Projection scanning -- anyone tried it?

[bernard_L]

The software tells me I'm getting 6400 ppi, which comes to about but I've seen many conflicting reports on what the actual optical resolution of the V850 might be.

You would be better off basing your decision on actual data. 6400ppi is just vaporware. Given enough memory, I could even resample these images to, say, 25600 ppi.

Here is actual data from my V700, which is not radically different from the V850 (cold cathode source versus LED). These are MTF charts. Obtained by running the V700 (via vuescan) at 1600 ppi nominal, with near-optimum height adjustment. Using the slanted edge method, but that does not matter.

So, at 1600 ppi, the contrast (modulation factor) is down to 30% at 0.2 cycle/pixel. Corresponding to 0.2x2x1600=640ppi. Someone disagrees about the derivation of 640 ppi, feel free; I have seen this kind of forum discussion before and no thank you.

With properly taylored usm restoration, the mtf can be boosted where the response is non-zero, giving this:

So, being a little optimistic, one might say that the response goes out to 0.4 cycle/pixel, resulting in 0.4x2x1600=1280ppi. My guesstimate of 1600ppi in post 13 above was actually generous to the V700/850. So from a 11x17mm negative, you can get 0.47Mpx.

 

[Donald Qualls]

You would be better off basing your decision on actual data.

Manufacturers don't like to give out that "actual data" and most of us lack the knowledge and tools to generate it.

What I'm getting here, though, is that I'm better off with a fifteen year old 6 MP DSLR up to about 6x4.5 negative size, in terms of actual resolution, based on measurements I don't really understand (I'm not a mathematician, optician, or statistician). Epson's got quite the scam running here (none the less, it's a lot more convenient to scan on the flatbed than with a copy stand, backlight source, and film holder/advance unit, even if I had any of those items).

 

[koraks]

Your Epson will max out at around 2100-2300dpi in the real world (so a little shy of 1 rather fuzzy megapixel from an 8x11neg). You may have to scan at 4800dpi and downsample to 50% in order to get the best resolving power, although in my experience with an older model (they didn't change much, fundamentally), scanning at 2400dpi will get pretty much the best you can get with the scanner anyway.
A flatbed scanner is a poor match for tiny negatives due to the inherent fuzziness and limited resolution of a flatbed scan. These machines are great for scanning sheet film and medium format, not so much smaller formats.

We have a Micro Nikkor for the Nikons

That's the right tool for the job; go and make a setup with a suitable light source using the Micro Nikkor and enjoy far better quality images than you'll ever be able to get from the flatbed or a projection 'scanning' setup.

 

[250swb]

But he said "I have access to are Nikon D70 and D90, neither one of which will beat the pixel count I can get from my V850 scanning the negatives directly.". But I do agree that either 6 or 12MP is much more than you want out of such a tiny negative unless all you want is grain. Yes he already got more than 12MP out of his negative scanning normally with his Epson V850.

The pixel count the OP 'thinks' he's is getting from his D850 is a marketing illusion all manufacturers take part in. Scanning at 6400dpi doesn't make a more detailed file, it is simply the biggest interpolated file the V850 can create. The maximum optical resolution of the V850 is 2600 at best, 2300 if not using Silverfast software, and some say it's more like 1600dpi.

Here is an explanation for you, scroll down to 'Image Quality of the Epson V850'

https://www.filmscanner.info/en/EpsonPerfectionV850Pro.html

in which the 6400dpi myth is exploded.

As for 6mp or 12mp being 'more than you want' for copying 35m that is far from the truth, I use a 45mp camera and yes I see grain, individual grain if I zoom in, which is good because it is grain that creates the image, otherwise all you have is mush. But the OP will get a better sharper image with the D90 (and macro lens) than he will with the V850.

 

[Chan Tran]

The pixel count the OP 'thinks' he's is getting from his D850 is a marketing illusion all manufacturers take part in. Scanning at 6400dpi doesn't make a more detailed file, it is simply the biggest interpolated file the V850 can create. The maximum optical resolution of the V850 is 2600 at best, 2300 if not using Silverfast software, and some say it's more like 1600dpi.

Here is an explanation for you, scroll down to 'Image Quality of the Epson V850'

https://www.filmscanner.info/en/EpsonPerfectionV850Pro.html

in which the 6400dpi myth is exploded.

As for 6mp or 12mp being 'more than you want' for copying 35m that is far from the truth, I use a 45mp camera and yes I see grain, individual grain if I zoom in, which is good because it is grain that creates the image, otherwise all you have is mush. But the OP will get a better sharper image with the D90 (and macro lens) than he will with the V850.

My flatbed is rated only at 600dpi and doesn't scan film so I never scan film with it. To scan film I use the Minolta Dual Scan IV which is rated at 3200dpi and it gives me about 13MP file from a 35mm film. I do find that's about all the detail a piece of 35mm film has even with Ektar or Kodachrome.

 

[koraks]

My flatbed is rated only at 600dpi and doesn't scan film so I never scan film with it. To scan film I use the Minolta Dual Scan IV which is rated at 3200dpi and it gives me about 13MP file from a 35mm film. I do find that's about all the detail a piece of 35mm film has even with Ektar or Kodachrome.

I currently have access to a Flextight Precision II which scans 35mm up to 6300dpi. I also have the same Scan Dual IV that you have. Comparing scans of both, I would agree with your statement, but with the caveat that if you go up a few notches (or even very many notches; see e.g. @dokko's work), a very good negative/slide may hold some more resolution than this. However, that's really under optimal conditions and with a marginal net gain in actual resolving power. From a more practical perspective for a home/small amateur lab perspective, I think 3200dpi is a reasonable upper limit of what you can get from a film negative. Anything beyond this tends to constitute only a marginal net benefit.

 

[Donald Qualls]

Okay, so for light source I have an old smart phone and (recently acquired) a Surface Pro 5 tablet computer with 12.3" 2k screen; I'm sure both of those will have visible screen pixels by the time I fill a frame in the D90 with a Minolta 16 12x18 frame, never mind 8x11. I've read about someone getting a "tile light" from home improvement store -- a flat panel LED intended to replace older technology ceiling lights -- but I don't recall reading how well it worked. Then for a copy stand, I've got nothing, really. Both of my sturdy tripods are huge, for large format, and the small one I have isn't rigid enough for this kind of use (and won't permit aiming the camera down between its legs anyway).

 

[koraks]

I'm sure both of those will have visible screen pixels by the time I fill a frame in the D90 with a Minolta 16 12x18 frame

Try a piece of milky white plexiglass. Or even a few layers of OHP inkjet film if you have any. The latter can put at some distance (or two layers with a spacer in-between) to increase the diffusion effect.

the small one I have isn't rigid enough for this kind of use

Even a wobbly tripod with a cable release should do OK. Small problems with parallel placement can be resolved by the combination of depth of field (accepting a little diffraction on the digital capture side) and using digital tools to straighten things back up again.

 

[Chan Tran]

Okay, so for light source I have an old smart phone and (recently acquired) a Surface Pro 5 tablet computer with 12.3" 2k screen; I'm sure both of those will have visible screen pixels by the time I fill a frame in the D90 with a Minolta 16 12x18 frame, never mind 8x11. I've read about someone getting a "tile light" from home improvement store -- a flat panel LED intended to replace older technology ceiling lights -- but I don't recall reading how well it worked. Then for a copy stand, I've got nothing, really. Both of my sturdy tripods are huge, for large format, and the small one I have isn't rigid enough for this kind of use (and won't permit aiming the camera down between its legs anyway).

Using a tablet as light source for copying film doesn't work. The pixels of the screen always shows as r,g,b dots.

 

[Donald Qualls]

Using a tablet as light source for copying film doesn't work.

That was my supposition. A diffuser might resolve that issue, as suggested by @koraks -- or a setup that puts the tablet screen far out of focus. And a cable release will solve a fair amount of wobbly tripod problems. Of course, then I have to find the batteries and charger for the D90, and locate the Memory Sticks (because Nikon switchedf from Compact Flash that I could read directly with my computers to Memory Stick, for which I'll need a new card reader, between the D70 and D90).

 

[koraks]

Yes, there's odds & ends to sort out, but where's a will, there's a way!

 

[Don_ih]

If you have an enlarger, you can turn it on its side, take the lens off, retract the bellows completely (or whatever there is), set your digital camera on a regular tripod, rig up some way to hold the film in the enlarger, use the diffuser of the enlarger as your good backlight source. No need to the camera to point down. You can do it all standing up.

 

[reddesert]

Flatbed scanners have always had exaggerated pixel count compared to dedicated film scanners - or in this case, compared to film, and to DSLRs that are the near equivalent of 35mm film.

Rather than worry about the actual pixel counts, one might ask, can a D90 capture an image that has as much detail as the negative of a Minolta 16 ? IMO the answer is yes. So the problem is getting the image onto the sensor with reasonable fidelity, which should involve a macro lens, some extension tubes or bellows (since you're working at a bit more than 1:1), a diffuser (old slide copier attachment maybe), and fixturing the whole thing for parallelism - a bubble level will help you here.

The D90 doesn't use a Memory Stick, that's an old Sony format. The D90 uses an SD card format that is near-ubiquitous, SD cards and readers can be had down to dirt cheap prices.

 

[250swb]

Okay, so for light source I have an old smart phone and (recently acquired) a Surface Pro 5 tablet computer with 12.3" 2k screen; I'm sure both of those will have visible screen pixels by the time I fill a frame in the D90 with a Minolta 16 12x18 frame, never mind 8x11. I've read about someone getting a "tile light" from home improvement store -- a flat panel LED intended to replace older technology ceiling lights -- but I don't recall reading how well it worked. Then for a copy stand, I've got nothing, really. Both of my sturdy tripods are huge, for large format, and the small one I have isn't rigid enough for this kind of use (and won't permit aiming the camera down between its legs anyway).

Knowing what you are trying to replicate in function is half the battle, Kaiser make a light pad (Slimlite Plano) specifically for 'digitising' negatives, here https://www.kaiser-fototechnik.de/en/produkte/artikel.php?nr=2454 and you can place negatives directly on it without any worry about pixels or an LED pattern showing up. There are other companies who make cheaper versions and I don't know if they work as well as the Kaiser model, but you'd have a returns window if they didn't.