Ultimate giclée lens?

[jonathan7007]

TrumpetPower,
Love the setup, interesting looking target for calibration. I miss my studio space...

You didn't say this outright but you must have decided that the lens profiles that adjust for distortions are not adequate. Or is it good enough for only certain art work? Does, say, Lightroom use Adobe profiles that change the Canon 50 f2.5 much at all? ...and did DxO have any different profiles that were better, worse?

Interesting warning on the shift and merge tactic. I was planning to try that to allow me to pitch local artists for this service. I have the 90 and 45TSE's. I do wish the 45 was sharper. A friend of mine has the 50 macro so will have to try the 45TSE side by side with the 50 -- pointed at the same material -- for comparison. 

jonathan7007

[TrumpetPower!]

Here is the one i'm thinking of, 85mm f/1.0 Repro-Nikkor.

That is an intriguing lens. Not at all what I'm looking for, of course, but still a very interesting diversion. I'll have to keep my eyes out for one in the wild. Thanks!

When I do copy work, usually for 3d painting/torn canvas stuff and some sculpture, I use one of these if I'm shooting on small format: http://www.hartblei.de/en/sr120m.htm

I've stumbled across those before, but always figured that, if I ever really wanted to go that route, it might make more sense to skip them and go straight to medium format...which is, alas, probably not a realistic move. Still, their 40/80/120 trio of lenses would be very nice to have...if I had an extra $12K laying around....

Otherwise, scanning back with a large format for the best quality in repro.

True, but you might be surprised at what you can do with 135 format. I can comfortably do 12x18@300ppi in a single frame, which is plenty of resolution for most of the stuff I do. The rare exception I can reasonably accommodate with stitching.

And...it's been a long, hard slog, but I get superlative color accuracy.

TrumpetPower,
Love the setup, interesting looking target for calibration.

It's my own creation. The majority of the patches are just printed from an iPF8100, but there's a ColorChecker in there with paints matched at Home Depot (and the spectral characteristics are the same as the original), plus another 24 patches of various "interesting" paints, plus a light trap, plus some PTFE. And a dozen pieces of wood. I have some ideas of what to do for the next version, but this is probably already better than anything on the market.

I miss my studio space...

And I wish mine were bigger! It started life as a garage. A previous owner converted it into something vaguely resembling an extension of the living room. I've run copper pipe around the perimeter of the tops of the walls and made drapes I can hang from them. I can go from a normal-looking generic empty room with white walls and (fake) hardwood floors to a white box to a black box to anything between.

...but it's only 10' x 20', with slightly more than 8' ceilings. I've got a big opening to the living room that I can use if I need more working distance, but that gets awkward to work in....

You didn't say this outright but you must have decided that the lens profiles that adjust for distortions are not adequate. Or is it good enough for only certain art work? Does, say, Lightroom use Adobe profiles that change the Canon 50 f2.5 much at all? ...and did DxO have any different profiles that were better, worse?

DNG color management is a cruel and unusual joke. I have the scars -- deep scars -- to prove it. And correct color trumps even the worst distortion you're going to get in a decent lens, by a wide margin.

My workflow corrects both for light falloff (and uneven illumination, should I screw up the setup) and for color. I don't touch geometry; with the 50 CM, there's no need.

I use Raw Photo Processor for raw conversion; Argyll CMS for color management; EquaLight for equalizing illumination; and Photoshop for whatever else (minimal sharpening, cropping, stitching, that sort of thing).

Elle Stone pointed me to a technique for nailing exposure in-camera and then perfectly nailing both exposure and white balance in post. I followed up that message with some refinements I've made to her workflow...but, alas, the text got stripped and only the screenshots remain. Anyway, the straight-out-of-RPP images I get are already very, very close to correct. I then equalize illumination with EquaLight. The first shot (after getting everything set up, of course) is the target, which I then feed to Argyll to generate a profile. For the artwork itself, I apply the profile in RPP before outputting it, and then equalize that with EquaLight. Then it's off to Photoshop for sharpening and cropping, and then back to Argyll for conversion to the printer's profile, and back again to Photoshop to actually print.

Interesting warning on the shift and merge tactic. I was planning to try that to allow me to pitch local artists for this service. I have the 90 and 45TSE's. I do wish the 45 was sharper. A friend of mine has the 50 macro so will have to try the 45TSE side by side with the 50 -- pointed at the same material -- for comparison.

If you're going to do a panorama shift with the TS-Es, you'll need to figure out a way to keep the lens fixed and shift the camera behind it. And, even still, you don't get all that much extra out of it.

What I meant was to leave the camera alone, and to slide the artwork on the table (or, in my case, floor).

There's definitely a big pent-up demand for quality giclée work, even here in the Phoenix metro area. The catch is that artists care primarily about color reproduction, and nobody outside of the national museums knows how to do quality color reproduction. If you can get to the point where it takes several long seconds of close scrutiny by the artist before spotting a part of the image where the colors aren't quite a perfect match, you might have some success.

Cheers,

b&

[dr croubie]

If you're going to do a panorama shift with the TS-Es, you'll need to figure out a way to keep the lens fixed and shift the camera behind it. And, even still, you don't get all that much extra out of it.

Should I say Hartblei to the rescue again? (too bad it's only for the 17 and 24mm so far, maybe when the 45 mk2 gets released it'll "support" them too)

[TrumpetPower!]

You do not need that ridiculously priced lens clamp to create parallax free stitched TS-E images, just slide the camera back the other way to the lens shift, even a modest body plate of the Arca Swiss variety has 24mm of shift capability.

Holy Hornswoggle, Batman -- that's an entire Grover Cleveland for an oversized tuning fork! It's half the price of a TS-E 24 II! I know a shade-tree mechanic who could fabricate something like that and wouldn't charge more than a Benjamin for it, maybe even half that.

privatebydesign is right. Count the number of ticks on the lens indicating mm of shift, and move the clamp an equal number of ticks the opposite direction. Done and done. Only potentially problematic if you want to do a vertical panorama, but you can use the dropout in your ballhead -- and who does vertical shift panoramas, anyway?

b&

[TrumpetPower!]

TrumpetPower!

If you really want to know how far you can go with this stuff, the story of the Unicorn and the Chudnovsky brothers is as good as it gets.

http://www.unicorns.co.za/tapestries/capturing-the-unicorn-how-two-mathematicians-came-to-the-aid-of-the-met.html

As should be no surprise for the New Yorker, that was a good story told well.

Fortunately, I'm nowhere near insane enough to tackle a project of that magnitude.

I also like to think that I would have done some small-scale experiments, discovered the fundamental flaws in the technique, and solved them before doing the full shoot. Specifically, I'd have made sure the tapestry was laid flat and not touched during the shooting, and I'd have come up with some sort of etch-a-sketch mechanism for suspending the camera over it. I'd probably have also thrown multiple cameras at it, too. The tapestries are about 8' x 12'...be conservative and shoot them in 8" x 12" sections, that's 144 exposures. Let's say it takes 30" to reposition the camera on the etch-a-scetch and make an exposure. Use three cameras simultaneously, and you could shoot an entire tapestry in under a half an hour. Double the cameras and get the per-shot interval down to 15" and you've got the whole tapestry imaged in six minutes, which would have easily solved all the problems the brothers had to deal with. Now, put the tapestry on a rolling bed with wheels that you can rapidly slide it into place, and have two other such beds set up, one with pure white and the other with a grid and a bunch of ColorCheckers (except you'd want something with way more colors, and heavy representation of the same and similar pigments used in the art). In half an hour, you've got the art imaged plus enough reference materials to perform any necessary corrections.

And that's assuming the technology they described in the article. Today...today, they'd be doing multi-spectral imaging using some sort of large format, probably even in 3-D, and I doubt that the imaging process itself will take more than a few minutes. The setup and post-processing and everything else, I'm sure, takes forever...but not the button-pressing part.

One assumes that the budget is not much of an obstacle for the Met....

b&

[dr croubie]

I know a shade-tree mechanic who could fabricate something like that and wouldn't charge more than a Benjamin for it, maybe even half that.

Count the number of ticks on the lens indicating mm of shift, and move the clamp an equal number of ticks the opposite direction. Done and done. Only potentially problematic if you want to do a vertical panorama, but you can use the dropout in your ballhead -- and who does vertical shift panoramas, anyway?

Well yeah, it's expensive. Hell, I've even just bought Medium Format lenses (like Zeiss Flektogon 50mm f/4.0 for Pentacon Six), hacked up a $10 chinese-ebay tripod clamp for some other lens, and stuck it on a $100 P6-EF shift-adapter.

And measuring and sliding also works, especially for landscape, i've done that too (Sunwayfoto make a nice sliding rail with measurements). But I wouldn't do it for macro or even A3-sized repro-work, you'll never get it *exactly* right, and errors are magnified ridiculously at short distances.