Slide Duplicator + 5D Mark II

<p>So what do people think about the quality attainable by slide duplicators (e.g. <a href="http://www.bhphotovideo.com/c/product/464076-REG/General_Brand_SDD_Digital_Slide_Duplicator.html">http://www.bhphotovideo.com/c/product/464076-REG/General_Brand_SDD_Digital_Slide_Duplicator.html</a> ) and >20 megapixel resolution cameras?</p>

<p>I have access to all the Nikon film scanners, but there are always so many problems (film flatness, pepper grain, depth of focus, scan time, etc.)</p>

<p>No idea of the kind of quality you'd get from this separate, but curious if anyone's tried a slide duplicator with a full-frame 20MP+ camera. I've searched photo.net but haven't seen any resolution comparisons.</p>

<p>Using this setup, I'd imagine you could get a lot of information from the shadows quickly, but taking multiple exposures.</p>

<p>Just throwin' it out there. Let's discuss if anyone has one or has any opinion!<br>

Thanks,<br>

Rishi</p>

<p>Hello...</p>

<p>Since you have a full frame DSLR you could use an old fashion slide duper with a T-Mount. They have small lenses of their own and a slide holder. If you are still on Sleezebay, after all they have done to their customers, the slide dupers are cheap there. In the 90s, I used to get better results making "internegs" for printing slides than my local photo shop which used a Nikon with a high line macro lens on a stand above a light table. I later bought a Repronar which was supposed to be the highest resolution slide dup creator but it only gave the slightest upwards quality difference. </p>

<p>I notice that the one on sale at B&H uses the camera lens to focus to the slide holder. I'm guessing it either has an additional lens or depends on the digital camera's lens to focus quite close. I also noticed that they said to use a tripod. Since the slide is firmly attached to the lens through a rigid tube and will move with the camera/lens, why does one need to use a tripod???? When I used my T-Mount slide duper I just went outside and pointed it a few degrees off the sun, hand held, of course. During a snow storm, I used a 500w blue bulb about two inches from the opal glass. I used auto exposure unless the slide itself was quite light or dark or was composed of something like a beach scene. Then I'd use exposure compensation and bracket. </p>

<p>The proceeding information is free and worth every penny. </p>

<p>Tom</p>

<p>Re: exposure & color accuracy, that is a problem easily solved using a Hutch color target or an IT8 target with a fixed light source.</p>

<p>Say I use a setup like this:<br>

<a href="

<p>There's a fixed light source. I then find the exposure for the target such that clear film (slide) maps to 'almost white' (say, 254, 254,254 for RGB); that should be my exposure for all slides, because it'll guarantee the best exposure of shadows that do not blow out highlights. Maybe bracket +/- 2stops to get more information out of the shadows.</p>

<p>Turn off all auto camera functions and whatnot.<br>

<br /> Create a profile based off the target shot.</p>

<p>Apply to all subsequent slide shots. Should get perfect accuracy color/exposure.</p>

<p>Basically, I want to know if this'd be <em>better</em> than a Nikon LS-5000/4000. By the time I solved all the LS4000 problems (by using AN glass to hold the film flat, but then a Scanhancer to diffuse the light so the LS4000 wouldn't pick up the AN texture, yada yada experimentation-that-makes-you-want-to-kill-yourself), my scans require 40 minutes per frame. <strong>40 minutes</strong> . The Scanhancer decreases light output that much. And that's also because I have to 4x multiscan (at the very least) so that the scanner doesn't introduce a tiny vertical pattern into scans (which I think, to my best guess, results from scanner vibration when it scans too fast?)</p>

<p>Anyway, without belaboring the point of the horrors of scanning with 35mm format Nikon scanners (oh, did I mention optical flare?), I'm just wondering if this set up would be better.</p>

<p>And you can do HDR easily.</p>

<p>Let us know if you try and find out. I've been curious about "scanning" 120 film using a DSLR and a 1:1 macro lens by taking and stitching multiple shots, as an alternative to flatbeds which leave a lot to be desired (in my opinion). I realize that for MF it might be more time consuming, but I'm wondering if it would also result in much better IQ.</p>

<p>Well I don't have any of the equipment :(</p>

<p>But tomorrow maybe I'll just go to the camera shop and try to take a picture of one of my slides with a backlight & a 5D Mark II w/ a 50mm lens with a 50mm extension tube... see what I can get.</p>

<p>Well FWIW, I tried this a while ago out of sheer frustration with the problems I was having with my Nikon LS5000ED. I used a Canon FD Duplicator 35, FD auto-bellows, FD 50mm f3.5 macro lens, FD-EOS convertor (the no-glass type), and EOS DSLR body. The body was borrowed from a friend and I don't remember the exact model but I think it was around 8-10 MPand not full-frame. Light source was a "daylight" compact fluoro.</p>

<p>The results were, in a word, c**p. Nowhere near the quality I was getting from the Nikon scanner, and I use the word quality lightly because the Nikon scans already left a lot to be desired. I shelved the idea there and then.</p>

<p>Now 8-10MP is not 20MP and I confess I'm not an expert at manipulating images from a digital camera so the images from a 20+MP full frame sensor in the hands of an expert may well give a much better result. It will be interesting to see - I've still got my FD slide duping gear for that very reason....</p>

<p>IMO Scanhancer introduces more challenges than it's worth (as you've explained)...no evidence that it serves a significant purpose (except, arguably, with badly exposed Velvia). Your LS5000 holds film plenty flat without glass unless somethings seriously amiss with your slide mounts.</p>

<p>If your slide mounts are worse than usual, re-mount in Wess or GEPE mounts (the glass versions don't help...according to tests reported here on P.N with the even-more-focus-troublesome Minolta 5400.</p>

<p>I do understand what you're hoping to accomplish...I did tens of thousand slide of duplicates with everything from Sickles and Forox duplication/animation cameras to Nikon and Canon with macro lenses and Pentax 6X7 with bellows and enlarging lenses. The Nikon Scanner's handling time is no slower than a slide duplicator's handling time, and it's scan time is less than a second before autofocus and Ice, still adds up to no more than a second WITH autofocus and Ice... you lose Ice with a slide duplication approach and only gain seconds, and no, Ice doesn't affect detail resolution or grain unless you use the heaviest settings. Compare that to the grief and hours added by Scanhancer.</p>

<p>I use a Nikon V, sometimes with the LH-3 strip film carrier. That accessory carrier is helpful with curvature at ends of rolls and with film that has skinny framelines (old Leica) or fogged framelines/stained base. The Nikon V is literally grain sharp corner-to-corner and Photoshop can deal nicely with the issues Scanhancer (which I have) trys to address. I don't have problems with slide focus, whether Kodak paper mounts going back to WWII (my parents) or Pakon plastic. Damaged slide mounts call for remounting, of course.<br>

The trick is to focus about 1/3 of the way from the edge...the Nikon's depth of focus is more than sufficient to deal with that curvature.</p>

<p>...incidentally, for MF...flatbeds perform a lot better with focusing carriers or glass-mount carriers (I have both from www.betterscanning.com). My old Epson 3200 flatbed with MF leads to prints that are sharper than most enlargers can deliver...but not as sharp as a Nikon 9000 can produce. Unfortunately that scanner doesn't offer Ice...eventually I'll try MF slide duplication/printing using 15mp Pentax with macro prime, since the Epson is ultra-slow with big film.</p>

<p>John,</p>

<p>Scanning film makes me wanna pop a cap in my skull.</p>

<p>One of these days I'll write a treatise on everything I've done/tested for the past year and a half, but, for now, let me just say that:</p>

<p>For LS-5000 & LS-4000 scans to be perfectly in focus edge to edge, though easier with the LS-5000 (i.e. sometimes works with no effort at all with the motorized feeder, but sometimes it doesn't), I <strong>have to</strong> use glass. Using glass, the film is so flat that the only variation in focus is due to the plane of the film/glass not being perfectly parallel to the axis of travel of the optics.</p>

<p>But that is fixable by sticking a card or two under the holder to get it pretty much so flat that the manual focus reading from one end to the other end in Vuescan only jumps from like .243 to 0.258... which I have found to be within the limits of even the LS-4000, which has a dismal depth of focus.</p>

<p>The Wess mounts are pretty good too, as they pull on the spokes of the film... but for some film, this doesn't work very well, and for other film, it tensions too much, creating ripples across the film.</p>

<p>The glass, so far, <strong>always works</strong> . Except for old really crappy film that has lots of curl... even then, I could probably solve it by adding some heavier glass.</p>

<p>All this comes at a significant price, however. That is, the texture of the AN glass.</p>

<p>Luckily, there's a (not so) simple way to deal with this: a Scanhancer. Held close to the light source, it <em>completely</em> gets rid of any evident texture due to the AN glass on both slide and negative scans.</p>

<p>At the same time, it helps to heal what I like to call 'holes' in negative film (those white specks you see all over the place), AND gets rid of the extremely offensive Fuji pepper grain (to me offensive, anyway).</p>

<p>But, the <strong>Scanhancer also comes at a price</strong> . There is light fall-off on the long edges of the film scan (the 36mm side), and IR dust removal introduces artifacts into the image along these same edges that make IR dust removal unusable using the Scanhancer.</p>

<p>See the artifacts below (look at the bottom of the neck for 'IR Light'):<br>

<img src="http://staff.washington.edu/rjsanyal/ScanningWoes/Scanhancer-IR_Artifacts.jpg" alt="" width="800" /><br>

<a href="http://staff.washington.edu/rjsanyal/ScanningWoes/Scanhancer-IR_Artifacts.jpg">Link to Full-Size Image</a></p>

<p>I believe this has something to do with the IR channel of the scan looking like this:<br>

<img src="http://staff.washington.edu/rjsanyal/ScanningWoes/Scanhancer-IR1vsIR16.jpg" alt="" width="800" /><br>

<a href="http://staff.washington.edu/rjsanyal/ScanningWoes/Scanhancer-IR1vsIR16.jpg">Link to Full-Size Image</a></p>

<p>I believe this problem can be fixed by getting the Scanhancer even closer to the LED light source of the Nikons... like one is able to for the Canon light source, pictured here with a Scanhancer on top (courtesy of Erik who sent this to Roger Smith):</p>

<p><img src="http://staff.washington.edu/rjsanyal/ScanningWoes/Scanhancer_CanonFS4000.JPG" alt="" width="800" /></p>

<p>We'll see.</p>

<p>Damn, I thought I was gonna keep this short.</p>

<p>Anyway, for all these reasons, and more (I want increased resolution... the Nikons just do not capture everything that's there in the film, as is evidenced by a light microscope and an Imacon 848). Hence I was actually thinking about the Minolta Dimage Scan Elite 5400 used off of eBay.</p>

<p>That already has a diffuser built into it that, apparently, doesn't cause these IR artifacts. As long as I can adjust the R, G, and B analog gains separately (I use this to get rid of negative film base casts as per the Erik Krause workflow), I think I'll be happy with it.</p>

<p>But, that also hinges on the ability to place some glass in the glass carrier of the DSE 5400... is that even possible?</p>

<p>Rishi</p>

<p>I do not beleive the 5400 I allows individual control over R, G and B as it uses a fluorescent lamp and not LEDs like the II or Coolscans.</p>

<p>Hi Roger,</p>

<p>Ed Hamrick just wrote me saying: "Yes, what Nikon calls "analog gain" is really variable exposure time.<br /> The Minolta can control the exposure time for each color."</p>

<p>So perhaps the Minolta is able to change exposure time of all the red, green, <strong>or</strong> blue sensitive pixels on the CCD? I believe the Minoltas use a primary color filter to separate the light.</p>

<p>On that note, I wonder if that's inherently inferior to separate red, green, and blue light sources. I doubt it, given that even the Imacons use one fluorescent tube as a light source. Besides, proper color profiling fixes most of these problems, so comments about 'accurate color' without proper IT8/HCT color profiling don't hold much merit.</p>

<p>Rishi</p>

<p>"I do not beleive the 5400 I allows individual control over R, G and B as it uses a fluorescent lamp and not LEDs like the II or Coolscans."</p>

<p>I believe Nikon scanners use white LEDs as the light source.</p>

<p>"what Nikon calls "analog gain" is really variable exposure time"</p>

<p>I always thought it was amplifier gain as implied by the name. Changing analog gain doesn't seem to affect scanning time which should be the case if actual exposure time was being varied. </p>

<p>No, the Nikons have separate red, green, and blue LEDs.</p>

<p>It IS interesting though that people claim that scan time doesn't change with changes in analog gain. I wonder if this is just subjective or whether people have actual real empirical data if this is the case.</p>

<p>Well, one way to find out: I'll test and report back.</p>

<p>For better or for worse, I just purchased a Minolta Dimage 5400 off of eBay. Hopefully it'll live up to (some) expectations? Just hope that one can scan it on its side, so that the weight of glass can hold the film flat.</p>

<p>Rishi</p>

<p>"No, the Nikons have separate red, green, and blue LEDs."</p>

<p>And Ir LED. Yes, you're right. I should learn to read the specifications first...</p>

<p>OK, I call bullcrap on whoever says that changing one of the red, green, or blue analog gains in Vuescan with the Nikon DOESN'T change exposure time.</p>

<p>I just scanned with the following settings on a LS-5000:<br>

<br /> RGB Exposure: 5<br>

Red analog gain: 1<br>

Green analog gain: 1<br>

Blue analog gain: 1</p>

<p>Preview took: 28 seconds</p>

<p>Then I scanned with the following settings:</p>

<p>RGB Exposure: 5<br>

Red analog gain: 1<br>

Green analog gain: 8<br>

Blue analog gain: 1</p>

<p>Preview took: 55 seconds</p>

<p>So that settles that. It's just increasing exposure time on the CCD. Just like Ed Hamrick said.</p>

<p>Once again, misinformation begets misinformation on this thing they call the 'inter-web'</p>

<p>:)</p>

<p>Cheers,<br>

Rishi</p>

<p>"So that settles that"</p>

<p>Not really.</p>

<p>I just did the same test with Nikonscan v4.0.2 on LS5000ED.<br>

Red analog gain: 0<br>

Green analog gain: 0<br>

Blue analog gain: 0</p>

<p>Scan time 25 seconds</p>

<p>Red analog gain: 0<br>

Green analog gain: 2 (maximum in Nikonscan)<br>

Blue analog gain: 0</p>

<p>Scan time...er... 25 seconds.</p>

<p>What is Vuescan doing that Nikonscan isn't?</p>

 

<p><strong>Oh snap.</strong></p>

<p>At least I had "gains <em>in Vuescan</em> " in my statement, else I also would've been a participant in spreading misinformation... whew :)</p>

<p>I think the more appropriate question, then, is 'What is NikonScan doing that VueScan isn't?' This is clearly a question for Mr. Hamrick.</p>

<p>Thanks for that test Lex.<br>

Rishi</p>

<p>Mr. Hamrick raises a good point, Lex. The Analog Gain may not be applied during Preview in NikonScan (though it IS applied in Vuescan during Preview)... so see what happens to the <em>scan time</em> when changing individual analog gains.</p>

<p>Here's Mr. Hamrick's response:</p>

<blockquote>

<p>Do the math:<br /> <br /> (5+1+8+1)/(5+1+1+1)*28 = 53<br /> <br /> It's definitely variable exposure time - no question about it. I've<br /> looked at the commands that NikonScan sends to the scanner,<br /> and they're in different units than in the user interface. I don't<br /> know, but NikonScan's units might be EV or something.<br /> <br /> Whoever did the test on photo.net screwed up. They were<br /> probably measuring preview time, but with NikonScan I think<br /> the "analog gain" setting might only be used with the scan.</p>

</blockquote>

<p>Want to give the scan, rather than the preview, a shot Lex? I don't have NikonScan on my Mac right now & last I checked could never get it to work on Leopard... and don't feel like switching to Vista right now!</p>

<p>Cheers,<br>

Rishi</p>

<p>No I was measuring actual scan time not preview time. Rishi, you changed my wording from "scan" to "preview" in the mail to Mr Hamrick. Check my post, I said "scan" time.</p>

<p>Now, to clear up the doubt concerning scan vs preview, Nikonscan DOES apply any change to analog gain to the preview as well as the scan.</p>

<p>So did I screw up? Well, not in the data presented, but maybe by making a conclusion without collecting sufficient data. As Spock said, ""Insufficient facts always invites danger, Captain." Lesson acknowledged.</p>

<p>The two 25 second scan times were obtained with an unexposed leader of colour negative film, a bit hasty perhaps but that's what I had lying around within reach for a quick test. OK, I have repeated the original test with the unexposed leader, the results are the same, 25sec and 25sec, within a second anyway.</p>

<p>Now with a correctly exposed colour negative:</p>

<p>M R G B<br>

0 0 0 0 29sec<br>

0 0 2 0 27sec<br>

0 2 2 2 19sec<br>

2 2 2 2 19sec</p>

<p>And now with a correctly exposed Kodachrome positive:</p>

<p>M R G B<br>

0 0 0 0 19sec<br>

0 1 0 0 22sec<br>

0 1 1 0 23sec<br>

0 1 1 1 24sec<br>

1 0 0 0 24sec<br>

0 2 0 0 27sec<br>

0 2 2 0 31sec<br>

0 2 2 2 35sec<br>

2 0 0 0 35sec<br>

2 2 0 0 65sec<br>

2 2 2 0 90sec<br>

2 2 2 2 105sec</p>

<p>Clearly there IS something going on with exposure times. Make of it what you will :-)</p>

<p> </p>

<p>My sincere apologies, Lex. You're right, I misread and misquoted you. I guess this is what happens when you're still up at ~3 a.m. working on ridiculous problems that should've been dealt with by manufacturer's of multi-thousand dollar products to begin with :)</p>

<p>Looking over your data, I think I have it figured out now. For the positive slide scan, scan times do go up with analog gain adjustment. But they go down with negative film, right?</p>

<p>That means Ed Hamrick must have been right when he guessed that the scale of the adjustment in Nikon's software must emulate an EV adjustment... so for negatives, if you want a brighter image, you decrease exposure time of the scan (a horrible idea, of course... better to do that in post-processing!).</p>

<p>So I guess it is just an exposure time thing, as Mr. Hamrick maintained... makes sense, especially given that he claims he's sending the exact same hardware calls.</p>

<p>Glad we got it all figured out :)<br>

Cheers,<br>

Rishi</p>