Scanning negatives as positives and converting in Photoshop

Preface:

I'm translating my friend's instructions from Russian to English, making
my own comments too. Please tell me if something is incorrect or unclear
there. I will add this to my site http://www.foto-group.com or
http://www.vkelman.com later on.
Although it sounds complicating, it is quite easy after you get used to it.

Examples:

My first examples (Fuji ZI645GA medium format, Fuji Reala)
http://www.photosight.ru/photo.php?photoid=2043792
http://www.photosight.ru/photo.php?photoid=2045567

My friend examples (fantastic color with Reala! But he read Margulis and
know much more about color managing then I do):
http://sergt.msk.ru/jpeg/album/Kodak%20400NC/slides/Image14.html
http://sergt.msk.ru/jpeg/album/Reala/slides/Image2.html

http://www.amazon.com/Professional-Photoshop-Classic-Guide-Correction/dp/0764536958

Instructions:

1. Scan negative as positive.

2. Do not do any color manipulations using scanner software.

3. Try to point a scanner to non-exposed space between frames (or near
film edge) and adjust scanner's color "guns" (that's hardware
adjustment, not a software!) in such a way that non-exposed space looks
as neutral as possible. Make sure not to cut edges of color histograms,
it is better to have some borders around them.
It is a first preliminary and rough step of eliminating negative's color
mask. Don't try to make it perfect on this step.
Not all scanners allow to do that. Nikon Coolscan 9000 does. You need to
find how to do that with your software. In NikonScan it's called "Analog
gain". In SilverFast click on "Options => Special => Lightness".

2. Try to set gamma to 1.0 on scanner software. Setting gamma to 2.0 or
similar value squeezes a tonal diapason and results in a lost of smooth
tonal regions. If histogram doesn't fit, make gamma 1.4, 1.6, etc...

5. Do not assign any color profile during a scan.

6. Download a special conversion curve film-negative.amp
(http://www.aim-dtp.net/aim/download/film_gamma.zip) and put it into an
appropriate Photoshop sub-folder (for example, "Presets").

7. Try to make non-exposed space between frames as neutral as possible
by adjusting colors. Make sure not to cut histograms. This step is
convenient to do using ColorWasher plug-in.
http://www.thepluginsite.com/products/photowiz/colorwasher/index.htm

8. Open scan in Adobe Photoshop. Go to Layers => New Adjustment Layer =>
Curves and hit "OK" to create a curves adjustment layer. Then use
"Load" button to load film-negative.amp (change file extension in
drop-down box from .ACV to .AMP to be able to see it). Hit "OK" to get *
dark and low-contrast converted image with bad colors *.
Note: you can use your own adjusted curve instead of film-negative.amp,
but it is easier to start with a standard one.

9. Assign Adobe RGB color profile or whatever you prefer.

10. Do Layers => New Adjustment Layer => Levels and adjust colors in a
regular way. Basically, use left and right markers separately on each
color histogram to cut empty spaces and then use central markers to
fine-tune result. You'll be immediately amazed by quality of resulting
colors.
(You can use Layers => New Adjustment Layer => Curves instead, if you
are experienced to work with it.)

11. Use your wisdom or Dan Margulis' one to make colors even better.

12. Send "thank you" to me

Vladimir Kelman wrote:
> From my experience Negafix + film profiles never worked fine. I use
> Nikon CoolScan 9000. Colors are awful and hard to fix. Negafix
> interface is overcomplicated and non-intuitive. It's not a Howtek
> problem, it is Silverfast problem. For Nikon film scanners a standard
> NikonScan software which does not allow to specify negative film
> profiles makes a better job of removing orange mask and automatically
> produces better colors. Recently I came across a clever technique of
> scanning negatives as positives...

Jake Conner wrote:
> Care to share your technique with the group? In my experience,
> negative profiles never get it quite right (I've used Nikon's,
> Epson's, Lasersoft's, Scanmate's, and Imacon's), I usually just align
> the endpoints for each channel to get it in the ballpark, then do
> the rest with curves in Photoshop...

Messages in this topic (0) Reply (via web post) | Start a new topic

.

Thank you for posting this. With this method I get colors are as good as those I get from
from an Ektachrome. I use a Coolscan 9000 with Silverfast 6.5.0r6 and scanned Portra 160
VC film.

There was some tweaking before it looked right for me.

-My scanner is calibrated for Ektachrome, I get better results with calibration on and the
profile embedded.

-To get the non-exposed space between the frames neutral I set the lamp control in
Silverfast (=analog gain in Nikon Scan) to White +9, R -10, G+2, B +8. Multi-Exposure was
off, it didn't work.

-In Silverfast a Gamma of 2.0 worked for me but not a gamma of 1.6. In the first
adjustment layer I used the film-negatif.amp curve but with a less strong correction (one
point on the RGB curve: input 77 output 114 but otherwise all the same curve).

-In the second adjustment layer I used the curve: RGB in0/out0, in56/out66, in132/
out195 and in191/255. The Red curve was in0/out0, in 101/out113 and in155/out255.
No correction on the green curve. The blue curve was in0/out0, in141/out133 and in255/
out255.

I am not a scientist and worked by visually evaluating the results.

Messages in this topic (0) Reply (via web post) | Start a new topic

.