Nikon Scan 4 and contrast increase in scan

  • Thread starter Thread starter Mike Engles
  • Start date Start date
M

Mike Engles

Hello

I have been investigating NikonScan4 with my LS50.

I conclude that whilst the preview can be made to look reasonable, Nikon
Scan whilst making the scan clips deep greys to black, hence increasing
contrast. With Kodachrome this can results in deep purple casts in the
shadows. I use no clipping in my previews,Nikon Scan seems to apply it
anyway. This happens in both Nikonscan as a standalone or import into
Photoshop.


Here are two links that have screen grabs, showing preview and scan.
They both have RGB measurenments.

http://www.btinternet.com/~mike.engles/mike/preview_scan4.jpg
http://www.btinternet.com/~mike.engles/mike/preview_scan5.jpg

You can also see most of the settings I have been using.

Mike Engles
 
| Hello
|
| I have been investigating NikonScan4 with my LS50.
|
| I conclude that whilst the preview can be made to look reasonable, Nikon
| Scan whilst making the scan clips deep greys to black, hence increasing
| contrast. With Kodachrome this can results in deep purple casts in the
| shadows. I use no clipping in my previews,Nikon Scan seems to apply it
| anyway. This happens in both Nikonscan as a standalone or import into
| Photoshop.
|
|
| Here are two links that have screen grabs, showing preview and scan.
| They both have RGB measurenments.
|
| http://www.btinternet.com/~mike.engles/mike/preview_scan4.jpg
| http://www.btinternet.com/~mike.engles/mike/preview_scan5.jpg
|
| You can also see most of the settings I have been using.
|
| Mike Engles


Since you see the problem even when you import to PhotoShop, the question
is whether the clipping/tingeing is being done by the NikonScan TWAIN
software or by the ROC software (from ASF). I'd first try to eliminate ROC
as the cause by rescanning the slides with ROC off and then comparing the
preview to the final scan. If you still see the problem, you could try a
different twain (such as VueScan).

You could also send a query and the images to the online Nikon tech support
folks - they've been very good about answering my questions in the past.

Jean
 
Jean said:
| Hello
|
| I have been investigating NikonScan4 with my LS50.
|
| I conclude that whilst the preview can be made to look reasonable, Nikon
| Scan whilst making the scan clips deep greys to black, hence increasing
| contrast. With Kodachrome this can results in deep purple casts in the
| shadows. I use no clipping in my previews,Nikon Scan seems to apply it
| anyway. This happens in both Nikonscan as a standalone or import into
| Photoshop.
|
|
| Here are two links that have screen grabs, showing preview and scan.
| They both have RGB measurenments.
|
| http://www.btinternet.com/~mike.engles/mike/preview_scan4.jpg
| http://www.btinternet.com/~mike.engles/mike/preview_scan5.jpg
|
| You can also see most of the settings I have been using.
|
| Mike Engles

Since you see the problem even when you import to PhotoShop, the question
is whether the clipping/tingeing is being done by the NikonScan TWAIN
software or by the ROC software (from ASF). I'd first try to eliminate ROC
as the cause by rescanning the slides with ROC off and then comparing the
preview to the final scan. If you still see the problem, you could try a
different twain (such as VueScan).

You could also send a query and the images to the online Nikon tech support
folks - they've been very good about answering my questions in the past.

Jean


Hello

It happens even when you do not use ROC, GEM or DEE, just curves.
I have another grab to show that.I will post it tomorrow.
I have sent the images to Nikon in the UK.

Mike Engles
 
i think you have discovered one of the reasons some people use other
software with the Nikon, particularly with dark KC. This was happening
with the 4000 and NS 3.x as well. i do not know the reason Nikon did
this and apparently still does.

Frank
 
degrub said:
i think you have discovered one of the reasons some people use other
software with the Nikon, particularly with dark KC. This was happening
with the 4000 and NS 3.x as well. i do not know the reason Nikon did
this and apparently still does.

Frank


Hello

Yes it did happen before. The problem might be worse, because the
natural contrast of the scanners has improved so much.

Also I have never liked the results fron Kodachrome/Vuescan.
Silver fast also has its problems.

Mike Engles
 
Mike Engles said:
.....

Also I have never liked the results fron Kodachrome/Vuescan.
Silver fast also has its problems.

Mike Engles

I've been scanning some 20-year-old Kodachromes for 3 days now with my
LS4000 and Nikon Scan 4, and it's a royal pain in the neck. Besides
being generally underexposed, the balance between the RGB channels shifts
drastically, and the only luck I have had fixing this involves increasing
the analog gain on the red channel (a lot) and the green channel (some)
while knocking down the gamma on the blue channel. But this leads to
spiky histograms unless I do it in 14 bit mode. Yesterday I only got the
spiky histograms in the blue channel (due to using curves in NikonScan);
today I get them in all channels if I even touch the gain. If someone knows
of a solution for this (other than buying Silverfast--I do have VueScan
but haven't tried it for this), I'd love to know about it.

Gary Hunt <[email protected]>
 
Gary said:
I've been scanning some 20-year-old Kodachromes for 3 days now with my
LS4000 and Nikon Scan 4, and it's a royal pain in the neck. Besides
being generally underexposed, the balance between the RGB channels shifts
drastically, and the only luck I have had fixing this involves increasing
the analog gain on the red channel (a lot) and the green channel (some)
while knocking down the gamma on the blue channel. But this leads to
spiky histograms unless I do it in 14 bit mode. Yesterday I only got the
spiky histograms in the blue channel (due to using curves in NikonScan);
today I get them in all channels if I even touch the gain. If someone knows
of a solution for this (other than buying Silverfast--I do have VueScan
but haven't tried it for this), I'd love to know about it.

Gary Hunt <[email protected]>




Hello

Try This.
Use the LCH curve to boost the curve by 2.0 in the luminosity channel.
Use analogue gain to give max white of 250, no high clipping.
Nothing else.

This shifts the black really high.
Scan in 14 bit into Photoshop if you have it.

This time the scan will be slightly brighter than the preview.

In Photoshop use curves to clip in a 'S' shape so that it looks good
again. You should have very little colour shift at the low end.

Use the individual channels to correct for any small shifts as well as
the grey picker..

If you do not have Photoshop, you can use the curves in Nikon scan in
the same way.

This by passes the slope limiting in Nikonscan as well as the the
problem with the scanner where its low end output is inconsistent.
 
I've been scanning some 20-year-old Kodachromes for 3 days now with my
LS4000 and Nikon Scan 4, and it's a royal pain in the neck. Besides
being generally underexposed, the balance between the RGB channels shifts
drastically, and the only luck I have had fixing this involves increasing
the analog gain on the red channel (a lot) and the green channel (some)
while knocking down the gamma on the blue channel. But this leads to
spiky histograms unless I do it in 14 bit mode. Yesterday I only got the
spiky histograms in the blue channel (due to using curves in NikonScan);
today I get them in all channels if I even touch the gain. If someone knows
of a solution for this (other than buying Silverfast--I do have VueScan
but haven't tried it for this), I'd love to know about it.

If you think that's bad you should try it with an LS-30!!! If you look
around this newsgroup you'll see lots of pulled hair on the floor and
numerous climbing scratchmarks on the walls... That's all mine! ;o)

But seriously, Kodachromes on Nikons are a nightmare although I was
under the impression that newer scanners - like the LS4000 - do a
pretty good job.

I've been trying to streamline the process because tweaking each slide
individually is just plain excruciating, as you've discovered. And
there just isn't enough time in this Universe to do the job properly.

My approach in broad strokes (still "work in progress") is to try and
get the maximum out of the scanner without any software adjustments in
Nikon Scan (no curves, brightness, etc). I leave all post-processing
to Photoshop. In particular this means:

1. Turn off Nikon Color Management! Very important! NCM just makes
matters *much* worse i.e. it *amplifies* blue!!!

2. Turn off Auto Exposure! Nikon Scan has a "random exposure
generator" ;o) which (at least in my setup) out of the blue (sic)
would produce diametrically different back to back scans without any
change in parameters. Even when it works, it creates a different color
balance in each scan which makes streamlining cast removal impossible.

3. To get rid of spiky histograms (I called them "weird" histograms
a.k.a. "comb" histograms - there's a thread on it in the archives) the
workaround is to set Interpolation to Bilinear and fix cropping. Once
spikes are gone stick with the same cropping. (Without going into too
much detail the spikes are the result of gamma conversion.)

I also have to do contrast masking but your scanner has sufficient
dynamic range so that's not applicable in your case.

I have no definitive suggestions on streamlining color cast removal
but the consensus (at least for difficult, very dark slides) seems to
be: assuming blue is level (0), boost red by about 1 Analog Gain and
green by about 0.2. (I know, the "about" bit really bothers me too.)

Another automatic approach (similar to Auto Levels) is to boost Analog
Gain until all three channels bump up against the right histogram
edge. This will give you a good color balance in most cases. However,
just like Auto Levels, in some cases (depending on image content) this
may actually introduce a color cast, but at least you will be getting
the maximum dynamic range.

Of course, like auto exposure, this will also create a different color
balance with each scan which makes streamlining (any residual) cast
removal later impossible.

That's where I'm at right now, still chasing my tail and cursing
Nikon... ;o)

As for VueScan, I for one am not exactly a fan... Even forgetting its
"interesting" user interface, its lack of controls (no individual
channel Analog Gain adjustment) makes it totally unsuitable for the
task.

In regard to color balance, VueScan is like a throwaway Instamatic
camera. Point and shoot. When it works, it works well, but when you
get a difficult situation (e.g. a very dark Kodachrome) you end up
with something looking like a false color image. Totally unacceptable
(for me, anyway, although many differ).

One VueScan plus is the ability to boost exposure more than NikonScan
(which is effectively limited to 4 AG). However, instead of using
Analog Gain i.e., EV values, VueScan insists on its own unique
"interesting" scale... :-/ And, of course, you still can't adjust
individual channel AG which renders this advantage meaningless.

Don.
 
Back
Top