Olympus-OM
[Top] [All Lists]

Re: [OM] (OM) OT Tina's scanning settings for K64 slides

Subject: Re: [OM] (OM) OT Tina's scanning settings for K64 slides
From: Peter Klein <pklein@xxxxxxxxxxxxx>
Date: Sun, 24 Jun 2012 17:51:51 -0700
Tina, we may be able to find a curve or profile that is a good starting 
point for you. But I have a feeling that the color balance is going to 
be highly individual, depending on the subject, the light, and even the 
age of the film.

Kodachrome is difficult to scan. It has its own rendition of colors, 
which is very different from other slide films.  We've come to accept 
them because for years Kodachrome *was* color, and we grew accustomed to 
the look. The difference between how our eyes see those dyes and how the 
scanner records them is greater than with modern E-6. The shadows are 
very dense and require careful boosting. And it may be that the 
different dyes shift differently with age. I've noticed in some of my 
old Kodachromes that some parts of the slide have a magenta cast, others 
a cyan cast. And shadows are always "too" bluish.

The following slide from New Hampshire's White Mountains in 1971 shows 
this very well. It's scanned with VueScan in multi-pass mode. Both the 
Kodachrome profile and the generic slide profile. Each had its good and 
bad points. I believe I ended up using the generic slide film profile 
and hand-adjusting to taste, because the Kodachrome profile made things 
look worse, not better.
<http://www.flickr.com/photos/24844563@N04/5521434976/sizes/o/in/set-72157626204203806/>
(Franconian Range from Cannon Mountain, NH)

I found that to fix the magenta casts, and I'd get things much to cyan 
overall, and vice versa. I could find no white balance that was even 
remotely neutral.

Kodachrome also tends to make Caucasian faces ruddy. Fix that, and the 
blues and greens go ugly. So in this one, I just got it to look as good 
as I could, and left well enough alone. When I tried to fix the ruddy 
skin or the magenta on the gray granite, other things went bad.
<http://www.flickr.com/photos/24844563@N04/5517030746/sizes/o/in/set-72157626204203806/>
(Overlooking Kinsman Pond)

Both these scans are the result of a lot of individual effort to get 
them to look reasonably like the original slide, which is very different 
than what E-6 slides would render.

To make them any "better" would require masking and color balancing of 
individual channels. Since I'm not scanning them for stock, only for 
personal record, I didn't think it worth the effort to go farther than 
what I did. I also found that I needed  a very different curve for 
high-contrast slides with much in the shadows than I did for subjects in 
even light.

A couple more--
1971, a very hazy day on Mt. Lincoln, looking at Mt. Lafayette:
<http://www.flickr.com/photos/24844563@N04/5508228287/sizes/o/in/set-72157626204203806/>

1982, me in Spokane, WA with my late friend, musicologist Hans Moldenhauer.
<http://www.flickr.com/photos/24844563@N04/5500473753/sizes/o/in/set-72157626204203806/>

--Peter
---------------
Tina wrote:

 > Thanks, Moose.  I really can't count on my memory of exact colors 
which has
 > been affected by the passage of time.  I remember the people themselves
 > vividly but a slight bit of blue or red in their skin or hair would be
 > beyond what I can remember!   I agree with you that what I have to 
work on
 > is matching the Kodachrome slides, most of which have never been 
projected
 > and have been stored in hanging file pages in dark slide drawers.  They
 > have not faded at all and look exactly like they did when I got them back
 > from Kodak.  It's true that Kodachrome has its own color palette which is
 > very different from E-6 films.  I'll be glad when I finish all of the
 > Kodachromes and get to the E-6's that I switched to in about 1993!
 >
 > I like what you did with the Vuescan software.  The third roll-over looks
 > very natural and I like that the shadows are so open.  Compared to the
 > original KR, your version has less contrast and is slightly cooler.  I'm
 > trying to duplicate your settings using a combination of Lightroom and
 > PSCS6 but adding a slight warmth and a little contrast.  The Maria 
Marisela
 > portrait was the first one I've tried and I found several better ones of
 > her after I worked on this one.  This has been a very busy week-end 
and it
 > looks like a busy week coming up but I'm hoping to find some time to
 > experiment more.
 >
 > In the meantime, my scanner is still going every hour I'm here 
storing all
 > of those raw KR scans away for further enlightenment!
 >
 > Tina
 >
 > On Sun, Jun 24, 2012 at 7:13 PM, Moose <olymoose [at] gmail.com> wrote:
 >
 > > On 6/24/2012 10:49 AM, Brian Swale wrote:
 > > > While watching the various examinations of Tinas scans, and 
suggestions
 > > > offered, I wonder (again) about the reflection of blue sky in 
very dark
 > > shiny
 > > > hair, and how much is a legitimate representation of the truth.
 > > >
 > > > Also, the degree of saturation of brown in skin which is habitually
 > > exposed a
 > > > lot to the sun. See below. The last one from Moose seems to have lost
 > > most
 > > > saturation, which surprises me.
 > >
 > > Brian, I think you misunderstand what I was doing with these most 
recent
 > > images. At first, I, and others, as in the
 > > links you make to Mike and CH's versions, made various efforts to take
 > > Tina's posted JPEGs and make them more realistic
 > > looking.
 > >
 > > I'm not sure why you choose to provide links to three images I posted
 > > individually. They are much better seen in the
 > > roll-over I posted. <
 > > http://www.moosemystic.net/Gallery/Others/Manley/First_Photo.htm> It is
 > > dead simple HTML, to which
 > > I can't see why you would have any objection, and makes viewing them
 > > together, in context - and most importantly - with
 > > labels easier.
 > >
 > > I believe you are making invalid assumptions about skin and hair 
colors.
 > > I've not been to the parts of Central America
 > > Tina visited. I have been to Costa Rica and many times to Mexico. 
It is my
 > > experience that people in different parts
 > > often have different body types, facial features, skin tones and hair
 > > color. My late wife, half Mexican had, to North
 > > Americans, black hair. She disagreed, saying, in effect, that in 
Mexico,
 > > where almost everyone has hair we would call
 > > 'black', many variations were recognized. Hair that is so black it 
seems
 > > to have some blue in it is one of those. Those
 > > of predominantly Olmec descent don't look like Zapotecs, or Aztecs, or
 > > Mixtecs, and so on and so on. In addition,
 > > different villages and areas have different mixing in of Spanish blood.
 > >
 > > The upshot is that I don't believe it is possible for someone who 
hasn't
 > > been there to make meaningful judgements, even
 > > speculations, about the 'truthfulness' of the hair and skin colors of
 > > people in Central American villages.
 > >
 > > For those reasons, and because Tina's professional work is known 
based on
 > > what is on the slides, my intent in this
 > > project has nothing to do with accurate depiction of the original
 > > subjects, " ... a legitimate representation of the
 > > truth." It is entirely about finding a simple, highly automated way 
to get
 > > scanning results that come close to matching
 > > the original slides. If blue highlights are in the original slide, I do
 > > not, for this work, care whether that is what
 > > the hair looked like to Tina's eye in the field, decades ago.
 > >
 > > The first image is not mine at all. It is the Raw scanner output Tina
 > > obtained from SilverFast. I presented it simply so
 > > everyone who has been struggling with her posted images and heard 
her talk
 > > about how dark the SF output is could see
 > > what she is talking about.
 > >
 > > The second image is also not about Moose correcting the image to 
something
 > > he prefers or thinks might be more realistic.
 > > It is simply what VueScan did with the SF output, using VS defaults.
 > >
 > > VS has a similar type of linear, unmodified scanner output 
available, and
 > > can 'scan' from those 'RAW' files. I often use
 > > this capability, scanning with the hardware and film into RAW files
 > > without making any adjustments. I can then scan from
 > > the files, making whatever adjustments I want. The first step is just
 > > mindless feeding of film into the scanner. The
 > > second is where corrections for film base, film color inaccuracy,
 > > exposure, WB, and so on may be done, as well as dust
 > > removal. The image may later be rescanned with different parameters
 > > without ever having to actually physically rescan
 > > it. So, for example Kodachrome that was scanned to RAW files with IR
 > > channel included before VS's new, improved ability
 > > to remove KR dust may be 'rescanned' for dust removal from the RAW 
files.
 > >
 > > The results in color correction, contrast, and highlights seemed to 
me a
 > > big improvement, as I would expect from
 > > application of gamma and correction for some KR idiosyncrasies.
 > >
 > > The third image is another VS 'scan', with adjustments to Black Point,
 > > Brightness and Curves in VS to lighten the
 > > overall image.
 > >
 > > It falls, to my eye, into a middle ground between dark shadows in 
bright
 > > sunlight and some visibility in those shadows.
 > > I won't know if it should be lighter or darker, more or less contrasty,
 > > until I get feedback from Tina.
 > >
 > > Further, I'm not trying to get this one image 'perfect'. I'm 
looking for a
 > > process that will work for the vast majority
 > > of the half-million slides she is scanning. These settings work pretty
 > > well for two of the samples she sent me. Neither
 > > is perfect, but both look at least decent and natural, while 
retaining all
 > > the histogram so that any ones desired could
 > > be customized for excellent printing, etc.
 > >
 > > The same settings don't work as well for the third sample, leaving the
 > > bottom of the histogram 'high and dry'. A quick
 > > pull down of the histo in Levels is all that's needed, but I'd like to
 > > find a more generalized solution requiring no
 > > individual intervention. The goal is a process that may be applied to
 > > batches of scans automatically.
 > >
 > > > To what degree is the acceptance dependent on Tins'a memory of 
what the
 > > colour seemed to be?
 > >
 > > As above, I'm not trying to deal with that, but with what's on the 
slides.
 > > Our picture of what the world and its peoples
 > > look like has been formed through looking at images made from 
Kodachrome
 > > slides far more than actual experience.
 > > Duplicating these KR slides in scans would be just wonderful, I think.
 > >
 > > I rather suspect that cinematographers for nature movie/videos have the
 > > National Geographic and other printed images
 > > from Kodachrome much in mind as they set the colors for their work. 
Paul
 > > Simon was both right and wrong. They have taken
 > > our Kodachrome away, but not the enduring color legacy in innumerable
 > > images and our memories.
 > >
 > > Voluble Moose
 > >
 > > --
 > > What if the Hokey Pokey *IS* what it's all about?
 > >
 > >
 > > --
 > > _________________________________________________________________
 > > Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
 > > Archives: http://lists.thomasclausen.net/mailman/private/olympus/
 > > Themed Olympus Photo Exhibition: http://www.tope.nl/
 > >
 > >
 > >
 >
 >
 > --
 > Tina Manley, ASMP
 > www.tinamanley.com

-- 
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz