35mm results for coolscan9000 and opticfilm120
OK folks, first of all, my apologies for the terrible delay.
I was surprised by some extra urgent work and the whole saga of upgrading my systems struck again after my last message.
A firewire card for the Coolscan9000 stopped working altogether at the worse of times (Xmas break!)
Have you tried to source a pci-e firewire card of late? Let me tell you: it's NOT an easy exercise!
I've got a spare one now, just in case!
Anyways, here goes.
Two images on 35mm Provia100F, taken around Oct 2013.
Used Silverfast for Opticfilm120 scan with the emulsion facing up ( sharper in my scanner).
Nikonscan for the 9000 (emulsion down) with my custom 35mm holder (ANglass and a metal frame) to keep things flat.
AF in the 9000 taken across various points, with an average final value.
The OF120 has fixed focus, only expedite way to change it is to move the film itself (emulsion down, emulsion up).
I tried to equalize as much as I could the colours, but I'll say this:
the colours in the OF120 examples are EXACTLY the ones of the original slides.
The ones for the Coolscan 9000 are off, mostly because I used DEE to try and recover dark areas: that slightly changes colours.
Not a radical change, but enough to be noticed in an A:B comparo like this.
The Opticfilm120 images are prefixed "inof", the ones for the 9000 are "in9k" with "n"being a serial number.
The crop (full size) names are self-explanatory, given the line above.
Opticfilm120: 5300dpi scans, downressed to same size as the Coolscan 4000dpi scans.
In a nutshell: 7500 pixels across largest dimension downressed to approx. 5660.
Sharpening was with Focus Magic, latest version.
The OF120 images can be sharpened a LOT more than the Coolscan ones.
I guess that is an artifact of the higher dpi not really resolving much more.
Of course: when not pixel-peeping the amount of sharpening and its artifacts are not as relevant.
I then downressed both whole images using Irfanview Lanczos filtering and those the "whole" images you see here.
First,a bush, whole images:
http://members.iinet.net.au/~nsouto/photos/dpug/i1of.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i19k.jpg
Now crops of the above, actual size and as similar as I could make them:
http://members.iinet.net.au/~nsouto/photos/dpug/i1ofcrop1.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i1ofcrop2.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i19kcrop1.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i19kcrop2.jpg
Note on the first crop how for the OF120 the yellow leaf on the bottom-right actually has more surface detail (the ribs) than the 9000, even though the little flowers appear sharper on the 9000. I don't know why. The same happens on the second crop on lighter densities.
The OF120 does a worse job of detail in the dark because it does not have DEE applied.
Likely something that could be corrected with selective curves - but I'd like to have a life!
Second a local chieftain in East Timor, whole images:
http://members.iinet.net.au/~nsouto/photos/dpug/i2of.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i29k.jpg
Now crops of the above, actual size and as similar as I could make them:
http://members.iinet.net.au/~nsouto/photos/dpug/i2ofcrop1.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i2ofcrop2.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i29kcrop1.jpg
http://members.iinet.net.au/~nsouto/photos/dpug/i29kcrop2.jpg
Note how the 9000 DEE can make a much better job of the dark set eyes than the OF silverfast can, although it changes the overall colour cast.
Likely things could be improved if I knew how to better use Silverfast, or with extensive masking with an image editor.
But the results are clearly much easier to get with the DEE stuff. In fact, I've found over the years that I use DEE on just about every single portrait I scan on the 9000: it makes for a much better image, although care must be taken with its tendency to change colours at apparently random.
Have a long look, it's all there. Any suggestions for further tests are most welcome.
I also have the latest 8 images in my gallery at dA, taken with Fuji Xtra400 (colour negative) and scanned with the Opticfilm120 using vuescan:
Dead Link Removed
The photos were taken during the 2012 Xmas period and only scanned now
(that's how long my scanners have been out of action due to pc issues...)
My conclusions at this stage?
For portraits with slide film - particularly very difficult ones with lots of contrast - I still like the 9000 and its DEE better.
For everything else, I can't honestly say one is better than the other.
The OF120 with vuescan does a brilliant job with Fuji Xtra400 film: the colours are natural and don't look cheap like they do with the 9000. I guess Ed Hamrick's profile for that film type is a better one than the default 9000 profile for colour negative.
For slide film, I prefer the 9000 although the OF120 is very close with Silverfast. Vuescan is not as good for slide film, I'm afraid.
I still have to pass some feedback on this to Ed Hamrick, I'm sure he can fix that at some stage. In a nutshell, the infra-red image cleaning for slides in vuescan makes everything fuzzy. Not for colour negatives, though. I suspect a problem of some sort in the calculations and process.
The slide film colours with the OF are a lot closer to the originals, while DEE on the 9000 definitely changes them.
DEE is great for slide film portraits, but for lansdcape or nature it can be quite annoying.
The time saved with the fixed focus of the Opticfilm is a factor in its favour. The 9000 is terribly fidly to get the exact averaged ideal focus point. And of course, that fiddling destroys the batch scanning workflow. The 9000 is slightly faster overall in the scans, but then again the OF was doing 5300dpi while the 9000 is doing 4000!
With the Opticfilm, I just batch prescan and colour correct, then scan the lot in one go.
Particularly with vuescan, this is very efficient and fast.
With the 9000, I have the additional step of figuring out the exact ideal focus for each frame.
While the 9000 may be marginally faster in the actual scan, its fiddly focus detracts from the overall time. And this is with AN glass holding things as flat as possible. Without it, it's a nightmare!
Anyways, here it is.