PDA

View Full Version : Scanning via DSLR--any updates?



Darin Boville
30-Nov-2011, 14:51
There's been a few threads on this over the years but I haven't seen anything in a while, so...

What is the current thinking on "scanning" 4x5 or larger negs with a DSLR? We're getting in very high megapixel DSLRs and more about to be announced. Last I heard we were talking about some sort of jig or animator's stand which would move the 4x5 over a light source, DSLR stationary, and then picts stitched together. Lots of theoretical discussion.

So I'm wondering--is anyone doing this? Is it working out?

--Darin

Ben Syverson
30-Nov-2011, 16:11
I have all the parts, but haven't had the time to put them all together yet...

From my initial tests, the biggest problem is dust. There's no way to get an infrared pass. The detail you can resolve is better than an Epson flatbed, but unless you're trying to take 4x5 to 40x50, the difference is not earth shattering. So the Epson wins out in terms of practicality, simply because of ICE.

The most promising application of DSLR scanning is 35mm and 120, since it's a smaller area to dust bust, and less stitching is required. I don't have a 120 scanner, and the Epson throws away half the information in a Mamiya 7 negative, so that's my main interest in DSLR scanning...

sully75
30-Nov-2011, 18:03
Something to do 120 and 35mm would be pretty awesome.

Tom Monego
30-Nov-2011, 19:05
As was said before the problem is dust, you spend longer dedusting than a scan takes. In the end with 35mm the results with an Epson at 3200ppi and a 10 mp D200 on a Bessler slide duplicator were very similar. I was shooting jpegs for speed (lack of in reality). Shooting 4x5 with a 20+ mp camera may be a different story. You need a good macro lens and a copy stand to do it right. Lens and alignment are key features.

Tom

Oren Grad
30-Nov-2011, 19:13
Getting the focus right is not trivial. Even the best magnified live view implementation on a DSLR these days doesn't approach what you can see with a good grain focuser in the darkroom. It doesn't take much of an error to throw away a whole lot of information from the original.

Henry Ambrose
30-Nov-2011, 19:30
I did this regularly and often many years ago with scanning cameras and flat art to be reproduced at much larger sizes. Sometimes it took half a day to finish one. Its doable but there's really no reason for it that I can see, not when you can put the 4x5 negative on a flat bed and get it in one piece. And the cost of a high end DSLR is not trivial.

Even the very best 35mm format macro lenses are not perfect in regards distortion. This makes stitching difficult and there is going to be some information "made up" as the captures are put together. This may or may not matter to you. Some scenes may not lend themselves to such treatment, so you have a process that is not 100% dependable, or so labor intensive that you never want to go there again.

The set up can be as simple as a copy stand and a light box. Sensor plane alignment and axis coincidence can be worked out to perfection by trial, and/or close enough by careful measurement in set up. Stopping down a bit helps with depth of field, but if too much, diffraction rears its ugly head.

By the time you go through all this monkey motion, you could have shot the scene with the high res digicam and saved a lot of trouble and money.

Ben Syverson
30-Nov-2011, 20:14
Focus is not the problem... If you have a rigid setup, you can use the live view to focus and then lock everything down.

f/8 is the optimal stop for DSLR copywork. It gives you a good zone of DOF, it's typically the sharpest or second sharpest stop on a macro lens, and vignetting has been reduced to near-negligible levels.

You still need to do lens corrections in post, but mostly de-vignetting. Distortion on a good macro at 1:1 can be astonishingly low.

The downsides have been discussed, but here's a huge advantage over scanning: the light source is so diffuse that you get a crisp, smooth "wet print" look. Scanners overemphasize grain, bubbles (pepper grain) and dust, because of the tiny light source.

Kimberly Anderson
30-Nov-2011, 21:28
Here's what I've been using lately...

http://tawayama.com/BSlegacyimages/D7000slidecopyingrig.jpg

Robbie Shymanski
30-Nov-2011, 23:36
In the past year I have hardly used my scanner and have only used a teathered DSLR on a modified enlarger for a copy stand. I do a archival work and there's a lot of bouncing between formats, everything from 8mm film up large prints. A scanner isn't always friendly for items that can't be pressed flat against a glass bed. For transparencies and negatives, I use a light table. Those are easy. Alignment via a good level has never been a problem, especially after shooting LF! Personally, I am a Luddite and my workflow is much faster via my copy stand setup than with a scanner as it is easier to move things around. Added, I am pretty good with a camera. A scanner is another piece of redundant equipment that I can survive without.

engl
3-Dec-2011, 09:11
I think there is going to be more interest in this type of scanning as drum scanners and high end flatbeds break down or wear out.

My Epson is nice for 4x5 but isn't giving me everything from my Mamiya 7 negs. I want to try DSLR scanning but my current camera and lenses are not really up for it. I think the best way to do it is as you describe with the camera and light fixed, moving the negative around. Additionally, I think contrast-detect autofocus could be used for better results, it is accurate even at 1:1 magnification (and beyond). The negative may not be flat and the contraption moving the negative around may not move the film perfectly in a plane. This could be dealt with by using smaller aperture, but that would drop some detail to diffraction. If the goal is very high detail scanning, I think a contrast-detect autofocused F2-F2.8 macro lens at F4 on a 18-24 megapixel body should make for some extremely detailed "scans".

As for dust, would some form of fluid mounting be possible? I'm not really familiar with how it works.