What is the best lens for 1x Camera-Scan?

[ColSebastianMoran]

A recent post asked "What's the very best lens for camera scan with my 50 MPx body?"

This got me thinking: What would define "very best?" Best cam-scan image quality when everything else is optimal? Good enough to give cam-scan results equal to the best? Best image quality given the rest of my setup? Best tradeoff of usability, fit to my setup, and cost?

Back story: Earlier in the digital age, I realized I could mount most any lens to my digital body and see quickly what it would do. I set out to buy a number of enlarging, movie, duplicating, macro, and legacy normal lenses plus macro gadgets. Back then, at 6MPx, half the lenses out-resolved the sensor. Things shifted at 12MPx, more at 24MPx, and now I have 50 MPx bodies.

At 50MPx, I have a handful of lenses that all perform at about the same highest level, all give ~80 lp/mm on the sensor (more than most any films). I'm sure that there are optical differences in these lenses, but I cannot see them at 50MPx and I doubt they would make a difference in cam-scan images.

So, what considerations for selecting a "best" lens? Here are some of my reflections after testing a lot of lenses and setups.

Tight Rig - You want your whole rig as close to rock-steady as possible. Any wiggle, looseness, any "play" will make it very hard to do manual focusing and can show movement during exposure. This suggests a lens with native mount to the body (no adapters) and a tripod-mount solidly fixed on the lens. You want a tight mount-camera-lens combination.

AF or Manual? - I have a really tight copy stand. I can adjust focus moving the whole camera-lens combination, find precise focus, pull my hand back and nothing moves at all. And, my Sony bodies have excellent focus-aids. With this rig, I prefer manual focusing, even with an AF lens. But, most copy stands aren't this tight. Everything shakes as you adjust position. MF by turning the focus ring jitters the image. Maybe your body doesn't have good enough focus-aids; the focus-confirmation dot on my Nikon bodies is not precise enough for cam-scan. Ditto a Canon 5DSR I've tried. For all these cases, I think you want AF.

So, with any wobble in the setup, I'm going to suggest a lens with native mount to body and AF. With a really tight copy stand and willing to do MF, then enlarging/copy lenses on tubes and adapters can work, better if tight.
Lighting? and shutter speed? - With any potential for vibration, flash is an easy recommendation but requires more setup. If your rig is really tight, and your body is free of shutter-slap, then ~1/3 sec on a light panel can work. Video lights give about 1/100th.

Film Flatness and aperture? - Film is not flat; the natural curve persists with any holder short of glass-mount. This affects lens choice. The highest resolving lenses are best at f/4. Lens in CoolScan 8000/9000 is f/2-something. Are you willing to do glass mount? Or focus stacking? If so, then a more exotic lens may can give you better results. If not, then I'll recommend f/8 to give better across-frame sharpness.

Resolution? - How much resolution do you want? How much detail is in the film? How big a print do you want to make? If you want ultimate resolution, you'll want to setup for 2x to a 50MPx body and stitch. This will give well over 100 lp/mm. The excellent test negative by Vladimir Serebryany has distinguishable lines to about 80 lp/mm in the center, 60 at corners, and that is with exotic film stock and a lot of care; this makes me think 60 lp/mm in cam-scan is plenty for most film images. Further how big do you want to print?

I'm very happy with 20x30" prints from my 24MPx bodies, even better prints from 42MPx, and processing gives further improvements ("Super Resolution" in ACR, Topaz DeNoise AI, etc.) If you choose a 24MPx body, you'll get fine cam-scan images, and the top dozen lenses in my tests all give the same result. At 50MPx, a half-dozen lenses all give the same top result at f/5.6 or f/8. There are more exotic lenses, but "Best" in my book, with my gear, doesn't go beyond these choices.

Alignment - This is a little nit for ease of use. I use a mirror for alignment. My script: Setup as normal, place a mirror on top of the film holder, move camera and lens back a bit so I can see the whole front of the lens, align until the lens is exactly in the middle of the frame, then move things down into position for 1x. (Thank you, Peter Krogh.) This requires refocusing the lens from 1x to something less. Here's the nit: I cannot do this with an enlarging/copy lens on tubes and adapter. Maybe with a focusing ext tube, but it's just less practical. Setting alignment in this way is far easier with a macro lens (with focusing helical) than with an adapted lens. Once again, this takes me to prefer a macro lens, even better one I can switch to AF.

Net for me: A high quality native-mount macro lens is my go-to choice for camera scanning. With my Sony bodies, that's the 70 f/2.8 Sigma Macro ART with a tripod mount ring (I use it in AF for alignment, then set for 1x and focus manually with the copy stand). Sony 90 is equal, even possibly better, at double the price; will need a tripod mount. I use flash to eliminate any issues of movement.

Also: On Nikon bodies, I use the 50 f/2.8 Sigma EX DG Macro in auto-focus.

A note aobut USAF target testing
Each Group number doubles resolution. Group 6 Element 1 is 64 lp/mm. G5-E1 is 32.
Six element changes in each group. Group 6 Element 4 is 1.41x G6-E1 or 91 lp/mm
Each element change is about 12%.
How much detail can film record? The lenses on our film cameras? For both, probably 50 lp/mm.
The Nikon Coolscan 35mm scanners are respected as high quality. They record 4000 pixels per inch which in practice results in about 60 lp/mm. At this resollution, most film grain shows clearly.

Here's the USAF target at 1x with two of my "best" lenses.
Test conditions: 1x to a 50MPx sensor, well off-axis to the mid-corner of FF sensor (extreme corner on APS). Best aperture. Shown at 400% screen grab.
This result is typical of my half-dozen best lenses.

My "Best" lenses Resolving G6-E4 at extinction 91 lp/mm to 50MPx 3.8µ sensor:
50 f/2.8 Sigma EX DG macro
50 f/4.5 Tominon
55 f/2.8 Micro-Nikkor AI or AIS
70 f/2.8 Sigma Macro ART
75 f/4 APO Rodagon-D 1x Copy Lens
75 f/4.5 APO Rodagon-D 2x Copy Lens, Normal Orientation
90 f/2.8 Sony Macro
100 f/2.x CoolScan 8000/9000 lens

 

[ColSebastianMoran]

Group 2 in my tests, all still excellent, but slightly less contrast or Resolution
Resolving G6-E3 81 lp/mm:
35 f/4.5 Tominon
60 f/2.8 Micro-Nikkor AF
75 f/4.5 Enlarging Ektar
80 f/4 Olympus Bellows Macro
80 f/4 APO Rodagon-N
80 f/5.6 Rodagon Fixed Aperture
80 f/5.6 Componon
80 f/4 Componon-S
90 f/2.5 Vivitar Macro w/1x Extender
105 f/2.8 Micro-Nikkor AF-D

All these lenses produce test result about like this. Show is the 80 f/4 APO Rodagon-N enlarging lens, compared here to the 70 f/2.8 Sigma Macro ART, one of my "best" group. You can see the difference in contrast and resolution in the test chart, but I doubt you'll be able to see a difference in a print.

 

[ColSebastianMoran]

Two more groups in my tests, but first please comment all this.
- Are these good tests?
- Do I have the right points about practical best for camera-scanning?
- What differences have you found in lenses at 1x?

 

[Oren Grad]

Informative and helpful posts - thanks for taking the time!

Re lens tests, these may be of interest:

https://www.closeuphotography.com/1x-test-2020

http://www.coinimaging.com/Lens_tests.html

A few observations from my own practice:

For digitizing 35mm and medium format negatives I'm currently using a copy stand setup with a Sony A7RIV, the Sigma Art 70mm macro, a glass sandwich negative carrier borrowed from my darkroom, and a Kaiser light panel. This replaces the Nikon 5000 and 9000 scanners that I previously used.

My copy stand is rigid enough to make manual focus in magnified live view practical, and that's what I usually do. I make the exposures on self timer with a few seconds delay, using electronic first curtain shutter.

I have a set of the Apo-Rodagon-D lenses, and tested the 75mm 1:1 on a Canon EOS 5DsR using an adapted bellows. But I found it difficult to get consistently optimal results because of a lack of finesse in the focus action of the bellows I was using as well as the limitations of the 5DsR's live view.

I have the Canon EF 50mm f/2.5 Compact Macro and have used it on various Canon bodies for copying prints and album pages, but I don't have the 1:1 adapter and have not tested it for film copying. My guess based on my print work is that it won't be anywhere near the same league as the Sigma Art 70.

As for how much resolution I want, the answer is: enough to faithfully render the grain character of Tri-X. By that standard, for 35mm film copied at 1:1, the A7RIV with the Sigma Art 70, even just in single-shot mode, is clearly superior to the Nikon scanners I used to own. If and when I can get still more resolution at a price I can afford, I will want it, for further improved quality both with a range of 35mm films and in copying larger roll film formats.

I haven't used resolution targets for testing - the character of film grain rendering is a highly sensitive qualitative indicator that tells me what I need to know for my purposes.

Looking forward to the possibility that I might be interested in upgrading to the high-resolution EOS R body when it finally arrives, I bought the Sigma lens in Canon EF mount and use it on the Sony with a Sigma MC-11 mount converter. I already have the Canon EF-RF converter.

 

[Freakscene]

I don’t really do any colour work, and when I do I still have access to a drum scanner.

For my own monochrome work I found moving to a system where focus was checked really properly once (by laser interference) and the system fixed in focus, then using a single wavelength light source with a matched lens made a huge amount more difference than messing around with lenses because of their resolution and contrast transmission.

If you can find and have a system that can use an APO El-Nikkor, however, you will really discover a world of difference.

Marty

 

[keytarjunkie]

If you are able to, I would recommend adding the new Sigma 105mm f/2.8 DG DN macro lens. It's a better performer than the Sigma 70mm, and even has a very slight edge over the Sony 90mm macro. I'm amazed at the sharpness I get down to the corners on a 35mm negative. The downside is some pincushion distortion, but it is not difficult to correct. At first I was using an old Nikkor 55mm f/3.5, one of my favorite lenses to shoot with but it really showed its age when put to use in this way.

 

[ColSebastianMoran]

... As for how much resolution I want, the answer is: enough to faithfully render the grain character of Tri-X. By that standard, for 35mm film copied at 1:1, the A7RIV with the Sigma Art 70, even just in single-shot mode, is clearly superior to the Nikon scanners I used to own. ...

Thanks, Oren, for all the good comments.

Yes, with a ~50MPx body (your A7Riv) we clearly out-resolve the 4000 ppi film scanners.

Have you looked at gain in your negatives at higher magnifications? I've done higher magnification shots simulating 100MPx and 400MPx. If you want to show the film grain, there's more to be seen, and seen more clearly, at higher MPx.

For my photography, I am usually working with noise reduction tools to smooth skies and flat areas, reducing grain, so it's a different mode of working.

For what it's worth, I find a good 24MPx cam-scan prints well at the sizes I want to use. Not sure there's more image detail to be found at higher MPx for 35mm.

 

[ColSebastianMoran]

For my own monochrome work I found moving to a system where focus was checked really properly once (by laser interference) and the system fixed in focus, then using a single wavelength light source with a matched lens made a huge amount more difference than messing around with lenses because of their resolution and contrast transmission.

If you can find and have a system that can use an APO El-Nikkor, however, you will really discover a world of difference.

Marty, wow. That's very interesting. Instead of broad spectrum, use a single color of light (easy to do these days with LEDs) for B&W. Can you suggest a color and a lens suited to that color?

And, APO El-Nikkor? Had to Google this one, but looks to be in the exotic price range with the printing Nikkors. What differences have you seen with these?

 

[Michalm]

I would say film scanner lenses that cover full frame image circle are likely going to be best suited for film scanning , being apo corrected and optimized to work around 1:1 have little or no distortion , even distribution of sharpness . Difficulty is how to mount them properly , and they typically have no iris and are around f2.5 . How about film holder with ANR glass to hold it flat? I would probably try to wet mount film especially scratched one .

 

[ColSebastianMoran]

If you are able to, I would recommend adding the new Sigma 105mm f/2.8 DG DN macro lens. It's a better performer than the Sigma 70mm, and even has a very slight edge over the Sony 90mm macro. I'm amazed at the sharpness I get down to the corners on a 35mm negative. The downside is some pincushion distortion, but it is not difficult to correct. At first I was using an old Nikkor 55mm f/3.5, one of my favorite lenses to shoot with but it really showed its age when put to use in this way.

Thanks for the comment on Sigma 105 DG DN. Sounds good, and Sigma certainly has a track record for good macro lenses. But, no, I don't have that one. Any chance you can show a comparison to 70 Sigma or the 90 Sony?

 

[ColSebastianMoran]

I encourage everyone not to worry if your lens looks tops or further down with the USAF glass target. It's a tough test and shows differences that might not show up in real images. Here's an example: Cam-scan of a good chrome.
- Left, 75 f/4 APO RodagonD 1x, one of my "best"
- Right, 75 f/4.5 Componar, well down the list.
- Sony A7, 24MPx at 1x, magnified screen grab @ 100%, corner.
Can you see differences? Yes
Will they be apparent in reasonable sized prints?

 

[keytarjunkie]

Thanks for the comment on Sigma 105 DG DN. Sounds good, and Sigma certainly has a track record for good macro lenses. But, no, I don't have that one. Any chance you can show a comparison to 70 Sigma or the 90 Sony?

There's a million tests on youtube from the talking heads. The Sigma offers higher center sharpness wide open than the Sony, and the corners look a little better too. At f/8 I doubt there would be much difference. But, the Sigma 105 is a great lens and I would recommend it. The barrel doesn't extend like the 70mm.

 

[Oren Grad]

Have you looked at gain in your negatives at higher magnifications? I've done higher magnification shots simulating 100MPx and 400MPx. If you want to show the film grain, there's more to be seen, and seen more clearly, at higher MPx.

I've tested the multi-shot modes of the A7RIV. For 35mm Tri-X they achieve a modest improvement at the cost of massive file-handling overhead. So normally I stick to single-shot mode. For 6x9 cm TX negatives they roughly match what the Nikon 9000 was able to do, but again with the file-handling burden. But since setups based on 100-150 MP medium format cameras are currently out of reach for me, that will have to do for now.

 

[Michalm]

There's a million tests on youtube from the talking heads. The Sigma offers higher center sharpness wide open than the Sony, and the corners look a little better too. At f/8 I doubt there would be much difference. But, the Sigma 105 is a great lens and I would recommend it. The barrel doesn't extend like the 70mm.

I think the trick is to find flat field field lens that is sharp at around f4 -f5.6 when using 42mp full frame or 24mp apsc camera because diffraction kicks in when stopping lens down furher .

 

[Michalm]

I encourage everyone not to worry if your lens looks tops or further down with the USAF glass target. It's a tough test and shows differences that might not show up in real images. Here's an example: Cam-scan of a good chrome.
- Left, 75 f/4 APO RodagonD 1x, one of my "best"
- Right, 75 f/4.5 Componar, well down the list.
- Sony A7, 24MPx at 1x, magnified screen grab @ 100%, corner.
Can you see differences? Yes
Will they be apparent in reasonable sized prints?

Indeed looking at this example it is hard to tell the difference. but there is other things to consider like sharpness consistency across the frame chromatic abberation , distortion. And depending on pixel density of any given camera sensor some lenses should be stopped down less than others.

 

[Freakscene]

Marty, wow. That's very interesting. Instead of broad spectrum, use a single color of light (easy to do these days with LEDs) for B&W. Can you suggest a color and a lens suited to that color?

I used a Zeiss S-Planar 50mm f/1,6, filter, and a 4360 A° Light source. Under these conditions it resolves 1400lp/mm, albeit on a high contrast target. It was phenomenal, but the lens can’t really cover a 35mm sensor in a single frame, because the image circle is 14.5mm.

Have you looked at gain in your negatives at higher magnifications? I've done higher magnification shots simulating 100MPx and 400MPx. If you want to show the film grain, there's more to be seen, and seen more clearly, at higher MPx.

I also thought this, and it led to my final approach of using a Zeiss microscope to ‘scan’ including stacking digital images through the film emulsion, but this took half a day to do one photo - I ended up with a 750mb file stitched from 6 stacks of 3 images each. It was the ultimate in both senses.

And, APO El-Nikkor? Had to Google this one, but looks to be in the exotic price range with the printing Nikkors. What differences have you seen with these?

Better colour, resolution, field flatness. Pretty much everything - they are by far the best enlarging lenses that were ever made, by a long, long shot. And now that darkrooms are archaic and other fields that were optical such as semiconductor manufacture are moving to other methods, it is unlikely these types of lenses will ever be bettered.

Marty

 

[Shac]

My choice is different - I use an Olympus Auto 1:1 Macro MC 80/4 (bellows) at f8 with a Sony A7R3. After using a Nikon Coolscan 4000 for years and then testing various macro lenses (e.g. Leica R60, Nikon F 50/3.5), I found the corner sharpness with the Olympus blew all these others away

 

[Greyscale]

I use an enlarger lens on a bellows and a focus rail.

 

[ColSebastianMoran]

My choice is different - I use an Olympus Auto 1:1 Macro MC 80/4 (bellows) at f8 with a Sony A7R3. After using a Nikon Coolscan 4000 for years and then testing various macro lenses (e.g. Leica R60, Nikon F 50/3.5), I found the corner sharpness with the Olympus blew all these others away

Olympus 80 f/4 Bellow macro is terrific. One of the few lenses actually designed for 1x copying.

 

[ColSebastianMoran]

I use an enlarger lens on a bellows and a focus rail.

Yes, in my experimentation with about 40 lenses, many enlarging lenses are quite good at 1x, most are quite good at 0.67x. Some are not.