SL2 Pixel Shift for Scanning thoughts

[pgh]

Advertisement (gone after registration)

Has anyone used the pixel shift on the SL2 for "scanning" 35mm or 120 negatives, specifically for the purposes of making large prints? 

I'm not really seeing, and not expecting to see much difference, I actually feel like it looks a touch worse to be honest, but that's probably a subjective reading, given that I'm essentially just trying to resolve grain at this point. I'm limited by my printer though, 22" wide.

I'm curious if anyone has found better results making a 30" or 40" or even 50" print from a 35mm negative using this pixel shift for the original capture - do you feel the prints hold up better than interpolating a single capture scan? Assuming the same well exposed negative.

As an aside, having just embarked on re-digitizing my film archive I'm quite pleased with this method, much better than the previous home method of flatbed scanning and honestly it comes quite close to the few negatives I have drum scanned, which was such a pain. This whole method of digitizing film makes shooting film feel viable again...that's another problem for another time. 

[LocalHero1953]

I used it when I first started scanning 35mm with the SL2-S. But it just produced bigger files with film grain spreading across more pixels. It didn't give me any better quality scans.

It worked a bit better with large format (4x5), but for the sake of smaller files, I now keep it for special scans where I want the best image for printing.

[pgh]

@LocalHero1953 Yea that’s the impression I’m getting. 
Have you seen an advantage with larger prints? Thats better than just interpolation?

[LocalHero1953]

I haven’t printed bigger than A2. I can’t be certain, but I don’t think it would make much difference. Grain is dominant over pixels for 35mm. 

[Stuart Richardson]

I think the issue is that you are gaining some detail, but also losing acutance. So it depends on your original. I think it is likely also limited by diffraction and the lens, and certainly on vibration. I found that wooden tripod I use for 8x10 is more stable than even my studio stand, which must way more than 50kg. It is better at absorbing vibration and is not as cantilevered out, so there is opportunity for flexion. At these levels of detail and magnification, the slightest vibration can make a difference, even from nearby traffic or machines in a building, such as ventilation. In my experience the camera based scanning with the SL2 and 105mm Sigma Macro is very close to the Epson V850 for 6x7. It has a bit less size, but a bit higher acutance. The Epson will outdo it for larger formats without stitching. I think the SL2 would certainly outpace it for 35mm. I did this test at a time when my lab's X5 motherboard died, and I was without the scanner for weeks/months, and I wanted to see what was better. Basically, it's my sense that the scanners tend to have the better workflows, better freedom from flare, while the cameras have better noise characteristics, more file flexibility, and higher peak sharpness or acutance. Now I have my X5 working again, I tend to stick to that because the workflow is a lot better (no glass, film flatness, don't need to worry about dust as much, and just overall very high quality results). But I have a lab and bought the X5 in 2010. If I were to spend the same money now, it would be on a Phase One or GFX based film scanning setup. Or honestly maybe even an IQ smart, but that is more because I mostly shoot 4x5 and 8x10 and those are better handled by scanners unless you have an institutional level budget to get a stitching Digital Transitions cultural heritage setup. But I think those start at at least 70k USD. 

In order: SL2 with Sigma 105mm Macro, a crop of that, then the Epson scan, then a crop of that. Both from an E6 transparency. I tried to color balance them pretty closely, but did not go crazy with it. 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited July 23 by Stuart Richardson

[Stuart Richardson]

P.S. If you look at the lower left corner, you will see the Epson has a bit more contrast. That is a flare issue. I think masking out all excess illumination and being in a very dark environment is pretty critical for camera scanning, and it basically shows up as loss of contrast. 

[Crem]

Advertisement (gone after registration)

2 hours ago, pgh said:

Has anyone used the pixel shift on the SL2 for "scanning" 35mm or 120 negatives, specifically for the purposes of making large prints? 

I'm not really seeing, and not expecting to see much difference, I actually feel like it looks a touch worse to be honest, but that's probably a subjective reading, given that I'm essentially just trying to resolve grain at this point. I'm limited by my printer though, 22" wide.

I'm curious if anyone has found better results making a 30" or 40" or even 50" print from a 35mm negative using this pixel shift for the original capture - do you feel the prints hold up better than interpolating a single capture scan? Assuming the same well exposed negative.

As an aside, having just embarked on re-digitizing my film archive I'm quite pleased with this method, much better than the previous home method of flatbed scanning and honestly it comes quite close to the few negatives I have drum scanned, which was such a pain. This whole method of digitizing film makes shooting film feel viable again...that's another problem for another time. 

I can't comment on the SL2, but for a while I did try the SL2-S with pixel shift for camera scanning 35mm film. Personally I think there was a slight improvement in color (non-scientific test), but it was minor compared to the workflow slowdown and huge files sizes. So I quit bothering with it. If I was scanning medium format or larger then I would definitely use it. The largest I print 35mm is 13"x19" (Canon Pro 310), but I like grain and I'm not pixel peeping film scans. I really should run a test comparing the same print with and without pixel shift.

Edited July 23 by Crem

[LocalHero1953]

45 minutes ago, Stuart Richardson said:

P.S. If you look at the lower left corner, you will see the Epson has a bit more contrast. That is a flare issue. I think masking out all excess illumination and being in a very dark environment is pretty critical for camera scanning, and it basically shows up as loss of contrast. 

I use the SL2-S and 24-90 at 90 for scanning 4x5 sheets - conveniently the sheet fills the frame at that FL, and since my 24-90 zoom creeps when pointed downwards, it eliminates that problem😁. I suspect there's a better 90mm for IQ - but I don't have one. I agree about flare though: I now do it in a darkened room after some bad experiences. For 35mm the Valoi Easy35 effectively keeps external light out.

Edited July 23 by LocalHero1953

[pgh]

1 hour ago, Stuart Richardson said:

I think the issue is that you are gaining some detail, but also losing acutance. So it depends on your original. I think it is likely also limited by diffraction and the lens, and certainly on vibration. I found that wooden tripod I use for 8x10 is more stable than even my studio stand, which must way more than 50kg. It is better at absorbing vibration and is not as cantilevered out, so there is opportunity for flexion. At these levels of detail and magnification, the slightest vibration can make a difference, even from nearby traffic or machines in a building, such as ventilation. In my experience the camera based scanning with the SL2 and 105mm Sigma Macro is very close to the Epson V850 for 6x7. It has a bit less size, but a bit higher acutance. The Epson will outdo it for larger formats without stitching. I think the SL2 would certainly outpace it for 35mm. I did this test at a time when my lab's X5 motherboard died, and I was without the scanner for weeks/months, and I wanted to see what was better. Basically, it's my sense that the scanners tend to have the better workflows, better freedom from flare, while the cameras have better noise characteristics, more file flexibility, and higher peak sharpness or acutance. Now I have my X5 working again, I tend to stick to that because the workflow is a lot better (no glass, film flatness, don't need to worry about dust as much, and just overall very high quality results). But I have a lab and bought the X5 in 2010. If I were to spend the same money now, it would be on a Phase One or GFX based film scanning setup. Or honestly maybe even an IQ smart, but that is more because I mostly shoot 4x5 and 8x10 and those are better handled by scanners unless you have an institutional level budget to get a stitching Digital Transitions cultural heritage setup. But I think those start at at least 70k USD. 

In order: SL2 with Sigma 105mm Macro, a crop of that, then the Epson scan, then a crop of that. Both from an E6 transparency. I tried to color balance them pretty closely, but did not go crazy with it. 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Thanks for this. 

To me, despite the slightly lower contrast the camera scan is evidently better. Epson has always had what I would call a sort of gummy look to the scans, more evident in 35mm color negative, that really is a hindrance to realizing the image in a satisfactory way. It was so frustrating it was part of the reason I stopped shooting film a decade ago. (I wasn't in a spot to afford more scanners and processing was hard to come by where I'd moved to as well). 

I still have the old V700, which for 35mm I have to say is pretty poor in comparison to what I am getting from my SL2 (hence redigitizing my archive). I'm using the Valoi Easy 35, Sigma 70mm macro and far as I can tell, does quite a good job keep the film relatively flat and flare out.

The biggest issue I had with the Epson though was color, I never could get good color out of it no matter what I did - the SL2 solves that problem and gives me more resolution on a 35mm negative. I'm making A2 size prints that I am quite pleased with, but am curious about going bigger. Once I finish my 35mm archives I'll dive into my (larger!) 6x6 archives and pick up the Valoi easy 120 for that. Then again, I rarely shot transparency, and when I did the Epson (or I) didn't have as much a problem with color. 

I'm not much of a pixel peeper in one sense, I know I'm quite at the limit of 35mm with an A2 print already, but I've seen some large prints from 35mm (obviously with evident grain) that I thought looked pretty beautiful in their own way - Lieko Shiga comes to mind. I don't have access to, and no interest in an X5 nor a drum scanner - those things made my life shooting film quite unpleasant - but I'm probably what you would call more of a volume shooter, even when I shot 120, and like low res scans at least from which to edit. Anyways, probably a few negatives I'd like to print large, knowing they will show grain it would still be good to make the print as good as possible given the limitations, perhaps I'll have to try a comparison with the pixel shift. 

 

[Stuart Richardson]

For 35mm I don't think there is anything that the SL2 with the right lens and right light source can't reproduce. As for a lens, the Sigma 105mm Art lens is a 1 to 1 APO macro lens that is as good as the 120mm APO Macro Summarit S. In this test I compared them and there was no difference. The Sigma is smaller, lighter and cheaper and has a native 1 to 1 versus 1 to 2. If you really want to improve the scanning, you can give it a go. 

I am with you though, the Epsons have always been unacceptable quality to me. The only reason I bought one was to scan 8x10, which I can't do in the X5. Here it can do better than the SL2 can do in single shot. I don't do much color negative, so I have not had to fuss much with the color. When I do, I will usually try Negative Lab Pro or Color Perfect. But in general I have always liked slides better. I have found it to be fine for that, but it does require tweaking. For 35mm I think any good digital camera scanning setup is likely to best it, as you will hopefully have a much sharper lens that will be able to focus perfectly on the film. I have been very frustrated with Epson for a long time, because I feel like they could have made those scanners really good, but instead they are just mediocre. I think it would have been so much better if they just made the lens a better lens that focused, maybe did XY stitching and made some more solid negative holders. Even if they doubled or tripled the price it would have been such a better deal, and I honestly think they would sell better. As it is now, I feel like there are disposable utensils with better quality plastic than what they use in the negative carriers. Everything about the scanner oozes cheapness, which is a shame, because I think if it were just fixed in a few ways it could be a really great, reliable tool for a huge swath of film photographers. 

[Stuart Richardson]

4 hours ago, LocalHero1953 said:

I use the SL2-S and 24-90 at 90 for scanning 4x5 sheets - conveniently the sheet fills the frame at that FL, and since my 24-90 zoom creeps when pointed downwards, it eliminates that problem😁. I suspect there's a better 90mm for IQ - but I don't have one. I agree about flare though: I now do it in a darkened room after some bad experiences. For 35mm the Valoi Easy35 effectively keeps external light out.

I think it is all about expectation. You are not going to get everything that is in a 4x5 out of it with a single shot of an SL2. The mismatch in aspect ratio doesn't really help either, as you are losing quite a bit at either edge. There is more than about 35mp of detail in most 4x5 photos, and multishot is also a bit of a compromise. The rule of thumb that I have found is that there is about 4000-5000dpi of usable detail in most films. Beyond that you are mostly enlarging the grain. But 4000-5000dpi from 4x5 is a pretty huge file. The X5 maxes out at about 2040dpi for 4x5 and that is roughly 75mp. That is good enough for sharp and lovely 100x80cm prints, and you can easily get away with 100x125 prints as well. But if you really want to get everything, a 4000-5000dpi drum scan or Creo/Kodak scan will get what you need. With the SL2 I think you can likely print a lovely 60x70 from 4x5 or even more depending on your preference and the quality of the scan. Ultimately it is always a matter of taste and discretion, but I would say that we are still a long way from digital cameras that can get all the detail and tone out of 4x5 or 8x10 in a single shot or even multishot. But when it comes to 35mm, any digital camera at 24mp or more with a good 1 to 1 macro lens should be able to get nearly everything out of it. 

I should specify too that what I am talking about is printing large photographs for museums and galleries, because that is what I do here. If it is a less demanding context, then of course the requirements become a lot more relaxed as well. 

Edited July 23 by Stuart Richardson

[misteracng]

I have been scanning 35mm with the sl2-s, multi shot is not really making a difference at this negative size, 24mp is plenty I also stay at f8, I feel f11 already shows diffraction and the scans are softer. Scanning medium format is different there is a clear improvement in detail with multishot. I have scanned 6x6 and 6x9 and it was better then a 60mp m11 monochrom but you need to zoom 100-200% to see it. So the SL2 multishot will be incrementally better but technique will really matter. No movement and turn of motion artifact compensation. Also I think it’s best to refocus after every frame. Any slight misfocus will negate any gains. Also the film will probably make a difference as well. To get the benefit of the sl2 multishot you will need very fine grain film, anything 400 speed is not going to show the difference, tmax 100, pan f 50, fp4 you can see the difference.

[Jon Warwick]

I find the difference in multishot (vs single shot) to be more apparent on the lower resolution 24mp SL2S than I did on the 47mp SL2. There is typically more aliasing with lower megapixel cameras, and multishot's goal is to eliminate aliasing - that in itself explains why there is more apparent fine detail in the multishot image. Still, to a lesser extent, aliasing in single shot certainly exists at 60mp and 100mp too. Test shots of (say) lettering on the back of a drink can that's about 15 feet away showed me the results of the different techniques ..... aliasing was quite apparent with a single-shot 100mp sensor, vs multishot on the SL2S that obliterated the aliasing. As an aside, a Monochrom camera like the M11M also has no obvious aliasing due to lack of color filter array.

I haven't yet tried stopping down to f11 or so, but i'd assume a bit of diffraction would eliminate the aliasing too on the 60 or 100mp cameras?? It's not clear if one could then sharpen up the image in post production, and get the best of both worlds (much less aliasing in single shot, but mitigate the diffraction with sharpening)?? Would that work, or has the fine detail been lost for good due to diffraction?

Edited July 24 by Jon Warwick

[Stuart Richardson]

I made a mistake once while film scanning with the Leica S006 that changed up my technique quite a bit. Basically, I was scanning an 8x10 for a client to make a modest print. I think 40x50cm. I was using my copy stand and the 120mm APO Macro Summarit S. I was shooting at f8 or f6.8 I believe. But for one shot I forgot to turn it from P mode to M mode, and the camera chose f2.5, which is wide open for that lens. That is a very very good lens, and lo and behold, the picture was sharper at 2.5 than it was at f8. Not just the center, but the corners too. 

There were two things going on: The 120mm APO Macro is not quite diffraction limited, but it is close, and the optical performance was better closer to wide open than at f8. But it was also the increased shutter speed. With the S camera, even with mirror pre-release and a studio stand, the release of the shutter causes enough vibration to affect your sharpness. 

So these days I am more likely to try to spend the setup time getting perfect alignment and then stop the lens down to f4 or so, rather than f8. If you are not using a dedicated macro lens, then you might need to stop down a bit more. I also highly recommend using the e-shutter and the countdown timer, if you are not already. I am not sure if it was clear in my first post, but I was using multishot for my film scans, and it does make a difference for larger negatives, but you do kind of trade resolution for acutance. The files get bigger and you get a bit more detail, but the detail is a little softer than it is in single shot. I am not sure how much of this is just magnification, but I have a feeling that more of it is introduced by the camera than by the lens, as least with the APO Summicrons and Sigma macro. 

[LocalHero1953]

I use the electronic shutter and 2 sec delay, and f/8. One thing I have yet to do is check the corner performance of the 24-90SL at 90, and its field curvature.

[Jon Warwick]

4 minutes ago, LocalHero1953 said:

….. its field curvature.

This is the issue that I have on other systems, particularly the M where I find many M lenses have field curvature.

Rather than my current approach of just stopping all the way down to f8 (even f11), I’m not sure what else to do to offset field curvature? …..is there a way to focus with an EVF to optimally mitigate field curvature at a more appropriate f stop (like 5.6)?

If I have a distant landscape, like a range of mountains that’s a planar subject to me, it’s often at infinity anyhow, so I’m not sure what else can be done but stop down to diffraction-inducing f stops that at least remove the field curvature?

[Stuart Richardson]

Yes, that is why they made macro and repro lenses flat of field. That is their job. In M mount my guess is your best choice would be something like the 90mm Macro Elmar or getting an R adapter and using the 100mm APO Macro Elmarit R with Elpro or 60mm Macro. But again, this is just to be optimal. I am sure the balance of stopping down a bit is also an option. Film scanning is a pretty specific application and it does benefit from equipment optimized for that task. 

Edited July 24 by Stuart Richardson

[Stuart Richardson]

18 minutes ago, Jon Warwick said:

This is the issue that I have on other systems, particularly the M where I find many M lenses have field curvature.

Rather than my current approach of just stopping all the way down to f8 (even f11), I’m not sure what else to do to offset field curvature? …..is there a way to focus with an EVF to optimally mitigate field curvature at a more appropriate f stop (like 5.6)?

If I have a distant landscape, like a range of mountains that’s a planar subject to me, it’s often at infinity anyhow, so I’m not sure what else can be done but stop down to diffraction-inducing f stops that at least remove the field curvature?

I struggled with this with the 70mm Summarit in the S system, as that kind of composition is so common here. I found that I had to stop down to 6.8 to get it to be pretty even, but f8 and f11 were more even, but also less sharp. My solution was to live with it and also rely more on the 120mm Macro for distant landscapes as it was much flatter of field, or the 45mm, which had more DOF and was also a bit less pronounced. One of the reasons I switched to the SL2 is that the APO Summicrons are just so good in this way. They just don't have to be compensated for in this way, and for me it was really freeing. Now I can shoot at f2 in low light and have infinity be sharp edge to edge. You have to watch DOF, but if the subject is contained within it, you are fine. I use the 50mm APO Summicron for copy work because I found it is equally sharp and flat of field as the two macro lenses I have, so if the reproduction ratio favors it, there is no reason not to use it. 

Edited July 24 by Stuart Richardson

[Jon Warwick]

37 minutes ago, Stuart Richardson said:

I struggled with this with the 70mm Summarit in the S system, as that kind of composition is so common here. I found that I had to stop down to 6.8 to get it to be pretty even, but f8 and f11 were more even, but also less sharp. My solution was to live with it ….

One of the reasons I switched to the SL2 is that the APO Summicrons are just so good in this way.

When you say your solution was to live with it, are you referring to stopping down to f8/11 to reduce the field curvature, even if diffraction was kicking in? Or you lived with the field curvature?

fwiw, I’m finding I’m probably happier losing some sharpness to diffraction (if anything, I see diffraction as more sacrificing a little acutance, rather than losing a ton of fine detail) …..but at least stopping down to f8/11 gives me a more even focus across the frame.

agree re the SL APOs. When I road tested  an SL3, it was immaculate across the frame at f5.6 with an SL 50 APO, really perfectly even in terms of centre and edge sharpness.

 

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Stuart Richardson]

I mean my solution was that I rarely used the 70mm wider open than 6.8. I used tripods or other lenses to make that happen. The only time I didn't was for certain portraits. It was of course very sharp on center, but for my landscape work it was not a 2.5 lens, but a 6.8 lens.