Re: basiccolor INPUT
Re: basiccolor INPUT
- Subject: Re: basiccolor INPUT
- From: Graeme Gill <email@hidden>
- Date: Sun, 06 Nov 2011 13:35:25 -0800
email@hidden wrote:
well I'm displaying a misconception then on my part. I hate to stray more
from the topic... but I assume then that input profiling has some method
of determining device performance, and therefore space size and
limitations, beyond the sample data sent and returned from the device?
Hi,
I'm not sure what you mean by "device performance", but profiling isn't magic.
A profiler simply takes the correspondence implied for each sample point
between device values (ie. RGB) and PCS values (ie. CIE XYZ) and casts them
into a standard format (ie. ICC). This implies casting it into a continuous
rather than discrete (sample point) form. Naturally the continuous mapping
will be most accurate when it is close to, or between sample points (interpolation),
and less accurate the further it gets from any sample points (extrapolation),
but fundamentally there are no limits to the domain or range of the mapping.
When cast into a particular representation such as an ICC profile, there will
be limits imposed by the encoding of the RGB and PCS values though, implying a
certain gamut, but this limit is not related to the test chart.
The gamut of reasonable accuracy will be largely set by the test chart, but
something like an ICC profile has no way of representing this, even if there
was a criteria for computing it.
Graeme Gill.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden