Scanning with a digital camera

[Archlich]

I am really enjoying learning more about people using a DSLR to make copies of film negatives and slides. I am assembling the equipment i need to do the same and hope to have good results in time.

However, is there any chance we could stop saying 'scan' with reference to a DSLR copy of a negative or slide? It's not a scan. It's a photograph of something, in this case another photograph.

It would be a scan if you were using a scanner, which you're not. Scanners make scans. Cameras make photographs, images, visual recordings, but not scans.

It's just inaccurate and somewhat grating. Like when a younger person refers to vinyl records as 'vinyls'. The plural of vinyl is vinyl when referring to records.

Sorry guys. I try not to say anything but it sets my teeth on edge.

It would be interesting to know what should the end result be called if a...scanner camera was used?

 

[pyeh]

In fact if you used an electronic shutter, would that be a scan?

 

[Huss]

Leica M7, 7A 35mm f2, Kodak ProImage 100, scanned with Nikon Z7. negativelabpro.com conversion.

 

[ColSebastianMoran]

For my present project, camera-scan of slides, choosing a light source. I have these options, all MR16:
- Solux 4700°K halogen (incandescent) bulb.
- Bulbrite 3000°K halogen bulb
- LED MR16 bulb, claiming 90+ CRI.

Anyone else have comparisons? I'm not seeing much of a difference.

 

[Huss]

Nikonos V, Nikkor 35 3.5, Cinestill 800 rated at 500, LR BW conversion, Nikon Z7 scan.

 

[ColSebastianMoran]

Nikonos V, Nikkor 35 3.5, Cinestill 800 rated at 500, LR BW conversion, Nikon Z7 scan.

Huss, that's a great shot! Thanks!

 

[Huss]

Thank you Sir!

 

[Huss]

Nikonos V, 35 2.5 Nikkor w/ waterproof glass removed, Fuji Superia 400, Z7 scan, negativelabpro.com conversion

 

[ColSebastianMoran]

How many pixels do we need? e.g. for 35mm?

How many pixels do we need? e.g. for 35mm?

How many pixels do we need to "get everything" that could be in the image for 35mm slides and negatives?

I have been assuming that 35mm film would have maximum of 6-10 MPx of image information. So, camera-shot (scan?) of 35mm slide or negative with 24MPx camera and a good lens would be plenty.

Then there's the idea that a full capture would show grain in B&W or the dye clouds in color materials. My latest with a Nikon D7200 and 105mm Micro-Nikkor doesn't show dye clouds. Further, a friend notes, "Your Bayer array camera may give you 24MPx file, but there's sampling, so you won't have that much image information."

What do you think? For really good slides/negs should I use multiple shots at higher magnification, then stitch? Let the dye clouds show, then blur them out in skies? I have lenses to capture 35mm at 2x, so I could...

My question, stated in generic terms: How many pixels per square inch of slide/neg is enough in a camera-scan?

 

[Huss]

How many pixels do we need to "get everything" that could be in the image for 35mm slides and negatives?

I have been assuming that 35mm film would have maximum of 6-10 MPx of image information. So, camera-shot (scan?) of 35mm slide or negative with 24MPx camera and a good lens would be plenty.

Then there's the idea that a full capture would show grain in B&W or the dye clouds in color materials. My latest with a Nikon D7200 and 105mm Micro-Nikkor doesn't show dye clouds. Further, a friend notes, "Your Bayer array camera may give you 24MPx file, but there's sampling, so you won't have that much image information."

What do you think? For really good slides/negs should I use multiple shots at higher magnification, then stitch? Let the dye clouds show, then blur them out in skies? I have lenses to capture 35mm at 2x, so I could...

My question, stated in generic terms: How many pixels per square inch of slide/neg is enough in a camera-scan?

I have this printed and sold this at 20 x 30 and it looks fantastic, even with your nose to the paper. You see grain but no degradation in the image quality.
So it can go bigger. With my D750 (24mp) I was happy with 20inch long prints. Most probably could have gone bigger, just didn't need to.

Nikon F, 45mm 2.8 GN lens, Fuji C200, Nikon Z7 scan, negativelabpro.com inversion.

 

[ColSebastianMoran]

Thanks, Huss. I'll record one vote for 24MPx as enough for a good, big print.

But, still doesn't quite answer my question: How many pixels to reach the point of diminishing (or zero) return in camera-scan of a good 35mm slide/neg?

 

[Huss]

Nikonos V, Nikkor 35 2.5, Fuji Superia 400, Z7 scan, negativelabpro.com plugin

 

[ColSebastianMoran]

How many pixels to reach the point of diminishing (or zero) return in camera-scan of a good 35mm slide/neg?

Did a little test on a pretty good 1977 Kodachrome 64 negative (35mm). One shot capture with an APS camera and good macro lens. Compared to a stitch of four shots creating 50MPx file. No additional detail in the stitched file.

I conclude that beyond 24MPx has greatly diminished returns for MY slides (my film, gear, and hand-held).

Caveats: Original shot on K64 with Nikon prime in good light, maybe 1/200 at f/11. With better gear, film, and technique, results might be different.

 

[jzagaja]

Problem with camera scan is high frequency information. Even if you use good macro lens like APO Rodagon 75mm DX2 (below) and get higher mag than drum scan, there is loss of fidelity. Hardly over 100lpm. Camera is limiting factor too - here Fuji XT2. Worn drum scanner is comparable then.

 

[ColSebastianMoran]

Hello Jack -- Good to hear from you again.

Yes, high frequency detail will be a problem. My gear for one-shot camera scans of 35mm measures at 64 cycles/mm (USAF 1951 glass target). At 2x for stitching, I get 128 cycles/mm, and it doesn't add additional detail for MY slides. Better film, gear, and technique in the original capture might change that.

Here's the comparison of one shot (left, 20MPx up-res'd to 50MPx) vs. stitch of 4 (right, 50MPx).

More: The single shot is as good as a 4000ppi scan (CoolScan V with VueScan). Again, for this slide and my many slides of about the same quality.

 

[ColSebastianMoran]

That's a small area of this overall scene:

 

[Huss]

I dunno, I seem to be getting sharper scans with my d850 set up and Nikon 60mm macro lens:

Full image:


1:1 enlargement from the middle left corner. You can see strands of hair on the lady's face in the foreground. And see detail everywhere:



And this is with Fuji C200, not some high end film. Hand held at low light as the sun had set.

 

[jzagaja]

XT2 has over 6000dpi (theoretical) sensor so more than enough but image looks too smooth. D800 is overklill. Anyone has Sigma SD? Do you make HDR with slides?

 

[Huss]

XT2 has over 6000dpi (theoretical) sensor so more than enough but image looks too smooth. D800 is overklill. Anyone has Sigma SD? Do you make HDR with slides?

If the image from the xt2 is ‘too smooth’ then the d800 is not overkill. the d850 is even better. I am printing 20x30 prints from 6x9 negatives with no
sign of grain and extreme sharpness. I print 20x30 prints from 35mm negatives where you see the grain but there is no loss in resolution/sharpness. And this is not from the ‘appropriate viewing distance’ cop out, but with your nose to the paper.

 

[Ronald M]

Leica M7, 7A 35mm f2, Kodak ProImage 100, scanned with Nikon Z7. negativelabpro.com conversion.

Just started looking at this thread , page 9 & 10 only. The scans from Huss with Z7 are the best so far.

I would just start with a copy stand, strobe, quality macro lens. That is really all the solutions are except for software.

If you are trying to archive images, freezer is best option and seal the film well. All digital medium deteriorates so it means recopying at some interval.