Digitizing Slides with a DSLR

Tanuki

I'd like to share the results of my initial trial of using a DSLR and slide copy attachment for digitizing 35mm slides.

Why am I using a DSLR to copy slides, you might ask? Why wouldn't I invest a few hundred dollars into a good quality film scanner? The reason is that I want to scan my family's slide collection which resides at my parent's house, half way around the world from me. Compared to a film scanner, a DSLR has two important advantages for me: speed of operation and portability.

According to dpreview.com, the Nikon D80 (my camera) can capture 8 stops of range in jpeg, and 10.4 stops of range with RAW capture if you really push the settings in ACR. I've read that many slide films carry a log density range of 3.2. If you do the math, 10.4 stops equates to a film log density of 3.1 for RAW, which is just short. So maybe it's a bit beyond the range of my D80, but hey, it's worth a try.

I started out with a conservative test by photographing a developed film leader from a roll of Fuji Velvia RVP (ISO 50, I think) so that half the frame was clear unexposed film and the other half of the frame was dark exposed film. I used a tungsten light source that was gelled to be daylight balance and my ISO was set to 100.

You can see the frame below.



I "developed" this raw file in Lightroom 1.0 using Exposure, Recovery, Fill Light and Blacks all set to 0, Brightness set to +50, Contrast at +25 and Point Curve set to linear. To my delight, I the dark side was reading at about 3% and the light side was reading about 95%. This is a much better result than I expected! I didn't even have to use the extreme develop settings in Lightroom that was used by dpreview to get the 10.4 stop range. It makes me wonder if something else is going on here... perhaps I'm getting some lens flare?

I did another experiment to take lens flare out of the picture (no pun intended). I went to a frame that was all dark and did a spot meter measurement, then compared to a spot meter measurement from a completely clear frame. The difference came out to exactly 8 stops, which seems to confirm my previous test.

So now for a test of an actual scene. I took a slide from that same roll, used the same develop settings in Lightroom, and I got this result.




Not too shabby, but a little flat. I did some adjustments to the develop settings, and came up with this result.



Comparing this to the original slide, I have to say that I'm pretty pleased with the digital copy.

I played with some extreme Fill Light settings in Lightroom to see how noisy the shadows were. To my delight, the noise wasn't bad at all. And just for fun, I tried the multiple exposure feature on the D80 (3 exposures with auto gain) to turn it into a multi-pass scanner, and it successfully reduced the shadow noise so that more detail showed through. However I probably won't go to this trouble for my family slide collection, considering that the results from a single exposure are more than good enough.

As far as I'm concerned, DSLR + slide copier + Lightroom = fast and decent results.

Any comments, questions or critique? I'm sure this is likely to turn into a debate about whether a DSLR is equivalent to a good film scanner, and might as well state now that I don't think it comes close on resolution and on features like Digital Ice. On dynamic range it probably falls only a litle bit short. But on convenience and maybe speed, it certainly has the advantage. For the family slide collection, it's more than good enough in my opinion.

If anyone is intersted, my copy lens is a Nikkor 50mm f1.8 with a Kenko 25mm extension tube. (At minimum focus distance, the 35mm frame fills the sensor frame almost exactly.)

pathfinder

Very interesting discussion.

I have read about using a Nikon CoolPix 995 3 MgPxl camera to copy slides several years ago.

Fixed slide holder/copiers can be used rather effectively and inexpensively to copy slides with dgital cameras. They do not offer all the contols and Digtial ICE that are incorporated in modern film scanners, but still can do a credible job.

Tanuki

If anyone's interested, I just posted another thread on the related subject of Digitizing Color Negatives with a DSLR.

Mike

pat.kane

Thanks for the write-up. At one time, I tried to scan a bunch of the family's slides, but gave up because of the effort and time involved. This is an interesting approach, though one I probably won't go to the trouble to duplicate--if I get the motivation again, I'll probably just pay to get the slides scanned.

JimW

I'd like to share the results of my initial trial of using a DSLR and slide copy attachment for digitizing 35mm slides.

Tanuki, this is a really great idea that I would like to try. Could you please tell me what you mean by a "slide copy attachment"? Thank you for posting this method.

Jim

Tanuki

JimW wrote:

Tanuki, this is a really great idea that I would like to try. Could you please tell me what you mean by a "slide copy attachment"? Thank you for posting this method.

Jim

Here is an example of a DIY version. It doesn't look that difficult to make. If you read the rest of my post, you'll find that a DIY version is probably your best bet, unless you want to just put your camera on a tripod and aim at a light table.

Another option is the Nikon ES-1 which connects to a 52mm filter thread, but only extends 69.5mm as detailed here, which may not be long enough for your lens. My Nikon 50mm f1.8 and a 25mm extension tube, for example, requires 95mm between the lens and the slide and would need some kind of additional extension. But good luck finding such an animal. The closest I could find is the old Nikon K-series extension tubes, which none of the used camera stores carry.

You can also buy attachments that have their own lens (search for "slide copier" on Amazon), but the qualilty is dubious. I bought one of these contraptions and was so disappointed with the focusing mechanism that removed the slide holder from the lens and made my own extension tube from PVC pipe to connect to my 50mm f1.8 and extension tube. I'll post a picture of my contraption if I have some time later tonight.

Sorry I don't have any better suggestions, but I've struggled with this myself for quite some time until I came up with my solution.

Tanuki

Tanuki wrote:

Sorry I don't have any better suggestions, but I've struggled with this myself for quite some time until I came up with my solution.

Okay, I've found the ultimate solution. You've got to see it to believe it. It's a slide projector hack with auto advance and auto dust-off. I love how he rigged the tubing to put the compressed air onto the front and back of the slide just moments before he copies it. I think he's gone overkill with the air compressor and solenoid though. This should work just as well with a can of compressed air or even a Pocket Rocket as the source of air into the tubing.

A second solution, which was what I was investigating when I stumbled upon the slide projector hack above, is to hack a stack slide viewer like this Pana-Vue or this Kaiser by removing the magnifying lens and have a tripod mounted camera looking down at the slide. I'm not sure the quality of the light source is very good in these cheapo viewers, so it might need to be replaced. It might be as simple as removing the back and placing it on top of a light table. And I would think there would be room inside to install tubing for the dust-off hack like in the ultimate solution.

I know some feel that it would be easier to send the slides off to be scanned professionally, but (1) my family's slides are irreplacable and I'm not willing to risk letting then out of my possession, (2) we could be talking about a lot of money to send thousands of slides out for scanning and (3) I'd miss out on the fun of doing this cool hack!

Mike

JimW

Mike - I think your original post is right on. If the dslr method renders very slightly less density and slightly more noise etc, then it seems to me that if necessary, this could be made up with good post processing methods. I’d think that your method and results would stack up very well against a slide scanner, short of a wet scan on a drum.

The Nikon ES-1 with the step down ring to my 60mm lens will work for me. I tried shooting the slides on a light table, but it was an inelegant solution. I have only about 500 slides to be digitized, so maybe I can dig up a used ES-1 here in NYC. Then I can use my flash for a light source and get it all done this summer.

That projector hack is cool though. And it reminded me that I want to shoot these slides with the camera tethered.

Those are a couple of very helpful posts, and popped up just when I was looking for this info. Thanks for posting them. - Jim

Tanuki

JimW wrote:

Mike - I think your original post is right on. If the dslr method renders very slightly less density and slightly more noise etc, then it seems to me that if necessary, this could be made up with good post processing methods. I’d think that your method and results would stack up very well against a slide scanner, short of a wet scan on a drum.

I agree that a modern DSLR at low ISO should stack up well against a single pass film scanner in terms of dynamic range and noise. I'd like to see a head to head comparison between a good DSLR copy technique and a prosumer scanner. The only direct comparisons I can find on the web involve P&S cameras with their tiny sensors and shooting jpg. It's just not the same.

JimW wrote:

The Nikon ES-1 with the step down ring to my 60mm lens will work for me. I tried shooting the slides on a light table, but it was an inelegant solution. I have only about 500 slides to be digitized, so maybe I can dig up a used ES-1 here in NYC. Then I can use my flash for a light source and get it all done this summer.

Are you shooting with a full frame sensor? If not, I would expect that the 60mm micro with ES-1 shooting distance will result in a cropped image of your slide. That's why I'm using the 50mm f1.8 with 25mm extension tube.

JimW wrote:

That projector hack is cool though. And it reminded me that I want to shoot these slides with the camera tethered.

Those are a couple of very helpful posts, and popped up just when I was looking for this info. Thanks for posting them. - Jim

I'm glad this is useful to you. Frankly I'm surprised more people aren't interested, as my Internet research on the subject has uncovered many, many archived questions about this on other forums. Maybe all the people who shot slides in the past have already purchased film scanners, or have given up on digitizing their collection of thousands of slides.

At the very least I expected some technically minded Dgrinners to have some intellectual interest in the subject.

Mike

JimW

Are you shooting with a full frame sensor? If not, I would expect that the 60mm micro with ES-1 shooting distance will result in a cropped image of your slide.

Nope, I'm using a Nikon D200, and I sure don't want to crop any image. However, I also have access to an older, but decent, manual focus 55mm Micro-Nikkor.

Also, I might have access to a friend's Nikon film scanner, don't know the model, but I think it's possibly a 4000? Might qualify as prosumer. If I can get access, I'll try and post a slide done both ways. Might be a little while though.

Tanuki

JimW wrote:

Nope, I'm using a Nikon D200, and I sure don't want to crop any image. However, I also have access to an older, but decent, manual focus 55mm Micro-Nikkor.

I think you won't be able to use the ES-1 with a 55mm Micro either, because the ES-1 doesn't extend far enough for the 1:1.5 magnification that's needed to fill a DX sensor with a 35mm slide. I think you'll need some kind of extension on the ES-1.

JimW wrote:

Also, I might have access to a friend's Nikon film scanner, don't know the model, but I think it's possibly a 4000? Might qualify as prosumer. If I can get access, I'll try and post a slide done both ways. Might be a little while though.

I'd love to see the results when you get around to it.

Mike