A reduction of the luminance of a TFT monitor from
e.g. 200 cd/m2 to 100 cd/m2 is hopefully never done
by the graphics card LUTs.
It's IMO done by monitor electronics, using monitor
control menues.
Nobody knows what's happening internally.
In my experience with EIZO CG19 this doesn't lead
to posterization because of reduced max. luminance.

The graphics card LUT or the higher resolution LUT in
the monitor (EIZO) has to apply only small corrections,
if the monitor is manually pre-adjusted near to the
target values.
This doesn't reduce the number of available levels
considerably.

Reducing the luminance (in limits) doesn't change the
chromaticity coordinates of the primaries very much,
if the monitor is technically OK.
The gamut is not reduced. It is assumed that the
observer is adapted to monitor white.
A secondary effect is named after a scientist, whose
name I forgot: darker samples APPEAR less vivid,
though the chromaticity coordinates are the same as
for brighter samples.

Finally I would like to prove some of my statements
by test results for Eizo CG19, p.36 here:
<http://www.fho-emden.de/~hoffmann/colcalc03022006.pdf>

The luminance is only 80 cd/m2, but the chromaticity
coordinates are near to those of sRGB (both sets are
different, but the actual set is not contracted towards
white).

Best regards --Gernot Hoffmann

Gernot_Hoffmann@adobeforums.com wrote:

> A reduction of the luminance of a TFT monitor from e.g. 200
> cd/m2 to 100 cd/m2 is hopefully never done by the graphics card
> LUTs. It's IMO done by monitor electronics, using monitor
> control menues.

Someone once told in this forum (Andrew Rodney as far as I recall)
that if one changes the "gain" of an ordinary TFT monitor (not the
LED ones though) by adjusting the individual RGB controls using
monitor control menues, the adjustment really happens in the graphic
card LUT. Isn't that true?

--
Regards
Madsen

Thomas,

the free GMB program Calibration Tester shows the
contents of the graphics card LUTs.
Altogether three programs, mentioned here on p.20:
<http://www.fho-emden.de/~hoffmann/measgamma10022004.pdf>

If the luminance settings of the monitor are above
the target value e.g.100 cd/m2 for a calibration
by instruments, then the graphics card LUTs are
trying to reduce the luminance (old example in the
doc, also p.20), for instance from 110 to 100 cd/m2.
Such a case would reduce the resolution a little.

As tested by Eizo CG19:
The adjustment of the monitor by monitor menues
doesn't affect the graphics card LUTs. In fact,
the calibration by GMB ProfileMaker5 guides the
user to adjust the monitor dynamically/interactively
so, that the LUTs contain finally almost no correction.
In this sense the old example above wasn't perfect.

This is the classical method, which I'm applying as
well to Eizo CG19, though this monitor offers a
calibration by ColorNavigator, based on monitor-
internal LUTs with 10 or 12 bits per channel
(leaving the graphics card LUTs at identity y=x).
For certain reasons I'm not using this feature.
One reason: then it would be impossible to have a
look at the contents of the LUTs.

Theoretically the result should be better for high
resolution LUTs, but I wasn't able to detect banding
or posterization after applying the classical method.

Best regards --Gernot Hoffmann

Thomas G. Madsen wrote:

Someone once told in this forum (Andrew Rodney as far as I recall)that
if one changes the "gain" of an ordinary TFT monitor (not the LED ones
though) by adjusting the individual RGB controls using monitor control
menues, the adjustment really happens in the graphic card LUT. Isn't that
true?




No. It is *possible* to control the monitor's settings, but only if it is connected to the CPU by a communication cable (usually DDC/CI-VESA compliant). I've certainly never heard that it's possible to affect the CPU's display card from the display (and I cannot think of a reason why that would even be desirable).

Apple's monitors allow for control of brightness from the CPU (the intensity of the backlight), but nothing else (not contrast, not the RGB signals).