Home-made negative scanner

[vicmortelmans]

For a bit of background, read the second part of this mail. I want to quickly propose a scanning technique I've been thinking on, to see if there's any valuable feedback from your side before I start implementing it.

Would it be possible to use a digital camera for scanning negatives? The camera is not able to focus close enough to take a (back-lit) negative slide filling the frame.

My idea is to project the negative on a sheet of calque paper (this should be easy using an enlarger setup). The digtial camera would be on the other side of the calque paper, at proper focusing distance, for taking a digital picture of the enlarged, projected image.

Any idea what the calque paper would do the image? Will it cause extra noise or other artefacts? Will the image brightness ratio's be identical to those of the negative? Will the brightness be high enough to operate at reasonable exposure times? What other material could be used? A matte screen (should be large!)?

I'll let you know if I succeed (or come back with another idea if I fail...)

Groeten,

Vic

------------

Background of my experiments:

For a while, I've been busy in my spare time (which is little) to create an home-made negative scanner. Commercial film scanners (or flatbed scanners with film scanning features) are (1) too expensive and (2) too slow. I want to end up with a system that allows to scan a whole roll of film *fast* and I don't require the utmost quality or resolution. I'll be very glad if I can use the scans for web-distribution and ordering 10x15cm digital prints.

The basic idea is that I already own a digital camera (Canon Powershot G2), which---in theory---produces images of sufficient resolution and quality (using 16-bit raw) for my requirements.

My first attempt was to add a close-up lens to the camera and shoot a back-lit negative. I found out that a lens of 10 diopter is needed to get close enough focus, but this lens (or the camera lens itself) causes massive chromatic aberration. Just stretching the histogram and converting to grayscale won't produce any quality at all. In theory (again!) there should be a way to correct this by digital postprocessing, but I fear it will become very complicated.

I've also investigated in how to process the scanned image, taking into account the relationship between negative density ratio's and actual scene brightness ratio's and I have measured and analysed the gamma-behaviour of this system.

 

[matti]

I have read about people using ordinary slide repro stuff with bellows and dSLR-cameras for this. It worked fine with b&w and slides. But was hopeless to get colour correction to work with negative colour film.

/matti

 

[Rob]

I was using a Nikon Coolpix camera to copy slides off my color correct lightbox to be able to display them on a TV as a slideshow. It was better than I thought it would be, colors were correct too. Edges were a little soft but it was doable for what I needed it for. I dont think you will have much luck building your own film scanner. Oh I did not have to use any closeup attachments, the nikon would capture the whole slide.

 

[sfb_dot_com]

You are of course completely barking mad - in a nice eccentric inventor kind of way 🙂

There might be an alternative: On eb*y at the moment (UK version) there is a Sony scanner from a print shop on offer. It's scan speed for a single frame is 9secs, and it can swallow a whole roll of film at one go. It also has hardware JPEG compression which means there is no demand on your computer's processor and is capable of scanning at 2,200 dpi giving a 6Mp equivalent file - more than enough for your needs.

The other flaw (well one of several actually) is going to be the transfer and download speed from your digicam, notoriously slow and battery hungry if you do it directly, and just another annoying step if you do it via a cardreader.

I've noticed on the online auctions, there are several bits of minilab up for grabs, and it seems to me that rather than reinventing the wheel, your best bet might be grab some of this purpose-built technology that is going cheap because of the wholesale shift to digital, and adapt it to your needs.

Best of British to you

Andy 🙄

 

[jano]

I have some MF film here and don't have a scanner for it, nor a darkroom to print. In the meantime, I made myself something that sort of works. Not perfect, but okay.

- found an old x-ray illuminating machine in the garage. Wasn't working, so I took it apart and somehow managed to put it back together and it works! Yay! I put this on the floor as a make-shift light table 🙂
- put the negative on the light table
- put a piece of glass on top the negative
- put my 6MP dslr atop a tripod, with a 100mm macro lens, shooting raw
- used f/16, mirror lockup, iso 100, and remote release with all lights off

Results are "okay" for b&w negatives. See attached and see here. A bit soft, probably because of film curling, and I also used heavy jpg compression to try to reduce size of the files. Taken with Iskra 2.

Negatives, on the other hand.... worked okay for fuji 800Z (girl @ camera store), but was horrible with kodak 160NC -- although I sorta like the effect. The tricky part is manually color balancing the whole thing.

For all types, could never get film flat because the white plastic screen of the "light table" is curved and the glass is, too. *shrug*

It's enough, though, to get an eyeball approximation and for small prints (I did 4"x4" with the b&w attached, using careful PS work and they came out decent). I dunno about projecting though... I new someone that projected the slides, snapped the shots with a polaroid, and scanned the polaroid in. Gah!

This may work with 35mm. Haven't tried it.

Jano

 

[jamiewakeham]

I tried shooting macro images of my slides with my 50/4 macro-takumar on my school's EOS350d. Simply couldn't get any decent results at all: whilst colour rendition was fine (brightly lit whiteboard about 2m behind the slide) I just couldn't get the image sharp. Stopped down for DoF, open wide for shutter speed, high and low iso, whatever. Soft as hell. Annoyed me lots, so I just borrowed a mate's DualScan IV 😉

Jamie

 

[David Goldfarb]

I've done it the way Rob describes above--Nikon Coolpix 990, copy stand and a 5000K lightbox. I don't print digitally in general, but this was a convenient way to get stuff in any format onto the web before I had a decent film scanner. I even used this method to produce a four-color postcard from a 5x7" transparency for a friend's gallery show once.

 

[wdenies]

In principle it shoud work. As a proof a German company is modifying slide projectors, to do bulk scanning of slides with a macro lens and DSLR.

 

[lubitel]

I never tried it but I was thinking if it isnt possible to somehow use a normal flatbed scanner? Leave the scanner top open and have some kind of lamp over the negatives? Has anyone tried something like that?

 

[Pherdinand]

If you project it onto a screen and then photographing the image on the screen, don't forget to turn your flash off🙂

OK, more seriously: It should be doable but you lose alot of illumination AND the texture of the projection screen will show up on the images, probably. But i'm curious about it myself.

By the way i tried a few different approaches, like, photographing the backlit slide and cropping later, photographing through a microscope or a magnifying glass, and latest, photographing paper prints. They all worked but the quality was far below acceptable.
The most serious problems were:
1. Reflections from the surface of the medium (slide-film-photograph)
2. Not enough light, ending up in too long sh speeds and too wide-open lens on the camera
3. Alignment problems - everything should be exactly parallel
4. Flatness problems for the medium, resulting in curved edges, curved objects on the image, strange reflections combined with no 1 and misfocus in some regions.

cheers

 

[matti]

If you use thicker paper you might get some fiber structure to the print, like if you use internegatives of paper with old copying techniques. Remember to 10-fold the exposure time, tough 🙂
/matti

 

[wdenies]

If someone interested I could do some tests with my dusty old Kaiser diaduplicator

 

[Kin Lau]

I never tried it but I was thinking if it isnt possible to somehow use a normal flatbed scanner? Leave the scanner top open and have some kind of lamp over the negatives? Has anyone tried something like that?

Doable, but the light from the scanner itself causes reflections. Some use put their small light table upside down on the scanner, and get usable results.

 

[vicmortelmans]

Thanks for all feedback. I accept the quote of being mad. The fact that so many of you respond to this kind of crazyness, also says something about your mental status...

I'll keep an open eye on ebay for old professional equipment, but continue my effort nevertheless.

With a proper digital SLR it should definitely work, as you can equip it with a decent macro lens or bellows. But I'm stuck with my Powershot G2, which is not suited for this kind of macro at all. And I'm not planning to invest in digital SLR for the time being.

Groeten,

Vic

 

[dmr]

I never tried it but I was thinking if it isnt possible to somehow use a normal flatbed scanner? Leave the scanner top open and have some kind of lamp over the negatives? Has anyone tried something like that?

The attached image was, believe it or not, scanned from a MF negative on a flatbed scanner. I placed a bright sheet of clean white paper behind it (this is a heavily used scanner at work, has marks on the white lid surface) and scanned normally at 1200 dpi.

Then I took the image in Photoshop, inverted, adjusted the levels, and spotted.

Amazingly, this looks much better than the old machine-processed print, which was all yucky and grayish.

This was a total slop-shot time exposure I made when I was very young (ca. 1967 I guess) with one of my dad's old Kodak folders. I placed it on a box my brother was carrying back and totally guessed at a time exposure.

I found the negative about a year ago, it was the only MF negative in a box of 35mm negs and slides, and nobody told me I couldn't scan it on a flatbed. 🙂

YMMV, of course. 🙂

 

[lubitel]

wow, that's interesting, because I also tried scanning MF on a flatbed and the result was terrible. Way too dark, to pull anything out of it. May be I should give it another try?

 

[dmr]

Here's a reduced version of the original scan. I lied! It was 2400 dpi and not 1200 and this took quite a while to read into Photoshop. 🙂

This image looks a darker than the original negative, which was somewhat underexposed but not too badly.

I did try scanning both with the emulsion side down and the emulsion side up. This one (emulsion side away from the glass) showed more detail.

Just for the heck of it, I have tried scanning color negatives on this same scanner, and the results were not good at all using the same method.

 

[wdenies]

Duplication works!
This afternoon I did a small test for comparistion between scanning and duplication.
Transformed the colour head of my old Kaiser enlarger into a slide duplicator
Duplicated a slide with the Nikon D70s - Nikkor 60mm macro in Amode at f16
The same slide was scanned with epson 4870 at 3800 dpi
Resampled the duplicate to bring it at the same pixelsize of the scanning

Conclusion:
Duplicate is a little bit unsharper than the scan (due to resampling?)
Speed : duplication is a speedy Gonzales compared to scanning

Wim

 

[wdenies]

BTW Vic your name sounds Belgian or Dutch

 

[vicmortelmans]

Duplication works!
This afternoon I did a small test for comparistion between scanning and duplication.
Transformed the colour head of my old Kaiser enlarger into a slide duplicator
Duplicated a slide with the Nikon D70s - Nikkor 60mm macro in Amode at f16

Wim,

What exactly was your setup? I understand you removed the enlarger lens and took a picture of the slide mounted in the enlarger? Or did you photograph the projection in some way?

Groeten,

Vic (Vlaams, in feite)