Re: filmscanners: pushing dynamic range on the nikon 4000ed

[Wayne Williams <wayne_williams@mac.com>]

hello bill,

again, thanks for your responses.  in thinking about your responses i'm
wondering if i still have an issue with the front end part of the process -
specifically the scanner profile and the setup of the nikon software.

what i see is that some of my scans match quite well to the slide after
importing into photoshop and assigning the scanner profile, and some miss
quite badly.

either the profile, the lighting, or my nikon scan settings are off.

you've helped me understand the issue of lighting and you are probably
correct that my light are closer to 3000 degrees - which is curious because
i would expect a yellow cast, not the blue one that i tend to see if my
prints are off.

i'm going to assume that the scanner profile is correct, monaco ez color
doing its job, until i eliminate the other variables.

so that leaves me wondering about the setting for auto exposure.  right now
i made my profile with auto exposure set to "on" - and this is how i make
all of my scans.  will this cause the scanner to read different images in
different ways such that the profile will have varying degrees of "fit".
how do you have this set?

thx,

w.





on 11/20/01 3:45 AM, Bill Fernandez at bill_sub@billfernandez.com wrote:

> At 11:34 PM -0700 19-11-01, Wayne Williams wrote:
>> ...using the gain to make the histogram adjustment versus the curve
>> adjustments
>> seemed to work well.  and it fit with the workflow suggested by monaco which
>> was to use the scanner in the same state as it was in when the initial
>> profile is made.  specifically not to use any of the color adjustments on
>> the scanner side - just pass on the raw bits.  so this assumes that
>> adjusting the gain will not shift the color characteristics of the scanner,
>> thus invalidating the profile...
> 
> BF: This is the recommendation I've seen/heard most often, but I've
> done some informal experiments both with my film and flatbed scanners
> where I adjusted the gamma and black/white points before scanning
> then applied my profiles and the results seemed pretty good.  I
> intend to do some more carefully controlled experiments someday when
> I get a chance.
> 
> 
>> i made two scanner profiles last night, one with the scanner set to
>> kodachrome and one for the generic positive setting.  i then scanned in the
>> kodak q60 target two ways.
> 
> BF: Check your Q60 target.  I suspect it's Ektachrome, in which case
> the Kodachrome film setting is not appropriate.
> 
> 
>> one with the scanner set to kodachrome and using
>> the kodachrome profile and then with the scanner set to positive using the
>> positive profile.  i assigned the correct profiles to both images in
>> photoshop after scanning via the twain interface.  i used a levels
>> adjustment in both images by setting the highlight off the brightest square
>> along the greyscale to eliminate variations associated with sampling from
>> different points within an image.  the images looked identical to me in
>> everyway - so i did not see the benefit of the kodachrome adjustment.
> 
> BF: I'm under the impression that the color characteristics of most
> color slide films are more alike than different, except for
> Kodachrome which is distinctly different.  Therefore you should only
> set the film type to Kodachrome when scanning Kodachromes, and use
> the generic positive setting for everything else.  It may be that
> your custom profiles are just doing a good job of compensating for
> your (perhaps erroneously?) using the Kodachrome setting.
> 
> 
>> perhaps i am missing something.  at any rate i'm using it based on your
>> advice.
> 
> BF: My advice is to use the Kodachrome settings ONLY when scanning
> Kodachromes.
> 
> 
>> also, i use elite chrome 200, not sure of the difference from
>> kodachrome or if it is significant in relation to using those profiles.
> 
> BF: I've never used Elite Chrome, but I suspect the optimal film type
> setting in NikonScan would be generic positive.  Again this is based
> upon my (anecdotal) understanding that Kodachrome's color
> characteristics are distinctly different than all other slide films.
> 
> 
>> this brings up a work flow question for me.  the folks at monaco said that
>> once i apply a scanner profile, assuming a correct monitor calibration, the
>> image should come pretty close to matching the slide on the slide table
>> without any photoshop adjustments.  in my experience i have a significant
>> blue shift even after i assign the scanner profile.  however, if i make a
>> good levels adjustment it then does match the slide quite correctly.  is
>> this your experience?
> 
> BF: I don't have a light table, so at present I either hold a slide
> up to one of my room lights or up to the screen (and use the monitor
> as a light table).  I'm finding that my profiled scans are a very
> close match when I use the monitor as a light table.  If your light
> table has a different brightness or color temperature than your
> monitor this might account for the discrepancy you're seeing.  So far
> it appears that the light tables that closely approximate D50
> illumination cost hundreds of dollars (GTI, Gretag MacBeth, etc.) and
> I haven't found one I can afford.
> 
> 
>> one final question, for this e-mail at least, i'm going with a straight 1.8
>> / 5000 degree setup.  this is how i've set my gamma in nikon scan, and how i
>> have calibrated my monitor.  i believe that my lighting is d50 halogen -
>> results seem good.
> 
> BF: As far as I know, halogen bulbs run in the 3000K's and are far
> warmer than the D50 standard.  The only halogen bulb I know of that
> comes close to the 5000K of D50 is the Solux bulbs at about 4,700K.