Re: Kodak IPB profiles
Re: Kodak IPB profiles
- Subject: Re: Kodak IPB profiles
- From: Joel <email@hidden>
- Date: Wed, 22 Nov 2000 10:23:23 -0600
At 12:57 AM +0000 11/22/00, neilB wrote:
We are aware that the black and white points of the scanned IT8
should measure at about 5 and 250. But whatever the endpoints of the
IT8 are set to when scanning, say as extreme as 245 and 15 when the
scan is checked in Photoshop, the profile always maps the data right
up to 0 and 255 when we profile to profile the IT8 scan to test the
profile. Is this a Kodak IPB 4 feature?
No matter how large the destination workingspace [even tried Kodak
ROMM] I find that in at least one channel a small amount of data is
being apparently clipped since it rests at 255 or 0.
I had expected IPB to map to 5 and 250 say for data integrity, it is
a rare occasion when a file needs to have black and white at 0 and
255.
I was considering buying ProfileCity's own scan profile generator,
anyone tried it please??
thanks
neil
I think this can depend on the rendering intent you have chosen. If
you are using Relative Colorimetric then you might see this as a
natural mapping - into any color space.
Opinions?
Steve Upton
Opinion #1:
Photoshop re-maps input/output data over the range of 0-255 during
any conversion which re-maps input/output data. A p2p conversion does
this: i.e. look at your levels (command-L/control-L) and histogram
before your conversion. Do your conversion, check levels and
histogram again. Normally your levels will adjust to the new space
but not necessarily over the entire 0-255 range. What alters the data
in the p2p is the (usual) space-to-xyz, xyz-to-space conversion in
the profile(s). Perceptual and black point compensation make for
broader re-mapping, relative with no BPComp allowing less, etc.
Hope this gets a thread going that attracts more concrete theories.
Good question.
--
Joel Johnstone
designtype
Winnipeg Manitoba Canada
email: work: email@hidden
color geek in residence, reality notwithstanding