Page 1 of 1

Digitising Film

PostPosted: Fri Dec 28, 2012 3:01 am
by surenj
I found two great tutorials on the topic that some members may be interested in - I may use it on my old film when I get organised one day!

How to use a Macro lens for the job
http://www.petapixel.com/2012/12/24/how-to-scan-your-film-using-a-digital-camera-and-macro-lens/


Why you don't need a dedicated scanner (Comparative review)
http://www.petapixel.com/2012/12/23/why-you-should-digitize-your-film-using-a-camera-instead-of-a-scanner/

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 9:55 am
by aim54x
now this is certainly a tempting situation....wonder how negative film will go? I suspect my 105VR + an extension tube will do the trick nicely. I just need to cut down the minimum focus distance a bit.....maybe I should hunt for a LONG hood for my Tamron 90mm f/2.5

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 3:35 pm
by Steffen
I'm half-way through digitising my slides that way. I'm using a 55mm Micro-Nikkor at about 1:1.5 (FX -> DX), and the ES-1 slide copying adapter, no stitching required. The quality is excellent.

Cheers
Steffen.

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 5:34 pm
by aim54x
Steffen wrote:I'm using a 55mm Micro-Nikkor at about 1:1.5 (FX -> DX), and the ES-1 slide copying adapter, no stitching required. The quality is excellent.


I think Patrick is playing with something similar

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 6:14 pm
by Murray Foote
The maximum resolution you will get from the best slide film (eg Velvia) where the image is taken with optimal technique is around 4,000dpi. Most slides will be significantly less than that. All colour negatives will be significantly less than that. Black and white negative films will vary greatly but the best (eg Panatomic X) might be better than 4,000dpi.

Recent flatbed scanners such as the Epson V700 tend to have a stated resolution of 9600dpi but this is a theoretical resolution. Actual resolution is probably no more than 2,000dpi, maybe 2,400dpi if you're lucky (there's a German site that gives accurate tests but I don't remember what it's called). This is not as bad as it sounds because the resolution of most images on film will be significantly less than 4,000dpi. The V700 can give a digital image that will produce a sharp print at A3+.

I'm sceptical about the claims for low resolution of the scanner. If he just grabbed a scanner to use for the test it's likely he didn't know what he was doing. For example, the V700 has two lenses, one for slides in holders, another for slides on the glass. If he used the wrong lens his results wouldn't be surprising.

Scanning is more difficult with colour negative and there are even traps with black & white negative. Quality of results will depend on choice of software and your skill in using it as well as the physical properties of the scanner. Generating or obtaining (eg from Wolf Faust) scanning profiles for different films may also help. Silverfast is likely to be the best software for colour negs and black & white negs but you need something like Silverfast AI Studio and that's not cheap. I would guess that scanning colour negs in Silverfast is likely to lead to better colours than copying with a digital camera. However, for digital copying, if you still have a film camera and access to the right kind of colour negative film you could shoot a Colorchecker target and generate an import profile in Lightroom, which might well give you accurate colour. Where your film has deteriorated with age, Silverfast and Vuescan have easy interfaces to recover from that which work well (at least in the case of Silverfast); manual corrections in Lightroom or Photoshop might not be as easy or accurate.

His claim that you can't get good quality from scanning with a V700 is probably wrong. And even though scanning is always time consuming, it is likely to be much faster than digital copying because with a good flatbed scanner you can scan around 36 images at once (if they're in strips, less if in slides).

However, what he says about scanning doesn't invalidate his claims for digital copying. Also, the blending method in the first link is not necessary if you have a D800. A 35mm slide is 1x1.5" and a D800 image corresponds to 1x1.5" @ 5,000ppi - more than enough provided you can come close to filling the frame. Conversely, an image from the 12MP D3s corresponds to 1x1.5" @ 2,800ppi which should be sufficient, certainly good enough for an A3+ print. You could combine multiple images if you really felt the need but in most cases it would make no difference. The 16MP D7000 yields 1x1.5" @ 3,300ppi which should also be fine. For copying 6x6, 6x9 or 5x4 there might be more of a case for combining exposures if you want optimal quality. Even with the D800, a single exposure to copy a 5x4 only gives 1200ppi which may be OK but is considerably less than full resolution.

An option if you want to copy film using a digital camera is to pick up an old Bowens Illumitran. Quite a few of these are available on EBay. They were designed for duplicating slides and have an internal flash light source. I have one that accepts 5x4 film and perhaps the time has come to make use of it. Other models may only accept 35mm film. The ES-1 adaptor would be an easier way to go but the Illumitran would give you a standardised light source.

One last thing: I wouldn't believe his implication that shooting RAW is optional. You don't want to clip colours and tonalities.

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 7:34 pm
by surenj
Here are some comments by the original photog about his Epsom Scanner process. FYI and may help make up your minds whether he knows what he's talking about. I certainly have NFI.

http://www.addicted2light.com/ wrote:addicted2light Alex • 4 days ago −
First: sorry to all of you, I was away the last few days!

I will try to respond to everyone here.

1) the height of my film-holder has been carefully calibrated, and I even used a piece of anti newton glass to keep the film flat.

2) the Epson results are not that bad, they look terrible because compared against a better technique. Keep in mind you're looking at really really small portions of the frame!

3) yes, a couple of flashes give you a good alternative setup; in this case the hive-like white material of the transparency adapter of many scanners makes a really good "base", because it will spread the light evenly

4) the process is very fast after you get used to it: I spend maybe 10 seconds making 3 to 6 shots (for medium format films) and 1 minute or so (this will depend by how powerful you computer is) to assemble them

5) yes, an Imacon maybe can give you better results; but you will spend more time scanning, and you will have to have 3.000 euros or more laying around. I don't! And a "GOOD" drum scan, from a reputable service, will set you back 50 to 200 euros FOR A SINGLE FRAME! This without factoring in the shipping costs back and forth and the risk of loosing your precious photos in the process.

6) If you shoot only 35mm films than you have alternatives, like the Coolscans; but if you use medium, or worse: large, format than you're out of alternatives.

I am not a fanboy, and I've never been. So if something doesn't work good enough I like to simply state the fact! Happy holidays to everyone :)


Information about Author.
About the author: Gianluca Bevacqua is a fine art photographer based in southern Italy who runs the website Addicted2light. This article was originally published here.

Re: Digitising Film

PostPosted: Fri Dec 28, 2012 11:54 pm
by Murray Foote
I had read that but on reconsidering I now think I may have been wrong about scanner resolution. He says he was using a 5D Mk2 and that would produce a copy equivalent to 1x1.5" @ 3,800ppi which is near enough to the maximum film resolution. So the resolution should be significantly greater than the V700 scanner. Even so, that resolution applies to fine grained slide film with the camera on a tripod, accurately focused and probably using mirror lockup. I would think that many slides would not come close to 4,000ppi in which case difference may be moot. It's still the case that a V700 can give you a good A3+ print.

Re: Digitising Film

PostPosted: Sun Dec 30, 2012 3:51 am
by surenj
aim54x wrote:wonder how negative film will go?

I am about to try with some good ol APS film! :surrender:

Re: Digitising Film

PostPosted: Sun Dec 30, 2012 4:32 am
by Steffen
I suppose the camera would be on a tripod in any case for reasons of practicality, but that has no impact on sharpness since the camera/lens/tubes/ES-1 are a rigid setup. I'm using a studio flash head to illuminate the diffuser on the ES-1, and set the camera white balance from shooting an empty slide frame in the ES-1. The only quality-diminishing mistake I could make (apart from not framing the slide right and square) is getting the manual focus wrong. The DOF is very shallow at this magnification ratio, thus the AF focus indicator, combined with shooting at f/8, gives me sharp reproductions every time without much effort.


EDIT: Here is a detail from a (slightly under-exposed) Fuji Sensia 100 slide:

Image

The setup was D7000, 55mm f/2.8 Micro Nikkor, 1/125s, f/8, ISO100, ES-1, studio flash head. You can clearly see the film grain in this crop, which is good enough for me and the job of digitising all my slides.

Cheers
Steffen.

Re: Digitising Film

PostPosted: Sun Dec 30, 2012 4:09 pm
by Murray Foote
I calculate that to get full resolution on a perfect high-res slide (@4,000ppi) you would need a 24MP camera, for very good resolution (3,000ppi) a 13MP camera and for good resolution (2,000ppi) a 6MP camera. I would think that a 13MP camera would deliver perfectly good results for most slides and even a 6MP camera is probably good enough for most output requirements.

For larger formats it's not so easy. For full/ very good/ good resolution (as above) you would need 59MP/ 33MP/ 15MP for 645; 134MP/ 75MP/ 33MP for 6x6, 6x7 or 6x9; and 383MP/ 215MP/ 96MP for 5x4 - unless you combine multiple images. Combining exposures be easier said than done if you have lots of sky or water.

Requirements will be less critical for colour negatives (including chromogenic b&w film) and coarse grained slide film because the resolution is less than fine-grained slide film. As Steffen suggests, focusing would be critical and it may help to use live view (if your camera has it).

The other loose end is dynamic range but you should be able to detect any clipping in Lightroom or RAWDigger (there will be more latitude than the camera histogram or Lightroom histogram shows). You're going to be copying at base ISO where dynamic range is greatest but it will be less for older cameras and smaller sensors. I suspect it's not an issue but if it is it would be for slides and perhaps b&w negs but not colour negs. (Colour negs capture a wider exposure range than slides but store it on film as a narrower band of colours and tones). Noise shouldn't be a problem on almost any camera.