Scanning with a DSLR

[cmo]

I own a Canon EOS 5D2, a bellows, a slide duplicating attachment and an excellent macro lens, and I use this as a scanner for my b/w films. It works great

When it comes to color negative films things are slightly different because the color masking makes corrections more difficult.

How doo you handle this?

 

[iansand]

perhaps you could use a filter on your lens, not sure what strength all the same

 

[maarten m]

do you use a flash or other constant lightsource?
take one picture of an empty peace of film, set this as your manual WB. should work.

mm

 

[cooltouch]

I logged in here looking for answers to the exact same question. I'm curious if cmox has tried this method of setting the WB using an empty piece of film. Seems on the surface at least, it should work. What I've tried doing is, in my post processing software, selectively remove this orange tint from the image. Or, after reversing the image, removing the cyan from the positive. The results haven't been all that impressive so far. The photos end up with a rather strong posterized look to them.

I'll have to read up on setting manual WB on my camera.

So, has anyone tried this? Results?

 

[pellicle]

When it comes to color negative films things are slightly different because the color masking makes corrections more difficult.

How doo you handle this?

use a scanner and tweak the colour channels. I've never had a problem with the mask, nay I can usually see no trace of it

 

[cooltouch]

Problem with using a scanner though, as opposed to a DSLR as a slide duplicator, is a big loss in image resolution. My Epson 4990 puts out maybe 2000 ppi, whereas my lowly 10.2 mp camera puts out about 2700 ppi. So this is a technique worth knowing better, methinks.

 

[glhs116]

I might be tempted by this technique if I had all the slide copying and macro gear myself. I made a stab at this in a much more redneck fashion using my cheap Zoom-Nikkor and lightbox. I found the following:

You need to shield the lens from all extraneous light (I made a tube and taped around the frame on the lightbox). Your bellows and slide copying attachment would take care of all of these considerations.

The next consideration relates, I think, to the problem you are having. The problem is the "white balance" mindset. Colour negatives have a large imbalance between the channels. Digital colour balancing is just a matter of boosting the weaker green channel and much weaker blue channel. This is far from ideal. If you are ending up with a colour cast you cannot easily get rid of the most likely explanation is that you are clipping one of the colour channels. Red will be very easy to clip on the highlight end (which become your shadows when inverted). Blue will be easy to clip on the shadow end (which become your highlights when inverted) because it is so weak. Especially on a Canon since Canons are tuned to give those deep inky blacks.

My suggestion is to set a "daylight" white balance and obtain an even exposure for all the primary colours by using actual screw-on colour correction filters. If you do it this way you will be able to obtain a full-range exposure for all three colour channels on the sensor and much more easily prevent clipping any colour channels.

The difference with a scanner is that it will automatically set different exposures for the different colour channels when used in the "negative" scan mode. On many scanners you can also take manual control of this such as the Epson, the Konica Minolta software and the Nikon software. On the DSLR all colour channels are exposed together by a single shutter so you either need to balance the channels with coloured filters or take three black and white pictures of the same film through three primary colour filters and combine them in Photoshop.

Sam

 

[pellicle]

Problem with using a scanner though, as opposed to a DSLR as a slide duplicator, is a big loss in image resolution. My Epson 4990 puts out maybe 2000 ppi, whereas my lowly 10.2 mp camera puts out about 2700 ppi. So this is a technique worth knowing better, methinks.

your right, the res of a good dslr will perhaps exceed a 4990 ... but not the control it offers and the ability to put 6 strips of 6 exposures on the glass and walk away while it goes about its business. I've never seen anyone do a good job of capture with colour negative on a DSLR but you may be the first. Primary problem seems to be the lack of signal you get in the blue channel. Perhaps if you could tailor your light source to suite? Dunno

When I want higher res scans I use a Nikon LS-4000

Its not that I haven't fiddled about for quite some time on it ... but in the end I bought a better scanner

 

[glhs116]

I've never seen anyone do a good job of capture with colour negative on a DSLR but you may be the first. Primary problem seems to be the lack of signal you get in the blue channel. Perhaps if you could tailor your light source to suite? Dunno

Pellicle, read my post. I was addressing that very issue. Since you can't vary the exposure per channel you need to either make a filter pack to neutralise the colour imbalance or use filters to take a properly exposed photo of each colour channel.

Sam

 

[pellicle]

Hi

Pellicle, read my post. I was addressing that very issue. Since you can't vary the exposure per channel you need to either make a filter pack to neutralise the colour imbalance or use filters to take a properly exposed photo of each colour channel.

I did ... and I thought I didn't say anything against that. Sorry if I did not mak a point that I agreed with you.

Actually I made the mistake of thinking I was replying to the OP ... as I'm trying really hard to not reply to arguments as it just gets me into strife.

 

[glhs116]

I hope we are all trying to help OP as well as anyone else who finds the thread later. The reason I didn't think you had read my post or had read and didn't agree was that you said the primary problem was the lack of signal in the blue channel and my post directly addressed that issue in terms of using a DSLR.

In fact, here is a not-totally-horrid result I got that way with afor-mentioned lightbox and sticky tape setup:

http://www.flickr.com/photos/samagnew/4713716315/

and another here:

http://www.flickr.com/photos/samagnew/4714351622/

The second one shows some of the light-leak problems I was having. A bellows and slide copy setup would eliminate such issues.

Sam

 

[TareqPhoto]

I am wondering if using DSLRs can give fine results, how to compare that with flatbed scanners such as Epson V750? or Nikon 9000?
What about using a digital medium format as a scanner?

 

[pellicle]

In fact, here is a not-totally-horrid result I got that way with afor-mentioned lightbox and sticky tape setup:

http://www.flickr.com/photos/samagnew/4713716315/

quite good ... is that full frame or in tighter?

 

[glhs116]

quite good ... is that full frame or in tighter?

The camera was my trusty ol' D80. My only DSLR. The lens was the Zoom-Nikkor 35-105 "Macro" that came with my Nikon FE mounted on a reversing ring. This gave me the ability to adjust the position of tripod, light box etc. until I had one 35mm frame filling the viewfinder. I made a black paper mask on the light box with a 35mm frame sized window in it. I found I also had to make a tube of black foam and shoot through this "tunnel" because so much stray light came in from the sides.

This would be much easier with bellows and copier attachments. The key to a good result is getting a good exposure of all the channels by optically filtering the negative rather than relying on software.

What else? I don't have Photoshop so my raw file was imported as a gamma 1.0 tiff by PhotoLine. The negative convertion was with the ColorPerfect plugin using the ColorNeg module.

No ICE but with a better camera and better setup I can see this cranking out some very good results much faster than scanning. Since I don't have the right stuff and I do have my 9000 I haven't experimented any further with it.

 

[pschwart]

what is needed is a side-by-side comparison with an equivalent scan using the 9000.

The camera was my trusty ol' D80. My only DSLR. The lens was the Zoom-Nikkor 35-105 "Macro" that came with my Nikon FE mounted on a reversing ring. This gave me the ability to adjust the position of tripod, light box etc. until I had one 35mm frame filling the viewfinder. I made a black paper mask on the light box with a 35mm frame sized window in it. I found I also had to make a tube of black foam and shoot through this "tunnel" because so much stray light came in from the sides.

This would be much easier with bellows and copier attachments. The key to a good result is getting a good exposure of all the channels by optically filtering the negative rather than relying on software.

What else? I don't have Photoshop so my raw file was imported as a gamma 1.0 tiff by PhotoLine. The negative convertion was with the ColorPerfect plugin using the ColorNeg module.

No ICE but with a better camera and better setup I can see this cranking out some very good results much faster than scanning. Since I don't have the right stuff and I do have my 9000 I haven't experimented any further with it.

 

[glhs116]

Give me a bellows and a real macro lens and I bet I can beat the 9000 (except for ICE and except for ultimate resolution) with my D80 even. With my current setup, the 9000 does much better.

Sam

 

[pellicle]

Give me a bellows and a real macro lens and I bet I can beat the 9000 (except for ICE and except for ultimate resolution) with my D80 even. With my current setup, the 9000 does much better.

Sam

I doubt that a better lens would double the quality you are getting. However any time you'd like to do a comparison I would be pleased to participate with my Nikon LS-4000. While not the 9000 it has similar spec save for the restriction of 35mm only (while the 9000 does 120 roll).

Particularly if we are talking negative I would be very interested to see the outcomes. If I had a 5DII I would have done this comparison already (if I would be still exposing 35mm negative, now its only 120 and 4x5).

 

[glhs116]

A better lens will mostly make me feel better. It bothers me to have whatever lens distortion is already in a shot compounded by additional distortion at the digitization stage. This consumer grade zoom I'm using plus the necessity of getting the camera and lightbox geometry perfect pretty much guarantees I increase the distortion in the image noticeably.

The bellows is what I think would make the massive difference. Essentially zero light leakage as well as full control of the geometry I think would make a massive increase in quality possible.

Of course my 10 megapixel D80 is never going to pull 24 megapixels out of a 35mm frame like the 9000. It is never going to include ICE. And it will always suffer a little from being a bayer based sensor unlike a proper scanner.

On the flip side, grain and defects are remarkably suppressed with this setup whilst sharpness remains very high. So that reduces the need for ICE. Also, once you got everything set up properly (which may take some time) you can scan film pretty much as fast as you can hit the shutter button. That is about one million times faster than my 9000. I have two rolls unscanned right now and I'm trying to figure out where I'll have the two hours or so a roll to scan them (2000dpi) in the upcoming weeks.

Sam

 

[pellicle]

Also, once you got everything set up properly (which may take some time) you can scan film pretty much as fast as you can hit the shutter button.

even with a bellows and a dedicated setup I'm willing to bet that its going to take an amount of time.

either way, my offer to scan a neg (or strip) of your film with my 4000 stands if you wish to do a comparison

 

[TareqPhoto]

I will join later to do the comparison with my Medium Format, i will have a bellow and maybe Large Format lens, or if no bellows then just my digital MF, i would like to see quality out of the setup against say Nikon scanners, i don't have Nikon scanner at the moment but only Epson V750.

 

[David A. Goldfarb]

I've been getting good enough results with my 5DII and Canon FD 35/2.8 Macrophoto (this is a high magnification lens that comes in RMS mount like a Luminar or a Photar) that I don't have any scanners anymore. I use various other lenses for larger formats, but the 35mm Macrophoto is great for 35mm and medium format, if I need that much resolution. I do everything on a copy stand with strobes for prints or a lightbox for transparencies and negs. I don't shoot very much color neg, so I haven't really sorted that out.

I find the camera gives me more exposure options than a scanner, and if I want higher resolution than I can get in one frame, I can always shoot multiple panels and get as many megapixels as the film is capable of resolving and as I have patience for, and I could even shoot more than one exposure and use HDR for a neg or transparency with an unusually wide density range (I think I've only done that once or twice).

For instance I had a slide of a historic event that I made in 1989 and that we wanted to print 4-color offset about 2.5 x 3 inches in a medium-gloss brochure, and the page design that worked involved about a Minox-sized bit of the full 35mm slide, so we used the crop from a 21 megapixel full frame image to lay out the page, and then I went back and put on a few more extension tubes and made a 21 megapixel image of approximately the crop we needed. Of course we were bumping against the grain limits of the film at that point, but we were able to pull out a lot more detail than we originally had. For that size and type of output, of course, in theory one shouldn't need that much resolution, but the more you can get from the film original, the more detail you'll have in the final image.

 

[cmo]

David, I use a similar setup with an 80mm f4 Olympus bellows head. Results are superb.

What light source do you use? I find that I need a lot of light to achieve short exposure times and consider buying a super-bright 160 LED video light plus a piece of frosted glass as a 'lightbox on steroids'.

 

[David A. Goldfarb]

I'm just using a small Logan 5500K lightpad, so there's probably a little noise penalty to be paid from the longish exposures, but it doesn't seem too severe.


Sent from my iPod touch using Tapatalk

 

[df cardwell]

Years ago, I set up my Canon D30 on a copystand and used it to proof and digitize sheet film. Today, I use a Nikon D700 for both sheet film and roll film, and haven't touched my film scanners in over a year. I recently made a large palladium print from an old 35mm TX negative and pulled several tricks from the hybrid bag to make the picture I wanted:

1) I made an enlarged 6x9 image on a sheet of 8x10 film.

2) I copied the enlarged positive with the D700 with a 60mm Micro-Nikkor.
At this step, it was easy to record detail from the original negative (that no scanner that I could ever afford)

3) which made a digineg that held the lovely, crisp, Rodinal grain and printed easily

My light source for 4x5 and 8x10 film is a homemade plywood box with a 'old' Omega color head inside.

For roll film, I use a Beseler duplicator.

AS for technique, what David said.
It is easy and a pleasant way to work .
(unlike scanning, which has never, in 20 years, has ever been fun)

Here is a typical proof of a 4x5 image that takes advantage of the RAW potential of current DSLRs and software. Good proofs, and easy.

 

[markbarendt]

I just did a down and dirty try, never having played with this.

Held a negative up to a light and shot it handheld with my D200 and a 50mm.

Opened it in adobe raw white balanced off a piece of jewelry, still a negative here.

Opened in PS added a curves layer with the negative preset (not color neg).

Added a second curves layer and used the gray point eyedropper.

It was actually reasonably balanced.

Guessing a reference frame with a gray card on film should allow you to do this once for any given lighting situation.