[argyllcms] Re: AW: Re: Question about viewgam
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 25 Feb 2021 21:19:02 +1100
Hermann-Josef Röser wrote:
Hi,
An input device like a scanner or digital camera (for a fixed exposure time)
delivers R,G,B triples and the range for each is (if normalized) restricted
to the range 0 to 1. So there are fixed limits, or do I misunderstand your
statement?
That's the output of the sensor, not its input. It's input is
the light hitting it.
All possible R,G,B-values completely fill a cube with side
lengths 1,1,1. If I use the ICC profile to transform the RGB-triples to the
Lab-space, the surface of this cube projects to a well defined volume in
Lab. ??
Yes, but an input profile maps from the color of the light hitting
the sensor to its RGB values.
[Of course typically in using it, the inverse of the profile is used to
convert the RGB values back into the colors.]
So you could in principle throw every possible spectrum at
the input device, and (if it properly adjusted) you should
never hit an RGB limit.
So what does the volume in Lab-space, I see in the output of viewgam,
represent?
Nothing very useful. Putting device limits into the inverse of the
input profile gives you some sort of (probably inaccurate) extrapolation
of colors outside (i.e. possibly imaginary, possibly folded back) the possible
input to the sensor.
Cheers,
Graeme Gill.
Other related posts:
