Home
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

More Scan vs Dupe Pics
View previous topic :: View next topic  


PostPosted: Thu Jan 21, 2010 11:34 pm    Post subject: More Scan vs Dupe Pics Reply with quote

With the recent discussions on film scanning happening here, I thought I'd do some more tests and comparisons and then show the results here.

I decided to compare my Epson 4990 against my duplicator rig, which is a contraption I put together that consists of a 55mm f/3.5 Nikkor, a small extension tube, the duplicator tube only from an Opteka "digital" duplicator, and the slide/negative carrier, which has position adjustments. After reading that Orio got better results when scanning his film with his Epson when the film was placed directly on the glass, I decided to give it a try, both with the film holder and without. Well, actually, even when I scanned the slide on the glass, I set the film holder on top of it, so the 4990 would be able to properly meter the photo. (Epson scanners determine exposure based on a slot molded into the top of the film holder.)

Here's the dupe rig I use:



The slide carrier is on the left and the negative carrier is on the right. I removed the piece that holds the diffusion panel from the front of the Opteka tube, and fortunately the slide and negative carriers slide over the tube's flange. It's a little fiddly, but once things are set, it works very well.

I selected a slide that was both unmounted and critically sharp. This particular slide was taken in 1986 using Fujichrome 100. I mention this because 1986's regular old Fujichrome is probably grainier than any of the various flavors of modern ISO 100 Fuji slide film.

I used a Canon F-1 and a Canon FD 50mm f/3.5 macro lens. To light the subject, I placed a white handkerchief on a glass table and then located a
Vivitar 285 strobe underneath. As I dimly recall, I just guessed at exposure, and then bracketed (because I have three slides of varying brightness in my files).

This is the image taken with the duplicator:


Here's a scan using the Epson 4990 @ 3200 ppi with the strip film holder:


And a scan with the slide placed directly on the 4990's glass surface. Also at 3200 ppi:


I did some post processing -- removed dust spots and bumped up the contrast somewhat. I didn't do any sort of sharpening, however.

It's pretty hard to tell the difference between the three images isn't it? Well, at 900x600 resolution for the web, that's understandable. 100% enlargements are needed to show the differences. So, following are 100% crops of each image.

Dupe 100% crop:


Epson 4990 with film holder 100% crop:


Epson 4990 with film on the glass 100% crop:


Note the reflections of the white surface in the lens. With the dupe, you can clearly see the weave of the handkerchief. With the crop of the image where the film was mounted into the holder, resolution isn't sufficient to resolve much in the way of detail of the weave. But the last crop above, the one where the slide was placed directly on the glass, shows some slightly more detail of the weave. Not nearly as much as the dupe, but it's there. Also, if you look at the writing on the lens bezel, it's pretty obvious that the last image is a bit sharper than the middle one, although it's noticeably softer than the dupe.

Before I tried scanning images directly on the glass, I tried shimming a mounted slide first. That is, I located it higher than it normally would sit. I cut up some (flat!) waste film into strips and used strips as shims. With one film thickness, the difference was negligible. With two film thicknesses, it became a bit more obvious -- the image was getting softer. With three film thicknesses, it was beyond question.

There is a drawback to scanning with the film on the glass, however. Newton rings appeared. You can see them in both of the above "glass scan" images, although they are somewhat faint. So, if I were going to make a habit of scanning my unmounted slides and negatives on the glass, I would need to get some ANR glass for this -- and I'd probably have to make what Epson calls a "film area guide" so the scanner will expose properly.


Last edited by cooltouch on Fri Jan 22, 2010 1:50 am; edited 1 time in total


PostPosted: Fri Jan 22, 2010 12:54 am    Post subject: Reply with quote

That's great for the duplicator. I have the Asahi Pentax slide dup coming soon. This is great news.


PostPosted: Fri Jan 22, 2010 1:55 am    Post subject: Reply with quote

TBaker wrote:
That's great for the duplicator. I have the Asahi Pentax slide dup coming soon. This is great news.


I'm guessing that your camera's sensor is not full-frame? Let us know if you can do full frame dupes of slides or negatives with your duplicator setup.

I would like to get a bellows with slide duplicator attachment, but I'm concerned that I might not be able to get all the way down to 1:1 with it.


PostPosted: Fri Jan 22, 2010 2:17 am    Post subject: Reply with quote

I have the 40D and on the fence about getting a 7D or 5D MKII.

Funny thing is I won a slide dup and he sent me the Asahi bellows. So he said keep the bellows and he'll send me up the slide dup. Can't wait.

Right now I have a Canoscan 8400 and it does a so so job with them but scanning negs I find too much a chore after do PP.


PostPosted: Fri Jan 22, 2010 7:12 am    Post subject: Reply with quote

The dupe wins, but I'm not seeing enough difference in the scanner images to bother with the on-the-glass scans. It seems like a lot of extra work for very little improvement.


PostPosted: Fri Jan 22, 2010 7:34 am    Post subject: Reply with quote

nice comparison, the Epson 4990 do a great job
you forgot to mention that a scan at 2400dpi is 7 minutes per slide on the Epson vs 1/125s on the dupe

last year, I made a comparison between the Nikon 9000 ED and the 40D
http://forum.mflenses.com/best-slide-scanner-t4661.html
Nikon 9000 ED on left


a dupe scan with a 5D is much better than with a crop like the 40D
the 5D show less diffraction at 1:1 and the sensor filter is weaker giving sharper images
http://forum.mflenses.com/ektar-high-tech-scanner-samples-t24360.html


PostPosted: Fri Jan 22, 2010 5:29 pm    Post subject: Reply with quote

TBaker wrote:
I have the 40D and on the fence about getting a 7D or 5D MKII.

Funny thing is I won a slide dup and he sent me the Asahi bellows. So he said keep the bellows and he'll send me up the slide dup. Can't wait.

Right now I have a Canoscan 8400 and it does a so so job with them but scanning negs I find too much a chore after do PP.


Well, you'll need both the bellows and the duplicator to dupe slides. Since they're Pentax, obviously you'll need the correct adapter for your 40D, and for a lens you should probably go with a macro -- you'll get your best results that way.

If I were in your shoes, I'd hold out for the 5DII. It's full frame -- the 7D's a crop body.


PostPosted: Fri Jan 22, 2010 5:32 pm    Post subject: Reply with quote

martinsmith99 wrote:
The dupe wins, but I'm not seeing enough difference in the scanner images to bother with the on-the-glass scans. It seems like a lot of extra work for very little improvement.


I'm inclined to agree. The only test left is to see if there is a sweet spot somewhere between the glass and the level at which the film holder holds the film strips. I'm inclined to think there isn't, but then I'm reminded of the images at betterscanning.com that shows the difference between a standard Epson film holder and one of theiradjustable ones. But they don't make one for 35mm anyway, so it's sort of a moot point.


PostPosted: Fri Jan 22, 2010 5:35 pm    Post subject: Reply with quote

The duplicator is sharper but introduces visible (and horrible) digital noise.

I'm glad to see that with your test too the film on glass is noticeably sharper than inside the holder. Like I said, that was exactly my experience, and AFAIAC I'm not going to use film holders anymore.


PostPosted: Fri Jan 22, 2010 5:52 pm    Post subject: Reply with quote

poilu wrote:
nice comparison, the Epson 4990 do a great job
you forgot to mention that a scan at 2400dpi is 7 minutes per slide on the Epson vs 1/125s on the dupe

last year, I made a comparison between the Nikon 9000 ED and the 40D
http://forum.mflenses.com/best-slide-scanner-t4661.html
Nikon 9000 ED on left


a dupe scan with a 5D is much better than with a crop like the 40D
the 5D show less diffraction at 1:1 and the sensor filter is weaker giving sharper images
http://forum.mflenses.com/ektar-high-tech-scanner-samples-t24360.html


On my computers, with either USB 2.0 or Firewire ports, at 3200 ppi, it only takes about 3 minutes, not 7. At 2400 ppi, it's only about 2 minutes. The time element is something to consider, but for the sake of image sharpness comparisons, I didn't see the need to mention it.

I find you 9000 ED/5D comparison very interesting. What was the resolution setting on the 9000 ED? The 5D's image is obviously sharper.

Your example brings up a dillema I'm faced with: should I continue to digitize my slides and negatives with my XS's APS-C sensor and 10.1mp, or should I just wait until I can afford a 5D or 5DII? I've already digitized most of my archive with my 4990, only to decide that the detail wasn't good enough, so recently I've begun the process all over again, this time digitizing with my XS. But then I began asking myself -- will it make any real difference? To answer this question, I have taken some pics of some of my sharpest slides by setting a light box on top of a copy stand, then placing the slide on the light box, and using my entire set of extension tubes with my Micro-Nikkor and my XS (using Live View and the self timer) to take pics of very small areas of the slide. In this way I was able to resolve all the detail the slides had to offer. Comparing these results to the 1:1 dupes I had taken, I can see that there was still more information to be captured that was beyond the resolving power of my XS at 1:1. So the answer to my dillema is that, if I want the sharpest dupes possible, I should just wait. But I hate waiting. Cool


Last edited by cooltouch on Fri Jan 22, 2010 6:05 pm; edited 1 time in total


PostPosted: Fri Jan 22, 2010 5:56 pm    Post subject: Reply with quote

Orio wrote:
The duplicator is sharper but introduces visible (and horrible) digital noise.

I'm glad to see that with your test too the film on glass is noticeably sharper than inside the holder. Like I said, that was exactly my experience, and AFAIAC I'm not going to use film holders anymore.


Some, if not most, of the digital noise is probably due to the fact that I slightly increased sharpness of the RAW image in Canon's DPP software, but did not use the noise reduction routines that it has. I could have smoothed out the noise, but it would have been at the expense of sharpness. I decided, in this case, to go for sharpness instead of smoothness.

So, when you scan your film strips on the glass of your scanner, are you using ANR glass atop the strips to eliminate the Newton Rings/Moire patterns?


PostPosted: Fri Jan 22, 2010 6:13 pm    Post subject: Reply with quote

Orio wrote:
The duplicator is sharper but introduces visible (and horrible) digital noise.

I'm glad to see that with your test too the film on glass is noticeably sharper than inside the holder. Like I said, that was exactly my experience, and AFAIAC I'm not going to use film holders anymore.


These samples seem to make my choice easier. I think a flatbed scanner is good enough for me, both for slides and 35mm negatives.


PostPosted: Fri Jan 22, 2010 7:58 pm    Post subject: Reply with quote

Michael wrote:
Your example brings up a dillema I'm faced with: should I continue to digitize my slides and negatives with my XS's APS-C sensor and 10.1mp, or should I just wait until I can afford a 5D or 5DII?

the scan was made 2 years ago with the 400D
Nikon 9000 ED scan at full resolution resized to 10Mpix
the slide is a sensia II developed in 1997, it doesn't have much more than 10Mpixels of details
you can continue to scan with your APS-C with great results


PostPosted: Fri Jan 22, 2010 8:14 pm    Post subject: Reply with quote

poilu wrote:

the scan was made 2 years ago with the 400D
Nikon 9000 ED scan at full resolution resized to 10Mpix
the slide is a sensia II developed in 1997, it doesn't have much more than 10Mpixels of details
you can continue to scan with your APS-C with great results


Oops, sorry I misread. But you mean 40D and not 400D, right?

As for your last statement -- I dunno. I was able to pull out more detail with the close-ups of the slides than I could get with my 1:1 dupes. So, I'm thinking that, while the 10.1mp dupes are certainly good, I can still get better ones with a higher resolution DSLR.

I'll have to post some pics to show what I'm getting at, but they are on a different computer.


PostPosted: Fri Jan 22, 2010 8:24 pm    Post subject: Reply with quote

Michael wrote:
Oops, sorry I misread. But you mean 40D and not 400D, right?

I got 3 scanners, the 400D, 40D and just now the 5DII
the comparison was made with the 400D, same sensor as your APS-C but without liveview, a pain to focus at 1:1
I made the mistake when I label the pic 40D, my memory start to fail Smile


PostPosted: Fri Jan 22, 2010 9:26 pm    Post subject: Reply with quote

cooltouch wrote:
martinsmith99 wrote:
The dupe wins, but I'm not seeing enough difference in the scanner images to bother with the on-the-glass scans. It seems like a lot of extra work for very little improvement.


I'm inclined to agree. The only test left is to see if there is a sweet spot somewhere between the glass and the level at which the film holder holds the film strips. I'm inclined to think there isn't, but then I'm reminded of the images at betterscanning.com that shows the difference between a standard Epson film holder and one of theiradjustable ones. But they don't make one for 35mm anyway, so it's sort of a moot point.

It's a shame that they don't as I think there would be a market.

Anyway, thanks for saving me some money. Cool


PostPosted: Fri Jan 22, 2010 10:37 pm    Post subject: Reply with quote

cooltouch wrote:


So, when you scan your film strips on the glass of your scanner, are you using ANR glass atop the strips to eliminate the Newton Rings/Moire patterns?


I currently don't have a piece of ANG that is large enough to cover the whole strip. So I just place the two little pieces over the first and last frame and leave the others "naked"
Which is perfectly ok, in fact the "secret" of the sharpness of this procedure is not in the ANG glass, it's in the proximity to the lower glass (that of the scanner) and in the flatness of the film that is caused by the pressure of the ANG glass.
In fact, if you don't mind moving pieces around, you don't really need the ANG glass. All you need is a piece of something that has thin thickness like that of glass, and similar weight. In other words something that is heavy enough to keep film flat and thin enough not to cause the lid of the scanner to stay higher than normal (which would be bad because it provides the upper illumination.
So say you have two pieces of aluminium, they are allright just the same, the only difference is that they will not let you scan the frames that they cover - but they will flat down the film so the free frames will scan just as perfectly.
Example: try with two heavy coins. Such as two 2 Euros coins. They will work. If film is very bent, use two coins.
Of course covering with ANG glass would assure 100% flatness.
The coins, maybe 95% flatness. But in most cases it will work.


PostPosted: Sun Jan 24, 2010 5:40 pm    Post subject: Reply with quote

Cooltouch:

Thank you for your efforts in providing these tests. I had been looking at Coolscan prices on ebay, and LS5000 prices are about 50% more than they were about 3 months ago. But your tests have convinced me that I would rather copy film in a DSLR.

However, rather than use a 50mm macro lens, I am thinking of using a 50mm or longer enlarging lens attached to a bellows. Any suggestions as to the best enlarging lens for the application?


PostPosted: Sun Jan 24, 2010 7:30 pm    Post subject: Reply with quote

pablo wrote:
Cooltouch:

Thank you for your efforts in providing these tests. I had been looking at Coolscan prices on ebay, and LS5000 prices are about 50% more than they were about 3 months ago. But your tests have convinced me that I would rather copy film in a DSLR.

However, rather than use a 50mm macro lens, I am thinking of using a 50mm or longer enlarging lens attached to a bellows. Any suggestions as to the best enlarging lens for the application?


Not really, although I suspect an enlarging lens will work well. Back when I used to have a darkroom I had a Nikkor EL 50mm that certainly did a good job for print making. I also had a Schneider Componon and a Rodenstock somethingoranother. These two lenses also did a good job.

Unless you already have a bellows that you can attach an enlarging lens to, though, you can most likely buy an old Micro Nikkor and extension tubes often for less than the cost of bellows alone.