Re: Camera Raw
Re: Camera Raw
- Subject: Re: Camera Raw
- From: Graeme Gill <email@hidden>
- Date: Mon, 24 Feb 2003 14:26:06 +1100
Chris Murphy wrote:
>
What we need to understand is that the ICC spec is a work in progress,
>
and when it comes to the input profile class, was designed with
>
scanners in mind, not digital cameras. This means the likely
>
possibility that the spec is simply insufficient in the area of
>
characterizing digital cameras; and it doesn't mean that it won't work
>
in certain situations, it just means it's not nearly as reliable as it
>
needs to be.
I'll put it another way. ICC profiling works within a certain
context. That context is one in which white points and "tonal
range mapping" from real world light levels, to a limited range
color reproduction device values, has already been performed.
Both transformations are the subject, to some degree or other,
of artistic choice. ICC profiles and the software that
supports their creation and use doesn't handle either
of these operations.
Input devices, particularly digital cameras, have to allow
these operations to be performed. On top of this is the
transformation from the camera sensor colorspace to a
device independent colorspace. In the CGI world (which
also has to deal with mapping synthetic "real world"
light levels to limited range RGB spaces), there is
still considerable work going on in developing semi-automated
tonal mappings, that simulate the human visual system, so
as to be able to convert real world light levels into captured
images that subjectively look like we would have perceived them.
There has also been a great deal of work in developing
automated white point estimation algorithms in the digital
photography world.
The challenge is out there to combine all these elements
together. It seems desirable that there be a way of capturing
and representing an input devices color behaviour in a way that
is valid over its full exposure and sensor dynamic range, and
independent of the desired white point of the final image.
This would then allow a workflow in which the raw camera data can
be converted into device independent space, white point
transformed, and finally automatically and/or with artistic
guidance tone mapped into the more limited color spaces
generally used for image storage prior to reproduction.
I gather there are various offering out there now that
attempt some of this, but perhaps there is still a way to
go before it is all mature, and there is (for instance)
a standardized way of representing the color behaviour of
an input device in a sufficiently flexible way.
Graeme Gill.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.