Re: Scanner Profiles?
Re: Scanner Profiles?
- Subject: Re: Scanner Profiles?
- From: email@hidden
- Date: Fri, 12 Oct 2001 20:34:03 +1100
Dave writes:
>
Once I profile my scanner I then apply that profile to my untagged
image in photoshop and save that as my raw file. <
If the scanner software does not use the profile and does not tag - then
yes.
>
I then convert to my working space to edit the image and save that
as a final master.<
Spotting and scratch removal are often done in the raw scanner RGB,
before this 'clean' untoned file is finally saved with the device tag.
This requires Photoshop 6 or another editor which will preserve the
scanner tag.
For more general post scanning editing, such as endpoints,
gradation/toning, colour balance etc - then yes, converting from raw
device profile to your preferred RGB working space is the suggested
workflow for most users. A second conversion would take place to the
output profile or workspace.
>
What if going straight to working space gives me a good result ?
Should I avoid the scanner profile conversion ? <
This would mean that you do not trust the profile.
You would be assuming that your chosen neutral RGB workspace is the same
as the device or scanner space. This is not often the case.
Some consumer gear capture with sRGB as a default - so simply opening
into sRGB is the ideal method for this subset of users.
Most other workspaces do not match the device.
You need an input and an outpt profile - to correctly describe the
scanner RGB to the workspace RGB.
If you simply open a file without converting, it will be _assumed_ that
this is the assigned profile.
Many users like Adobe RGB (A98). But not many devices RGB is close to
this 'idealised' flavour of RGB. Our EverSmart Supreme is close to A98
in raw RGB - but we still use the profile for the scanner in the scanner
software.
If simply assigning a neutral RGB workspace to a scanner or digital
camera file - ColorMatch RGB or Apple RGB might be better, since they
are monitor type spaces which accept raw unprofiled RGB input better
than the wider gamut spaces, which do not often work well without a
correct input conversion.
I think there might be more workflow and output issues with simply
assuming a RGB, which does not require a conversion - than converting
from device RGB to workspace RGB. If the input is correct, the results
should be better than simply making an assumption on the RGB.
But this is only the 'textbook' answer - if your 'pre ICC' method works
for you...but then why go throught the process of custom profiling?
Hope this helps,
Stephen Marsh.