Anyone trying pixel-shifting camera scanning?

[bags27]

I currently use my Fuji GFX-R (MF 50 mps) with a Mamiya 625 f/4 120mm macro to scan my MF (6x6, 6x7, and 6x9) negatives. Works well. But always looking for better.

I read about a Large Format shooter who so liked his pixel-shifting Olympus micro 4/3 for scanning (I imagine he stitched several) that he actually got rid of his drum scanner. Wow! That's quite a claim.

I could swap my Fuji for a used pixel-shifting Panasonic S1R (FF, 47 mps) to pair with my Sigma L mount macro lens. I'm thinking of renting it first to compare. But just wondering if anyone's tried this technology for film scanning.

Of course, if that's great, the new pixel-shifting Fuji GFX 100-S might be even better, but that's not being discounted from $6k, and there has to be a limit to how effective larger sensors are going to be, right?

thanks!

 

[MattKing]

Thread moved from "Digital Negatives" to the Scanning and Scanners sub-forum.
The "Digital Negatives" sub-forum relates to the inkjet or laser printed negatives used by many people contact printing using Alternative and Traditional printing methods like cyanotypes.

 

[bags27]

Thread moved from "Digital Negatives" to the Scanning and Scanners sub-forum.
The "Digital Negatives" sub-forum relates to the inkjet or laser printed negatives used by many people contact printing using Alternative and Traditional printing methods like cyanotypes.

Thanks and sorry. Of course!

 

[4season]

Off and on, I've fussed around with camera-scanning, including 80 mp pixel-shifted using Olympus Pen-F + Canon FDn 50 mm macro lens. To date, have not been particularly blow away, because at least with color slide or negative film, I quickly reach a point where I can see no additional details being resolved. I think black and white negatives offer much greater potential in this regard. And for grins, I think I might like to see just what sort of killer resolutions I might achieve with a combination of great optics, high resolution b&w film, and careful technique.

 

[bags27]

Thanks much! I was in fact thinking of grain in MF B&W. I spend so much time (and money) getting low/no grain negatives, but sometimes I'd like to have my scans reflect what little there is. Thinking pixel-shifting is the last frontier from everything I've tried.

My biggest gain so far was actually switching to a monstrous copy stand--probably a rebadged Beseler enlarger. I thought my original kit was stable, including an electronic trigger, but when I used this new stand I couldn't believe how much sharper my 135 scans were.

 

[McDiesel]

Yes, I have been experimenting with Fuji GFX 100S and Sony a7R IVa. Don't even know where to begin... Let's start with the difference between the two.

I couldn't find a usable macro lens for the GFX platform. There are no true macro lenses that are native to the mount, and I believe that the auto-focus is a must have capability. Without auto focus the scanning experience goes down the drain approaching the sad state of scanning with a film scanner. I only have 2-3 minutes to spend per roll before the agony kicks in, so the Fuji was out. Also, Fuji is really weak on software. They don't have robust tethering applications like Canon or Sony do, you have to use awful plug-ins to tether, just FYI. Everything below comes from my experience scanning with the Sony.

Pixel shift raises the bar for lens quality tremendously. Even without pixel shift, I am yet to find a macro lens that doesn't lose detail in the corners. I have tried Canon 100 f/2.8L macro, Sony's 90mm macro, cheaper Ronikon and older Nikon macro lenses, and finally found the champion: Sigma 105mm f/2.8 DG DN Macro Art. Like other lenses, it doesn't offer resolution parity between the corners and the center on a 60MP sensor at any aperture. But if you "zoom out" a bit and only use 80% of the sensor area, you get a pretty even grain pattern: center vs corners. That is before pixel shift.

With pixel shift enabled, you do get a sizeable, measureable, real world resolution boost, but it's tricky: your copy stand has to be absolutely rock solid. Even my dog running around causes enough micro-vibrations to throw it off. Heavy trucks on the street - same thing. Basically I have to wait until late at night before I can actually scan this way. And even with this Sigma lens the resolution in the center is higher than in the corners.

How practical is this resolution boost? For 35mm it's complete an utter overkill even for the fine-grained ISO 100 films. There's no additional detail to reveal, so I always turn it off when scanning 35mm. 60 "bayer megapixels" is more than enough. I downsample my scans to be 5,000 or 5,600 pixels on the wider side. This makes the grain look tight and crisp. I haven't seen any scans online that are higher quality than this method.

Now, when it comes to scanning medium format, pixel shift - again - doesn't help revealing any additional resolution on film. Maybe it would have for transparencies or other ultra fine-grained emulsions, but for HP5+, Fuji 400H Pro, and some FP4+ that I shoot in 6x6 on my Rollei, there's no extra resolution to be revealed. I tried pixel-shifting and also sticthing two shots - nope, no additional detail. 60 bayer megapixels is enough for one-shot. However, the grain quality goes up substantially when stiching or using pixel shift.

TLDR: pixel shift is useless for 35mm, but gives better grain quality for medium format. Pixel shift is tricky to use. Requires extremely high quality optics and impossibly stable shooting platform. No resolution boost vs bayer 60MP, but helps to reduce grain aliasing making grain look better for pixel peepers.

And finally, this needs to be added: I actually stopped using it entirely. The fragile nature of pixel shift slows down my scanning, while the benefit is only apparent when pixel peeping medium format negatives, which I don't do much.

 

[bags27]

Hey--thanks incredibly much for this. This is super helpful, because I already have the Sigma 70 macro in L mount and could always swap it for the longer 105 macro to combine with the Panny SiR. Since I have a massive copy stand (and I agree that subtle micro-vibrations defeats the best scanning, so, I, too, do my scanning at night), really good backlighting, etc, I'm thinking to give this a go with my MF negs.

BTW, not sure I agree with the virtue of autofocus. I used to do this with my Leica CL and the Sigma 70. Since going to the Fuji and Mamiya manual, I get better scans. Using the 1951 USAF resolution chart (a guy has made 35 mm negs of it) right on top of target negs allows me to get as good or better focus, I think, as AF. But yes, it is tedious. I just tell myself: it's not mindless scanning, it's macro photography. Sometimes I can even believe that.

I've given up with stitching. Just too much trouble, and it takes me away from cameras and into the computer. I'd rather do scanning where I'm working with a camera.

Thanks again...super helpful!!

 

[4season]

With pixel shift enabled, you do get a sizeable, measureable, real world resolution boost, but it's tricky: your copy stand has to be absolutely rock solid. Even my dog running around causes enough micro-vibrations to throw it off. Heavy trucks on the street - same thing. Basically I have to wait until late at night before I can actually scan this way. And even with this Sigma lens the resolution in the center is higher than in the corners.

Interesting!

Have you seen any benefits in using only electronic shutter, and switching off IBIS?

Room vibration / noise: With the aid of an audio analyzer, I discovered that I've got a surprising amount of <30 Hz energy in my room, though you'd never guess just by listening. Even with minimal traffic and quiet neighbors, it never goes away entirely.

Random Sony gripe: Not a particular fan of needing to use Imaging Edge software to generate the pixel-shifted image. Olympus does that in-camera! Love pretty much everything else about A7R4, though.

 

[4season]

I have a pretty good idea of how my legacy b&w images will look at super high resolutions: Too much @#$ camera shake "Back in the day" when I had a Fuji GSW690 III, here and there I'd get hints of what was possible, with detail galore. The other 98% had just enough motion blur to annoy the cr@p out of me, though when printed at normal sizes, they looked great. Hope you folks used a tripod more than I did!

 

[GLS]

Yes, I've recently switched to the Sony A7RIV and Sigma 70mm Art and routinely use one 16-shot pixel-shift multi shot to digitise my 120 and 4x5 film. I used to use a D810 and Tokina 100mm macro, and stitch together multiple frames at about 1:1.5 magnification.

Both methods give good results, but the 16-shot PSMS definitely has the edge (provided a good setup and techniques are used), and is so much faster and easier to do. I disagree with @McDiesel that there are no benefits to resolution using PSMS with 120 film vs a single 60 MP shot; there certainly are when using high resolution films. I have tested it with Agfa Copex Rapid, for example, and the difference is obvious.

If I was severely nit-picking with PSMS on the Sony, I would say it isn't perfect. I have observed that it can produce tiny artifacts on high contrast edges, which look a little bit like interlacing. You have to be viewing the 240 MP image at 100% (or even 200%) to see them though, and if the image is downsized slightly they disappear completely, so as I said it's a nit-pick, but worth mentioning. I suspect it may be at least partially due to micro-vibrations during the capture process. I already have a good Kaiser copy stand, but something beefier might help here. At this level of resolution the tiniest imperfection in technique is revealed.

Finally, the need to use the Sony software to output the PSMS file isn't ideal, and it can't output a RAW file (16-bit TIF is the best it offers). There is a third party program (PixelShift2DNG) which can produce a RAW file, but I found when using it that it otherwise didn't do as good a job as the Sony software (the files had noticeably less detail). Using a camera that handles all the processing for you would make workflow easier, but there are trade-offs elsewhere (resolution with the S1R, or cost with the GFX). But really, the extra few minutes to process each PSMS image with the Sony software isn't a big deal IMO.

 

[Raghu Kuvempunagar]

What happened to @Old Gregg? He had some interesting and informative posts on pixel shift for scanning. All those posts have vanished!

 

[bags27]

Yes, I've recently switched to the Sony A7RIV and Sigma 70mm Art and routinely use one 16-shot pixel-shift multi shot to digitise my 120 and 4x5 film. I used to use a D810 and Tokina 100mm macro, and stitch together multiple frames at about 1:1.5 magnification.

Both methods give good results, but the 16-shot PSMS definitely has the edge (provided a good setup and techniques are used), and is so much faster and easier to do. I disagree with @McDiesel that there are no benefits to resolution using PSMS with 120 film vs a single 60 MP shot; there certainly are when using high resolution films. I have tested it with Agfa Copex Rapid, for example, and the difference is obvious.

If I was severely nit-picking with PSMS on the Sony, I would say it isn't perfect. I have observed that it can produce tiny artifacts on high contrast edges, which look a little bit like interlacing. You have to be viewing the 240 MP image at 100% (or even 200%) to see them though, and if the image is downsized slightly they disappear completely, so as I said it's a nit-pick, but worth mentioning. I suspect it may be at least partially due to micro-vibrations during the capture process. I already have a good Kaiser copy stand, but something beefier might help here. At this level of resolution the tiniest imperfection in technique is revealed.

Finally, the need to use the Sony software to output the PSMS file isn't ideal, and it can't output a RAW file (16-bit TIF is the best it offers). There is a third party program (PixelShift2DNG) which can produce a RAW file, but I found when using it that it otherwise didn't do as good a job as the Sony software (the files had noticeably less detail). Using a camera that handles all the processing for you would make workflow easier, but there are trade-offs elsewhere (resolution with the S1R, or cost with the GFX). But really, the extra few minutes to process each PSMS image with the Sony software isn't a big deal IMO.

Thanks so much. Really helpful. Yes, I had heard reports of artefacts, too. But good to know they're entirely manageable.

An advantage of my using a Panasonic S1R is that exports in DNG and that it takes the L mount, which I use for my Leica digital equipment (though I no longer shoot digital). But it seems as if the long-awaited Sigma Foveon FF L mount camera may actually become a reality. If that is so, it opens up new possibilities.

 

[Les Sarile]

I currently use my Fuji GFX-R (MF 50 mps) with a Mamiya 625 f/4 120mm macro to scan my MF (6x6, 6x7, and 6x9) negatives. Works well. But always looking for better.

I read about a Large Format shooter who so liked his pixel-shifting Olympus micro 4/3 for scanning (I imagine he stitched several) that he actually got rid of his drum scanner. Wow! That's quite a claim.

I could swap my Fuji for a used pixel-shifting Panasonic S1R (FF, 47 mps) to pair with my Sigma L mount macro lens. I'm thinking of renting it first to compare. But just wondering if anyone's tried this technology for film scanning.

Of course, if that's great, the new pixel-shifting Fuji GFX 100-S might be even better, but that's not being discounted from $6k, and there has to be a limit to how effective larger sensors are going to be, right?

thanks!

Regarding the claim of getting rid of a drum scanner in favor of an Olympus, I suppose there may be many factors to consider not the least of which is which drum scanner and primary/secondary use of the DSLR.

I shot a 4 X 4 arrangement of 12233 rescharts with 35mm Kodak Techpan @ ISO25 processed in Kodak Technidol and scanned with a 14.6MP K20D, 36MP D800 and my Coolscan 4000dpi. The 100% crops of these are on the left. The big 100% crop on the right is about a 4.5X optical magnification and you can clearly see detail that these methods of scanning failed to achieve.

Resolution testing my SMC Pentax-M 50mm F4 macro lens by Les DMess, on Flickr

Be interesting to see how much more can be achieved by the newer pixel shifting technology.
I also just found out there is a Heidelberg Tango drum scanner service just down the street from me that can scan 35mm at 7250dpi providing a 10276 x 6850 image. Would be interesting to see that as well.

 

[Oren Grad]

Here are three captures from the same 35mm Tri-X negative. The top one is Nikon LS-9000, the middle one is GFX 100S with 75mm f/4 Apo-Rodagon-D 1:1 lens, and the bottom one is GFX 100S in pixel-shift mode with 75mm f/4 Apo-Rodagon-D 1:1 lens. For both of the GFX captures the Apo-Rodagon-D was set to f/5.6 marked aperture.

The GFX raw files were processed in Capture One with sharpening set to 0 and saved as TIFF files. All three TIFF files were opened in Picture Window Pro 8 for final processing. Minor curve adjustments were made to roughly match contrast and brightness to minimize discrepancies that might bias perception. The two smaller files were upressed to match the pixel-shift file using bicubic with sharpening set to 0. The crops here are presented at 100%.

 

[4season]

Here are three captures from the same 35mm Tri-X negative.

Do you have any similar test images taken using a higher resolution film such as Tmax? Although I see some differences in white balance, contrast, and the texture of the grain, actual image data being retrieved looks pretty much the same to me.

 

[Oren Grad]

Do you have any similar test images taken using a higher resolution film such as Tmax?

Afraid not, sorry.

 

[Les Sarile]

Do you have any similar test images taken using a higher resolution film such as Tmax? Although I see some differences in white balance, contrast, and the texture of the grain, actual image data being retrieved looks pretty much the same to me.

That's why in 35mm, I had to resort to Techpan as the scanning target as it had to exceed the scanning device. Of course before I did this test, I didn't know how much it was going to exceed it.

I am certain I also conducted this test with TMAX100 and Velvia so I will have to go through the archives.

 

[McDiesel]

My results were extremely similar to these, thank you @Oren Grad. No uptick in resolution, but more natural appearance of grain.

I disagree with @McDiesel that there are no benefits to resolution using PSMS with 120 film vs a single 60 MP shot; there certainly are when using high resolution films. I have tested it with Agfa Copex Rapid, for example, and the difference is obvious.

I mostly shoot HP5+, FP4+ and ISO400 color print films in medium format. I wouldn't rule out resolution benefits for finer grained films, but what Oren Grad shared above is what I've been seeing.

Also, speaking of micro-artifacts of Sony a7R IV. If you're shooting tethered, there's a setting in the Sony RAW converter for stitching pixel-shift frames, designed to reduce/eliminate these artifacts. I think it's sticky, i.e. once you dial in the setting that eliminates these stepped edges it will be reused by their tethered shooting app when it does automatic stitching.

Also, I forgot to mention that Sony has two pixel shift modes: 4-shot and 16-shot. The latter isn't worth the hassle in my case, but the 4-shot mode is more useful. It doesn't add a lot of scanning time, isn't as prone to micro-vibrations, and effectively cancels the artifacts caused by demosaicing of the bayer pattern. No extra resolution in my case, but noticeably better grain. That's what I use by default.

 

[4season]

That's why in 35mm, I had to resort to Techpan as the scanning target as it had to exceed the scanning device. Of course before I did this test, I didn't know how much it was going to exceed it.

I am certain I also conducted this test with TMAX100 and Velvia so I will have to go through the archives.

Well you've certainly made a point about what film is capable of under carefully controlled conditions. But even in the case of an extremely resolving film such as Technical Pan, I wonder if an inexpensive lens used for microfilm or microfiche might nevertheless easily out-resolve the film. Most obvious difference that I can see is that such optics aren't operating at 1:1 magnification, but ~15x-30x. Granted, this will likely require stitching multiple images together in order to capture the full frame, but unless someone's got a wideangle, low-power microscope, maybe that's just how it goes. Potentially a killer technique for Minox 8x11!

 

[4season]

Funny, but the more I think about achieving the ultimate high-res scanning solution, the more I think the answer lies in old-school technology from the era of microdots and the like.

 

[Andrew O'Neill]

What is pixel shifting?? Can someone point me to an article? Thank you.

 

[GLS]

Also, speaking of micro-artifacts of Sony a7R IV. If you're shooting tethered, there's a setting in the Sony RAW converter for stitching pixel-shift frames, designed to reduce/eliminate these artifacts

I know. I use a value of 0.5 for 16-shot images. If you overdo it the output becomes noticeably softened.

 

[Oren Grad]

More to chew on! Another set of copies of a 35mm Tri-X negative, this time using an A7RIV with Sigma Art 70mm Macro, set at f/4. From top to bottom, these are single-shot, four-shot and sixteen-shot captures. The multi-shot captures were converted to single DNG files using the LibRaw PixelShift2DNG utility. Raw files were converted in RawTherapee 5.8 and saved as TIFF files, then opened in Picture Window Pro 8 for final processing, which consisted solely of upressing the smaller files using bicubic with sharpening set to 0, and then cropping for presentation. Again, the crops are presented here at 100%.

 

[GLS]

What is pixel shifting?? Can someone point me to an article? Thank you.

It is where cameras with in-body-stabilisation systems on the sensor move the sensor around whilst taking multiple frames of the same scene, which are then blended together to create an image with increased resolution and/or colour fidelity (because the Bayer noise is eliminated).

Here is Sony's blurb about it:

 

[Les Sarile]

Well you've certainly made a point about what film is capable of under carefully controlled conditions. But even in the case of an extremely resolving film such as Technical Pan, I wonder if an inexpensive lens used for microfilm or microfiche might nevertheless easily out-resolve the film. Most obvious difference that I can see is that such optics aren't operating at 1:1 magnification, but ~15x-30x. Granted, this will likely require stitching multiple images together in order to capture the full frame, but unless someone's got a wideangle, low-power microscope, maybe that's just how it goes. Potentially a killer technique for Minox 8x11!

I used a used Pentax M 50mm f4 macro lens (bought used cheap) on a manual focus Pentax LX body to take that shot as I couldn't believe the super sharp results I was getting so I wanted to see just how good it is. At this point I still can't determine if the limitation is the film or the lens. I do know that digitizing methods I used are not capable of fully resolving the detail.