Actually, it's not necessarily a different thread, because "scanning" with a camera is a definite option, as a colleague and I demonstrated in some detail here, in particular the PDF download. I'd be interested to hear (and perhaps so would others), either in this thread, or another one, how you are approaching this and what you are finding.
Hi Mark,
I'm focused on digitizing my Kodachromes from the 80s and on, plus some older family chromes from the 50s and 60s. I started off with a Coolscan IV but ran into flare/blooming problems (even with the mirror cleaned) that really limited PP adjustments. I tried a Coolscan V (I had 3 aborted purchases of Vs and 5000s before I found one that did not have error indications on boot) but the V had even worse flare. (Digging around on the net revealed that the sensor covers in at least the IV and V series is plastic; replacing them with glass fixes the problem - not a cost effective solution. I don't know what the story is with the 9000s.) Anyway, I started off with the Nikon ES-1 and 55mm f2.8 Micro-Nikkor, but the edge/corner quality just wasn't there.
Right now I am using the Olympus Auto Bellows and slide holder attachment with the D800E and a monolight as the light source. (The Olympus bellows has the slide holder attached to the tripod mounting block instead of the rail, with the result that you can focus without altering the magnification ratio, nice if you need to focus stack for severely bowed slides; this also lets you use longer focal length lenses (e.g., 75mm and 80mm) than you can with the Nikon bellows which I also have tried. However, for reasons that utterly fail me, the Olympus slideholder itself only allows movement up and down, so I've modded the Oly slide attachment with a Konica holder to allow four-way movement.
I'm in the process of head to head testing of the Olympus 80mm f4 bellows macro and the Rodenstock Apo-Rodagon D 75mm f4 APO, both of which are optimized for 1:1 reproduction. I'm also investigating the workflow for using greater than 1:1 magnification and stitching (for that I'll be comparing the Oly macro and the Rodenstock Apo-Rodagon D 75mm f4.5 which is optimized for 2:1 and should be ideal for close to 1:2 reversed). Stitching is a bit of a pain and I'm not yet convinced that it yields quality gains at print sizes up to 17x25. (That said, when I was scanning with the Coolscans, I would typically make 3 or 4 scans at different focus values and focus stack and hand mask, so I'm used to dealing with multi-image workflows.)
I'm also playing around with slide mounting methods for digital duping. A lot of the slides I have are quite bowed which poses a DOF problem; also some old family slides are warped from excessive project or heat. The sweet spot for the above lenses is f5.6, IQ goes down a little at f8 and really starts falling off at f8 (as I understand it, at 1:1 a nominal f5.6 aperture is an effective f11, so you are already in diffraction territory even before really stopping down). Right now I'm using Wess AHX mounts, which are full frame plastic mounts with sprocket registration tabs which help to tension and flatten the slide, but only to a certain degree. A recent idea is to use Gepe plain glass mounts with the glass only on the non-emulsion side to see if I can flatten the slide some without affecting image clarity; at first blush I am not seeing Newton rings problems (I see them on the mounted slide but not in the "scans").
I am finding that the D800E can easily handle Kodachrome's dynamic range (with the caveat that I want the prints from the files to resemble the Kodachrome images, i.e., I am not trying to turn Kodachrome shadows into D800 shadows).
One final note, I really liked the grain acutance look of the Coolscan scans and experimented with duplicating in the D800 files it by removing the diffusion glass on the slide holder and using a condenser enlarger head as the light source to get more collimated illumination on the slide (think cold light vs. condenser B&W enlargements). Result: don't even think about using a condenser head, dust and scratches go off the charts with no real increase in grain acutance.
The intermediate bottom line is that I am getting "scans" that have the slide detail of the Coolscan V scan and improved dynamic range, but am not yet getting the crisp looking grain I liked in the Coolscan scans. (I have no interest in suppressing the natural grain of the chromes, to me that's an inherent part of film that I want to retain). So I'm looking at tweaks at the margins to see what more I can squeeze out. For example I'm finding that the Oly and Rodenstock 1:1 macros are so close (in tests on slides in glass mounts to get a perfectly flat test subject) that you have to look closely at 1:1 to see differences on screen, such I expect that with an open mount slide any lens difference will be subsumed by DOF and/or alignment issues. I suspect this is a light source issue rather than a lens issue, though.
I'm documenting my investigations, and will dribble them out the web as I go along.
Alan