About every two weeks I feel the urge the write something film-related.
Well, for today's episode:
I've been scanning much film. Processed 7 rolls a couple weeks ago and just
now getting the time in the evening to babysit the scanner. Maybe I should
pay my daughters to do this grunt work.
Much of the film has been ancient long-since expired Fujifilm NPZ 800
(always frozen) that I picked of three packages of for a song. It's been a
little hit-and-miss with this film. I believe the age has affected it a bit,
probably fogged a touch by xrays or who knows what. But the main problem
I've run into is blocked-up highlights. Scanning has proven to be a touch
pickier than I expected.
So, on two different rolls, I did photograph the IT8 target as well as my
entire test fixture. I tested the IT8 target with flash and in direct sun.
Wow, what a revelation. Folks, can you save yourselves a lot of grief and
just buy yourself an IT8 target to calibrate your cameras and films? You may
THINK something stinks, when in reality, it's YOU.
OK, so NPZ 800 has grain the size of golf-balls, right? Well, yes and no.
It's not so much that the grain is enlarged, because the Fujifilm and Kodak
consumer ISO 200 print films have grain about the same size. What is
different, though, is that NPZ is gritty. It doesn't have much chroma noise,
mostly luminance noise. This is opposite of, say Gold 200, which has color
dye clouds the size of footballs, but much less luminance noise. Frankly,
the NPZ is easier to deal with.
I wasn't happy with the way the colors shifted (specifically the greens)
with NPZ, but did discover an issue with the auto-levels not tracking the
base color correctly. Also, it is critical to get the right amount of
exposure into the scan because of how the highlights block up. When they
block up, it's quite nasty. This has been the most difficult aspect of
scanning this film. When left to its own devices, scanning software will
almost always throw a green cast over NPZ scans.
Feeling a bit lazy, and wanting to speed the process up, I decided to
shortcut a couple rolls of film as the images were family stuff not
requiring super-duper-perfection. As an experiment, I used VueScan and set
the scan resolution to 2000 dpi instead of 4000 dpi. Well, for some reason
Mr. Hamrick has his software setup with the Coolscan V-ED to scan at 4000
dpi and then scales it downward to the chosen resolution. Looking through
some things he has written, I think he has implemented some "special sauce"
in the algorithm, but whatever it is, serious kudos to him because the
resulting file is quite incredible and much easier to post-process than at
native resolution.
OK, maybe not incredible from the perspective of knocking the Nikon D3 off
the high-ISO perch, but definitely worthy of usage. The funny thing is the
resized file is very similar in size and character to a carefully groomed
E-1 image at ISO 800. The noise is gritty, but resolution is good.
In my test fixture, I did find it interesting that the resolution (of the
normal scan) is slightly better than E-3 and D2X files. Noise obviously
masks many of the nuances, but otherwise it was very fascinating that an ISO
800 film was matching details that E-3/D2X files were attaining at base-ISO.
In standard usage, I'm not really seeing this resolution advantage because
my cameras are usually handheld and, well, the OM-system is notorious for
shutter-shock. Also, I have to consider specific lens issues...
What about the IT8 target, you ask? What about the colors and dynamic range?
Scanned with either software, the IT8 target's colors pretty much gob smack
you with their accuracy and saturation. Reds are red, Greens are green,
Blues are blue, Yellows are yellow, Cyans? Yup, you guessed it--cyan. Oh,
and the Magentas are even magenta. Go figure! The entire gray scale from
white to black you can see every single step. Usually, you get a few of them
merging and that line between the patches disappears. Not so with these
scans. However, the noise is high at all brightness levels. Blacks aren't
really black because of the noise floor. Whites aren't really white because
of the noise floor. I did do some playing around with several
noise-reduction techniques, but the problem there is they tend to average
out the noise and the base brightness causing a loss of dynamic range.
That's kinda an issue because between the noise, the colors and densities
are correct. If you noise-reduce, then you lose contrast which then requires
you to bump and slide the curves.
Every single color patch on the IT8 target is clearly defined and looks very
good to the human eye. I haven't profiled it yet because of the noise issue,
but to my human eyes, the scanned image tracks the original IT8 extremely
well and has a characteristic I haven't seen before with digital cameras:
Each of the color ladders in the middle portion of the target maintain equal
brightness all the way up. Usually what happens is you'll lose separation
between two or three patches of a given color at some point goint up the
brightness scale.
Back to the scanned images, I was doing a bunch of people photography. It
never ceases to amaze me how wonderful skintones can be with these
professional films. Getting the skintones right is easy and natural. And
getting the skintones right doesn't turn your backgrounds into the uglies.
One technique I've been learning of late, is how to convert color images to
B&W. One of the techniques I've been playing with is taking the color image
and extracting the luminance channel and the red channel into separate
images or layers. Then recombining them using "Hard Light" or "Soft Light"
on the red channel. Stir and add salt and pepper to taste. The result, with
NPZ 800 is an image that looks a lot like Kodak Plus-X. It is easy to overdo
the mix and the shadows do go a little, uh, bold on you, but it is
remarkable what this little trick can yield. I've been told that PWP tends
to do this method of conversion a little better than other programs and in
my testing I do agree with that assessment.
Well, that's it for this installment.
AG
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|