Using a Sony A7 as a 35mm film scanner

Musicman

Musicman

Suspended / Banned
Messages
9,866
Name
Rob Telford
Edit My Images
No
Sony A7 'Scan' experiment by -cybertect-

I'd been toying with this idea for a while and this is my first attempt at using my Sony A7 as a film scanner - Sony A7 + Canon FD 50mm f/3.5 Macro lens + Canon FD Auto Bellows that I happen to have.

The effective resolution is about 6000dpi and rather better pixels than I get with my V700. It needed no sharpening beyond the default in Lightroom.

It's also a lot quicker :-)

Original shot with a Canon AV-1 with FD50mm f/1.8 and Kodachrome slide.

100% crop

Sony A7 'Scan' experiment 100% crop by -cybertect-

V700 scan for comparison - the colour with the A7 is more accurate and was a lot easier to get right.

Sun shadows by -cybertect-
 
That's a cracking result, very sharp and very clean indeed. I've often wondered what kind of results I could get 'scanning' film using my 5Dmk2, be interesting to try sometime!
 
Fantastic results. Pretty sick of how slow scanning can be, especially medium format...
 
What was your set up for this, it certainly seems to be easier and quicker than using a scanner. Excellent result.

Andy
 
Andysnap said:
What was your set up for this, it certainly seems to be easier and quicker than using a scanner. Excellent result.

Andy
Click to expand...

Yes, please share..
 
Well, it's as I said in the OP.

Sony A7 camera : FD 50mm f/35 Macro lens : FD Bellows : FD bellows slide adapter accessory : iPad

As I couldn't easily find my usual lightbox last night, I used my iPad as a light source :)

The Sony A7 is a 24 Mpx full frame mirrorless digicam, which allows me to use the 50mm at 1:1 magnification. Using the live EVF/screen on the camera allowed me to zoom in to 100% and focus on the grain, which made things easy.

It should work with any full frame camera with live view, though I suspect you could do it with a 1.6 crop body at a lower magnification set on the lens.

Pretty much the same setup as here with a 40D, in fact.

http://www.flickr.com/photos/hpulley/4408771957/

I had previously experimented with my Panasonic G2, but there wasn't enough travel in the bellows to get the whole of a 35mm frame in shot.

Otherwise it's exactly the same as you would have done for slide copying with film.

Other 35mm bellows with a lens with decent flat field reproduction should work. You could probably use a simple slide copier attachment instead of the bellows and macro lens (Jessops et al used to sell them, usually with a T2 mount) but I used the bellows and macro lens as I already had them.

I've yet to finesse how I set it up, but so far I'm getting about 19 Mpx 'scans' (about 5300 x 3500 pixels) when I've left a little room for cropping out the slide mount and any rotational errors.

Once you have it set up, it's a matter of seconds to pop the slide in and capture the frame. The downside is that you have no ICE, but as I've been mostly trying Kodachrome, that wasn't going to work anyhow. I've still got to work out how well I can deal with negative strips instead of mounted slides.

The A7 has huge dynamic range of over 14 stops, which helps capture the most amount of detail across the tonal range.

I've a couple of other attempts with the same setup I'll post up later.
 
That's impressive from using the iPad as a light source. I've done the same but assume you have a retina model as I've used an iPad 1 and the pixels show clearly on the image :0(

View attachment 2053

That's a 6x6 negative from a Kiev 80 using a Sony NEX with OM 50/1.4

Cheers
Steve
 
Yeah I thought of using the ipad as a light source but the pixels are very visible.
 
The slide mount for the bellows has a ground glass screen behind it, which will mitigate that problem ;)
 
Anyone want a Minolta Scan Dual III? Selling to fund purchase of Canon FD Bellow malarky.... :p;)
 
you could try tracing paper to blur the pixels perhaps?
also if your using a manual lens you can get the super cheap extension tubes off ebay, if u cant find cheap bellows
 
Musicman said:
I've still got to work out how well I can deal with negative strips instead of mounted slides.
Click to expand...

OK. I think I have this sussed now :D

For those that wanted to see, here's the setup (still using my iPad as a light source, hehe)



Neg strips slide in a convenient set of pegs between the slide holder and ground glass on the Slide Copying Rig.

I picked up a copy of ColorNeg for US $67, which handles turning a colour negative into a positive image with a couple of clicks, with presets for a number of common film types.

Here's a V700 scan I did last year, shot with my Canon A-1 on Superia 200.

I have always had terrible trouble getting the colour half way decent when scanning Superia, and this was still some way off my memory of the scene.


Bicycle on Mill Street by cybertect, on Flickr

And here's the A7 'scan', processed through ColorNeg and a little tidying up in Photoshop: levels, some mild contrast adjustment and spotting.

It was a breeze in comparison: faster, easier and the colour is a lot better.




Comparing 100% Crops

V700 scan with Vuescan



Sony A7 + FD 50mm f/3.5 Macro 'scan'



Look carefully, and you can actually read LONDON BOROUGH OF SOUTHWARK on the street sign, which must have been 70 or so metres away. Individual bike spokes and the cyclist's shoelaces are discernible where they were not with the V700.

I am now really curious to see how it stacks up against a decent film scanner like a Nikon Coolscan and how much more detail an A7r could extract :)
 
Last edited:
I was in maplin yesterday and I was looking at those big SAD lamps, I wonder if they'd be better than the ipad.
 
TBH, I'm not sure it's that critical.

I corrected for any colour cast from the iPad by taking a picture of a blank frame, with no slide in place, before I started digitising images.

Eyedropper WB in Lightroom came out at 5450 Kelvin with a Tint of +12. Lightroom's standard Daylight setting is 5500 K/+10, so it is actually very close to LR's daylight.

Applying the 5450/+12 WB to any subsequent images brought them into line with the now established 'white'. Any remaining colour casts in the digitised image should due to the original exposure conditions or the film itself.
 
So, passing the Oxfam shot this morning, and spotted an AICO Video Slide Copier for £2.99. It appears to be native 52mm thread, but comes with a number of adapter rings for different lens threads, including 46mm and 49mm. The latter should work with my Pentax lenses, and perhaps with extension rings I'll get the appropriate image size.

TBH I'm not quite sure why I bought this, as I already have two scanners, and it doesn't come with a light source (and I don't have an iPad). But at £3 it seemed churlish to refuse...
 
Just purchased a Minolta version, can't wait to have a play.
 
Wow great results if you also have a very good digital camera to use.
 
...and with a very good flat field enlarging lens the results might be even better.
 
PMN said:
A crop sensor isn't going to resolve more detail than a full frame so I'm not sure how it could be better, I'd love to get a setup like this and do a direct comparison between my 7D and 5Dmk2!
Click to expand...

Just comparing using a macro lens with an enlarging lens....and would think quite a few members with digital cameras would have an enlarging lens but not a macro.
 
PMN said:
I'd be very interested in seeing the results from a crop sensor camera to see how it compares, presumably the results would be softer but also presumably still pretty good!
Click to expand...
Teflon Mike tried it during the summer, it looked like a hellish faff and more of an a academic exercise than something you'd want to do for a roll of film.
 
I had previously tried a similar setup with my Panasonic G2 and the same FD macro lens.

It didn't work as well as with the A7 as the G2's 4:3 aspect ratio meant I lost a significant chunk of the available 12 Mpx resolution to letterboxing the top and bottom. IIRC, I was getting final file sizes in around 8 Mpx, compared to the 18 or so I am with the A7 and 10-12 Mpx with the V700. The G2 captures were much sharper than the V700, though.

I also couldn't get the bellows back far enough to see the full frame with a 2x sensor crop and 50mm lens, so I had to dispense with the bellows and just rest the slide on a light box and mount the camera on a tripod. This meant I had to reset the focus for every new slide, which was slow and I also had difficulty with glare and reflections of the light source off the front of lens. It also wasn't going to work with negs.

In addition to improved resolution, the A7 has better noise, dynamic range and colour than the G2, which makes the files easier to work with.

If the film is fairly clean (and I don't have terrible problems with dust) I can get away with spotting an A7 RAW file in Lightroom instead of having to take a TIFF into Photoshop.

An APS-C body with live view might do a lot better, though.

A couple done with the G2

Festa sul Ponte by -cybertect-

We wanted to be the sky by -cybertect-
 
FWIW, a couple of evenings ago I whipped through a box of 38 Kodachrome slides before dinner in about 20 minutes, and that was while explaining what I was doing to my six year old son :)
 
Last edited:
steveo_mcg said:
Teflon Mike tried it during the summer, it looked like a hellish faff and more of an a academic exercise than something you'd want to do for a roll of film.
Click to expand...

I remember that, I think he was almost doing a giant panorama of the frame and stitching them together giving a massive file. I didn't see the relevance of doing it like that (beyond it being a purely academic exercise) as it wasn't going to extract any more detail than doing a single standard shot and took probably 1000% longer!

Musicman said:
I had previously tried a similar setup with my Panasonic G2 and the same FD macro lens.
Click to expand...

Very interesting, they still look pretty damned good at web size!
 
PMN said:
I remember that, I think he was almost doing a giant panorama of the frame and stitching them together giving a massive file. I didn't see the relevance of doing it like that (beyond it being a purely academic exercise) as it wasn't going to extract any more detail than doing a single standard shot and took probably 1000% longer!
Click to expand...

Ah I see what you were getting at though, use a APS-C and just take one frame rather than the stitching. Would these slide copiers work for small framed cameras?
 
PMN said:
I wasn't really thinking that, I was just saying I seem to remember that particular experiment had many images stitched together to get the final frame and the benefits of that passed me by somewhat. :)
Click to expand...
Fair enough. :)
 
You must log in or register to reply here.
Back
Top