An interesting way to do ULF

[Leonard Evens]

In the latest issue of View Camera Magazine, Richard Phillips describes a way to do ULF using a DSL to digitize his large negatives. He puts the negative on a light box and photographs it with a full frame Canon DSLR using a 60 mm macro lens. It looked like something I might conceivably try some day---if I had a UHL camera, which I don't at present, but who knows what the future will bring.

What I wondered about was whether I could do anything similar with my NIkon D80. The 18-200 mm zoom lens I use with my D80 works pretty well for general photography, but it does produce significant distortion. In principle, it should be possible to correct that digitally, but it never works out to my complete satisfaction. I wondered if there are any lenses (macro or not) available for my D80 which are distortion free to a sufficient degree that I need not worry about it?

[Mark Sampson]

The 55/2.8 (manual focus) and 60mm/2.8 (autofocus) Micro-Nikkors are the standard lenses for copy work in Nikon mount. I wonder if a consumer-grade DSLR has enough resolution to properly copy a ULF negative, but "enough" and "properly" will vary with the user.

[Scott Davis]

I think it also depends on your level of patience and the precision of your copy stand for how many samplings of the negative you want to stitch together.

[Vaughn]

Very sloppy job of it -- 7x17 neg on a light box and me standing on a chair over it. Panasonic DMC-FZ30

Good enough for the web and my limited PhotoShop skills! LOL!

Old Barn, Humboldt County
Ritter 7x17 camera, 24" RD Artar
X-ray film, developed in Ilford Universal PQ Developer

[Leonard Evens]

The 55/2.8 (manual focus) and 60mm/2.8 (autofocus) Micro-Nikkors are the standard lenses for copy work in Nikon mount. I wonder if a consumer-grade DSLR has enough resolution to properly copy a ULF negative, but "enough" and "properly" will vary with the user.

The D90 yields a 4288 x 2848 px image. At 300 ppi, this could yield a 14.3 x 9.5 inch image, and at 240 ppi, a somewhat larger image. My D80 doesn't do quite as well. But it seems as if one could produce something comparable in resolution to an 8 x 10 contact print from an 8 x 10 negative. Phillips discusses making the equivalent of contact prints for his negatives and says the results are hard to distinguish from prints made by conventional techniques. In addition, one has all the tools available with digital photography. I don't suppose one can do as well in terms of resolution as would be possible for larger prints with an enlarger. But I am never going to have an 8 x 10 inch enlarger!

So if someone had an 8 x 10 or 11 x 14 camera with lenses, and no possibility of using a darkroom, this could be a feasible way to make contact print sized prints, or, with some loss of quality, larger prints.

In my case, I could in principle make contact prints from large bw negatives, so if I got the ULF bug, I would be able to play. If I wanted to make larger prints, without laying out the cash for a scanner capable of scanning such large negatives, this could provide an alternative. (Even if I had room for a large format enlarger, I am not physically able to deal with such a large device because of arthritis.

Right now I'm satisfied with 4 x 5, so this is just a daydream for me.

[Ken Allen]

The lens is the trick here. You have to have a macro lens that is excellent in the 1:1 range. no substitution! For example I tried use a Hasselblad H2/39 MP back w/CF120 Macro and it works beautifully on originals larger than 8x10, but gets significantly worse in each step down. By the time you get to 35mm, the quality is fail.

But the Canon EFS 60mm macro on an APS-C Type camera (even a Rebel) and the quality can be excellent. It's a trick to get the correct tone reproduction and I'm still working on: getting a stable setup with no vibrations, reducing flare, dust reduction, etc.. But I think with this lens, and improving DSLR chips scanners may be a thing of the past.

I used the Canon 60mm on a Canon EOS Rebel T1i to scan some circa 1990's Kodak Lumiere 35mm slides and compared the scans to a few different scans. I also scanned a resolution chart, which can be misleading because dynamic range may limit your ability to see detail because of tone differences, but the B&W resolution target is not affected by dynamic range. The resolution chart indicated that the Rebel was achieving 2048 ppi actual. My non-scientific observations are below:

1990's Photo CD Scan Vs. Rebel is the clear winner in all categories: Dynamic range is far greater with significantly more shadow detail, Color resolution is unbelievably better, the color compression on the PCD was terrible turning skin tones in shadow into a solid red. The Rebel scans faithfully reproduced subtle skin tones deep into the shadows!

Nikon Super Coolscan 5000: The Nikon scanner provided a little more detail on a resolution chart (2896 ppi), but lost detail in the shadow areas. Something strange from the coolscan is that it will sometimes render spatial details into the dark areas, but loose the ability to accurately render color transitions in those same dark areas. All the shadows become colored or colorless. The Rebel had greater dynamic range, again rendering shadows with detail and accurate color. The Rebel also has smoother more accurate overall color.

Imacon Flextight Precision II: Provided slightly more detail (3251 ppi), and accutance, but since it was mostly recording grain, the Rebel still managed to capture all of the image detail. Reading the resolution from a scanned resolution target is much different the referring to a scanner specification. The Imacon rendered colors into the shadows like the Rebel and better than the Coolscan, but the rebel still seemed to have slightly more color accuracy in those tough areas. The extra detail of the Imacon did not seem to buy much in this case.

Speed? I assume that once we get a good workflow the Rebel will scan at a rate of 10-100 times faster than any of the above scanners, besides maybe the Kodak PCD scanner, but that's not relevant.

Obviously, on larger format film, all of this will have to be retested for validity. I assume I'm hitting the resolution limit of the chip with this lens, so res may increase with more megapixels. But for 35mm slides and negative films, 35mm DSLR copy system with an excellent macro lens is a go.

[Nathan Potter]

Of course this all depends on how much detail you want to capture from the ULF negative. At 1:1 you'll record perhaps half the detail in the field of view and so on. So it can require a lot of DSLR images to cover a ULF negative it that negative is of very high quality (resolution wise). If you go to a net magnification you end up with an ever decreasing FOV so need even more frames to cover. The stitching can become a nightmare unless fully automated.

Nate Potter, Austin TX.

[c.d.ewen]

Leonard:

Why not use a process lens?

I've mounted my D90 onto the back of a 4x5 Horseman. I don't know what 4x5 you have, but attaching things to a monorail like the Horseman is easy. All attachments, lensboards, bellows, GG, mount onto the standards in the same way.

I've shot through various LF lenses with the D90 by mounting a Nikon to T-mount adapter onto a lensboard on the rear standard. I used a more elaborate attachment, but it's possible to just superglue the adapter to the lensboard, and you're ready to go. Once you're focused, you just use the rise and fall to move across the image.

Charley