Scanner, Step Tablet and Photoshop Curves

[Hugh Jass]

I have a T2115 Step Tablet from Stouffer. Can I use this to create a curve
in photoshop to correct for my scanner's (Epson 3200) response to Black and
White negative material? Does anyone actually do this? Am I wasting my time
or will I see a big improvement? Thx.

 

[Bart van der Wolf]

I have a T2115 Step Tablet from Stouffer. Can I use this to create a
curve
in photoshop to correct for my scanner's (Epson 3200) response to
Black and
White negative material? Does anyone actually do this? Am I wasting
my time
or will I see a big improvement? Thx.

Yes, it can be used and I have posted results from my Epson 2450 with
a similar (Kodak steptablet No.3) in this group
(http://www.google.com/[email protected]).

The diffuse lighting of the scanner will reduce the Callier effect, so
different grain structures between target and film won't have as much
an effect as they would with more collimated lightsources. As long as
you realize that the Optical Densities will undergo a potentially
non-uniform Gamma adjustment for the final file output (unless you
save as linear gamma files), they are a useful tool once calibrated.

Bart

 

[Hugh Jass]

Yes, it can be used and I have posted results from my Epson 2450 with
a similar (Kodak steptablet No.3) in this group
(http://www.google.com/[email protected]
4all.nl).

The diffuse lighting of the scanner will reduce the Callier effect, so
different grain structures between target and film won't have as much
an effect as they would with more collimated lightsources. As long as
you realize that the Optical Densities will undergo a potentially
non-uniform Gamma adjustment for the final file output (unless you
save as linear gamma files), they are a useful tool once calibrated.

How do I "save as a linear gamma file"?
I think after looking at your old post and plotting my data all I really
need to do is adjust VueScan "exposure" control so my theoretical and test
data graphs line up a little closer. For the most part the scanner is very
linear until about 2.7 and above. Any curve that I would build to adjust the
scanner would pale in comparison to the changes I tend to make using PS's
contrast and levels controls.
One thing I did noticed is when I scanned the Stouffer strip side ways
(rotated 90 degrees) rather than length ways I got different results. The
densities reached above 3 consistently, I don't know why this is though.
Something to do with the CCD?

 

[Bart van der Wolf]

SNIP

How do I "save as a linear gamma file"?

Depends on the software used. For my Epson scanner I use VueScan
exclusively. Saving as a 48-bit Raw will give me what I want. Whatever
you choose, it must be the same method as used for the actual scans,
otherwise the calibration makes no sense.

I think after looking at your old post and plotting my data all
I really need to do is adjust VueScan "exposure" control so
my theoretical and test data graphs line up a little closer.
For the most part the scanner is very linear until about 2.7
and above. Any curve that I would build to adjust the scanner
would pale in comparison to the changes I tend to make using
PS's contrast and levels controls.

Be very cautious with Photoshop's Levels control, as the center
control is not a simple Gamma correction. Photoshop is also not a
scientific (photometric) application, but rather a pre-publishing
image processor, where the looks of the image are more important (with
room for improvement) than the mathematical accuracy.

One thing I did noticed is when I scanned the Stouffer strip side
ways
(rotated 90 degrees) rather than length ways I got different
results. The
densities reached above 3 consistently, I don't know why this is
though.
Something to do with the CCD?

Depends on the settings and the actual differences you saw. The
scanner has the possibility to scan at a smaller stepper motor pitch
than the CCD sensor pitch which could cause a difference at the
highest resolution, maybe that's what happened, maybe it is jest a
scanner warm-up issue.
Anyway, you should average an area (the ISO recommends 64x64 pixels)
before taking density readings. Vuescan allows to average by using the
"File size reduction" on the Output tab. A value of 6 (or more) would
be adequate for that.

Bart

 

[Hugh Jass]

SNIP

Depends on the software used. For my Epson scanner I use VueScan
exclusively. Saving as a 48-bit Raw will give me what I want. Whatever
you choose, it must be the same method as used for the actual scans,
otherwise the calibration makes no sense.

What's the difference between saving as RAW and saving with the Color
settings on VueScan set to NONE? I'm scanning B&W 16 Bit, BTW. I also use
VueScan.

 

[Hugh Jass]

Are you "working" the RAW file right in your photo editing software? What's
the advantage to doing this?

 

[degrub]

Likely "none" has gamma applied by Vuescan. Raw is the bits direct from
the scanner.

 

[Bart van der Wolf]

SNIP

What's the difference between saving as RAW and saving with
the Color settings on VueScan set to NONE? I'm scanning B&W
16 Bit, BTW. I also use VueScan.

From the help file:
None
The black and white points aren't used at all, and the image is only
corrected for the CCD's color response (if the Media option is set to
Image) or by the film's color response. This image is gamma corrected.

Bart

 

[MPA]

I have a T2115 Step Tablet from Stouffer. Can I use this to create a curve
in photoshop to correct for my scanner's (Epson 3200) response to Black and
White negative material? Does anyone actually do this? Am I wasting my time
or will I see a big improvement? Thx.

http://www.littlecms.com/iphoto/
maybe these free tools can help.