How Great a Lens do we Need for Camera-Scanning?

[astrosecret]

the sigma does a really nice job. probably going to sell my canon 100mm usm for it. the canon doesn’t hold a candle to my coolscan 5000. the center is fairly sharp
maybe even sharper but resolution drop off quickly

 

[ColSebastianMoran]

I’m impressed with your results at f8 and even f11 considering that lens resolution drops when stepped down.

On printing. I use an Epson ET8550 having had enough of the troublesome and expensive P800. A 8x10 or A4 printed at 300dpi would barely need a 8MP file, and I’ve had surprising results with old 3 to 5Mp files. So these results are really encouraging for low tech scanning using readily available and cheap lenses. Seems that 5-18Mp files will do nicely for run of the mill printing.

Great topic!

Thanks. This was interesting to do, and happy that it's of interest.

Optics theory about diffraction: Extinction resolution in lp/mm from diffraction will be at 1600/f-eff where f-eff is the effective f-stop. At 1x, the effective f-stop is double the number set on the lens. So, f/8 set on lens is f/16 effective and the resolution is limited to 100 lp/mm. The sensor also limits resolution; a 50MPx sensor has an absolute resolution limit ("Nyquist limit") of about 125 lp/mm, practical limit is somewhat less.

I read resolution of 91 lp/mm and some detail a little finer with the 70 Sigma. The results look about the same from f/4 to f/8, so I think f/8 is best for camera-scanning to give a bit of DOF to handle film not perfectly flat.

My point is that while better lenses will produce better files, I think we get pretty good cam-scans with lenses down to Group 3. My example lens, the 75 Vivitar Flat Field Lens is between group 3 and 4 in resolution.

 

[ColSebastianMoran]

Your work really is helpful. I own several of the lenses tested including the 55 f/2.8 Micro Nikkor and the Vivitar 75 f/3.5. Was planning on using it and a Sony nex-C3 (16mp) to copy some old slides for small enlargements.
. Looks like almost "any old lens will do."

While I won't agree that "most any old lens will do"... I suspect at 16MPx, you won't see much difference between Group 3 and Group 1 lenses.

 

[ranger9]

I would say "flatness" (of field) is more importance than "greatness" (of sharpness) as beyond a certain point, all you're accomplishing with a sharper lens is generating a more and more detailed image of grain edges. My guess is that your cheapo lens that goes soft in the corners simply lacks field flatness. Sometimes you can get a compromise by focusing halfway or 2/3 of the way out, but of course that's fiddly.

 

[ColSebastianMoran]

I would say "flatness" (of field) is more importance than "greatness" (of sharpness) as beyond a certain point, all you're accomplishing with a sharper lens is generating a more and more detailed image of grain edges. My guess is that your cheapo lens that goes soft in the corners simply lacks field flatness. Sometimes you can get a compromise by focusing halfway or 2/3 of the way out, but of course that's fiddly.

May be true, especially for lenses designed for general use, not copying/enlarging. Note the Vivitar was labeled "Flat Field Lens."

Flatness of field and flatness of the film are both issues. I've seen a lot of film flatness issues. Shooting at f/8 helps both, but my best lenses are sharp across the frame (verified with Vlad's test film). For film non-flatness, especially mounted slides, even though my best lenses are sharpest more open, I'll shoot at f/8 and focus away from the center.

 

[kiemchacsu]

thanks for tremendous efforts and great work
i am now very happy with Nikkor 55/3.5 Ai version, which I replaced my 55/2.8 AIS
i still believe the 55/3.5 is great lens.

 

[ColSebastianMoran]

Let's take it one step further... For a print of roughly 8x10, how would an iPhone shot compare to the cam-scan with the 70 Sigma? Is it possible that the iPhone shot will be "good enough."

Here are files to examine:

The 70 Sigma file exported with sharpening at 3000 pixels wide (for 10" print): http://2under.net/images/70SigmaMA28-f8-Duomo-AR40516-3k.jpg

Same chrome, shot with the excellent ExoLens Macro by Zeiss on an iPhone 6s: http://2under.net/images/iPh6-ZeissExoLens-Duomo-IMG_0018-3k.jpg

I suggest printing or examining these two files to draw own conclusions. I've printed both files, and it's hard to find visible differences. I conclude that an iPhone shot with a good macro lens can make a nice 8x10 print.

Here's a screen-grab comparison, 100% from the 3000 pixel file. Shows differences, but these are hard to see in the 10" print.

 

[ColSebastianMoran]

... and without a macro lens, the iPhone is pretty limited. Closest focus with my iPhone 6s gives about 1500 pixels wide from 35mm image, enough for screen display. The iPhone 13 Pro Max has a close-up capability that can fill the frame delivering 12MPx. While that looks good on the phone, it won't make an 8x10 print.

 

[smetfilm]

Group 3 lenses. Still good; will be able to see differences on real images in very large prints or in pixel-peeping.
Note: When used a lower magnification or on APS body, results will be closer to lenses above.

Group 3, less resolution G6-E2 and MTF50 still around G4-E4 (3-Star)
50 f/1.8 Olympus Zuiko (on Vivitar Macro Focusing Teleconverter, low contrast, some CA)
50 f/2.8 El-Nikkor (Old style, metal body, huge numerals) Some CA
50 f/2.8 Componon-S
50 f/3.5 Olympus Zuiko Macro
50 f/4 Componon (less contrast, old, small lens)
50 f/4 El-Nikkor
50 f/4.5 Enlarging Ektar
55 f/2.8 Industar (s/n 92xxx from 1992; some CA, removed in LR)
55 f/2.8 Panagor Macro (also sold as Vivitar, Komine) Vivitar
55 f/2.8 Yashica ML Macro
55 f/3.5 Micro Nikkor (s/n 899x, late model, some CA)
55 f/3.5 Micro Nikkor (s/n 204x, early, compensating aperture, some CA)
60 f/2.8 7Artisans Micro version I (less contrast, some CA, APS only, version II of this lens is considerably better)
75 f/4 El-Nikkor
75 f/4.5 Enlarging Ektar, MTF50 G4-E4
75 f/4.5 Tominon
90 f/5.6 Komura Enlarging lens (less contrast)
100 f/4 Canon FD Macro (some CA, MTF50 at G4-E3)
100 f/2.8 Canon EF L Macro on 5DSR or on Sony (considerable CA)
100 f/4 Pentax Bellows (some CA)
100 f/4 Leica Bellows Elmar (G6-E2, good contrast)

Hiya, quick question: How did you mount the enlarger lenses? Did you mount directly to the camera (with L39-native mount adapter), or did you use bellows as an intermediary for focusing?
Recently I was lucky enough to find a great deal on a Durst M305 enlarger with the 50 f/4 El-Nikkor enlarging lens. Hoping to mount it to a Canon 1300D (aps-c, 18MP), and in the future some Fujifilm camera.

Thanks for the work, Richard. Really makes getting into camera scanning a lot easier.

 

[ColSebastianMoran]

Hiya, quick question: How did you mount the enlarger lenses? Did you mount directly to the camera (with L39-native mount adapter), or did you use bellows as an intermediary for focusing?

Hello back. How mount enlarger lenses? Every which way. I use a copy stand (Polaroid MP-4, rock solid), and I prefer to have a tripod mount on the lens or on an adapter to make things more stable. I try not to use the tripod thread on the camera body.

Tricky with enlarging lens or any lens without a focusing helical to get the right extension. Also, it's handy to be able to use the mirror alignment trick, but to do so, you want to adjust focus.

Bellows is handy. A thin L39 adapter on front of bellows should allow 1:1 focus with the 50 f/4 El Nikkor, but you won't be able to focus to a lower magnification and therefore won't do for camera-scan with an APS body. Bellows is more flexible with an 80mm lens.

There are a range of focusing extension tubes. On eBay, easy to find M42 focusing tubes and to add adapters to the front and rear. If you want to be able to adjust focus, this will be the easier way.

Here's a complete setup for a 75mm lens (75 f/4 APO Rodagon-D 1x). The Nikon PN-11 has a tripod mount, I used that to attach the stack to the copy stand.

Your 50 f/4 El Nikkor should be OK for camera scan at 1x. Not one of my top lenses, but it's good.

But, my net of all this: it will be easier for you to get a macro lens that focuses without adapters. For a Canon body, I can highly recommend the 70 f/2.8 Sigma Macro ART or one of the earlier Sigma EX DG Macros.

 

[Junakreiter]

Thanks for your testing !

I have been using a Nikon 60 f2.8 for some years.
As i have seen your test charts i thought there was room of improvement,
I tried some lenses:
Two Nikons 55 f2.8 Ais are sharp in the center, but not the corners.
And two Sigma 70mm both EX and Art gave me very good sharpness,
but again, in the corners, my Nikon 60mm is better.
Do you think i have tried faulty copys of that lenses ?

 

[ColSebastianMoran]

Thanks for your testing !

I have been using a Nikon 60 f2.8 for some years.
As i have seen your test charts i thought there was room of improvement,
I tried some lenses:
Two Nikons 55 f2.8 Ais are sharp in the center, but not the corners.
And two Sigma 70mm both EX and Art gave me very good sharpness,
but again, in the corners, my Nikon 60mm is better.
Do you think i have tried faulty copys of that lenses ?

 

[ColSebastianMoran]

All very good questions... I had access to a 60 f/2.8 Micro-Nikkor AF for a short while. I was only able to test at 1x in mid-corner with a 50MPx sensor. The results were excellent resolution (91 lp/mm on glass USAF target) with good contrast. 70 Sigma Macro ART was a little better contrast in same conditions. I no longer have access to this lens.

I have several 55 Micro-Nikkors. In my tests at 1x and 0.67x to 24MPx sensor, the 55 f/2.8 is top-notch, all the way to the corners, equal to the 70 Sigma Macro ART. Here at 0.67x, f/5.6, Vlad's Test Target, to 24MPx sensor (Sony A6000). This result G0-E2 (62 lp/mm) is typical of best lenses in these test conditions.

This thread was "How great a lens do we need?" and both these lenses will give excellent cam-scan results for real film shots to 24MPx sensors.

 

[ColSebastianMoran]

Thanks for your testing !

I have been using a Nikon 60 f2.8 for some years.
As i have seen your test charts i thought there was room of improvement,
I tried some lenses:
Two Nikons 55 f2.8 Ais are sharp in the center, but not the corners.
And two Sigma 70mm both EX and Art gave me very good sharpness,
but again, in the corners, my Nikon 60mm is better.
Do you think i have tried faulty copys of that lenses ?

But you raise a good question; what happens if we are more demanding? Now with a higher res camera body (Sony A7R-IV) I have the same test at 61MPx. In this, the 55 f/2.8 Micro-Nikkor is quite good, but not quite as good as the 70 Sigma. This is Vlad's test target, the resolution of the 70 Sigma is close to the limit of the target G0-E4 78 lp/mm. Here's the comparison 1x, f/8, 61MPx sensor.



I'm curious about your test conditions? Camera body/sensor? Test target or real film scene?

 

[Godfrey]

I don't have any of your listed lenses. For negative scanning, I currently use one of these:

Micro-Nikkor 55mm f/3.5 preAI (Leica M10-M, M10-R bodies)
Leitz Macro-Elmarit-R 60mm f/2.8 (Leica M10-M, M10-R bodies)
Leitz Macro-Elmar-R 100mm f/4 (Leica M10-M, M10-R bodies)
Leica Summicron-R 50mm f/2 (Leica M10-M, M10-R bodies)
Hasselblad Zeiss Makro-Planar 120mm f/4 T* (Hasselblad 907x/CFVII 50c body)
Hasselblad Zeiss Planar 80mm f/2.8 T* (Hasselblad 907x/CFVII 50c body)

I scan 35mm full frame, 35mm half-frame, Minox 8x11, 120 6x4.5, 120 6x6, and 120 6x7 format negatives.

What I find is:

- Edge resolution and contrast with all of them on all formats is just fine for my usual largest prints, which are 11x17 inch or 13x13 inch, considered evaluating printed results at proper viewing distances.

- The Macro-Elmarit-R 60mm has a bit of field curvature and rectilinear distortion in the 1:3 to 1:1 range; it seems best suited to macro work involving small 3D subjects rather than flat field work, although the distortion is easily corrected and the field curvature worked around by using f/11-f/16 rather than larger apertures.

- The Makro-Planar 120mm lens has the best performance (flatness of field and even illumination) in the 1:4 to 1:1 range I've seen, and larger formats (6x6, 6x4.5, 6x7) do particularly well imaged with it onto the 907x's 33x44mm sensor format.

- The Summicron-R 50mm lens is the best performer I have for imaging Minox 8x11 negatives, which require approximately 2.7:1 magnification to maximize filling a 24x36 format sensor with the full Minox frame plus the film rebate. It shows superior flatness of field and the best edge to edge illumination, contrast, and sharpness for this high magnification scan setting.

G

 

[ColSebastianMoran]

What I find is:

- Edge resolution and contrast with all of them on all formats is just fine for my usual largest prints, which are 11x17 inch or 13x13 inch, considered evaluating printed results at proper viewing distances.

Thanks for these comments, if you're happy with results from the 55 f/3.5 Micro-Nikkor, that reinforces my original thought in this thread. While we can chase corner resolution with better lenses, and I'll admit that I enjoy doing so, we can get "good enough" results in camera-scan from just "pretty good" lenses. Don't need the ultimate.

That said, I'm now leaning to recommend simple, straight-forward setups, modern lenses, that don't need adapters.
- 70 or 105 Sigma Macro ART in Sony and Canon mounts
- Manufacturers' macro lenses (that focus to 1x)
- 7 Artisans 60 f/2.8 Macro version II (better than ver 1) for most APS mounts

 

[Godfrey]

Thanks for these comments, if you're happy with results from the 55 f/3.5 Micro-Nikkor, that reinforces my original thought in this thread. While we can chase corner resolution with better lenses, and I'll admit that I enjoy doing so, we can get "good enough" results in camera-scan from just "pretty good" lenses. Don't need the ultimate.

That said, I'm now leaning to recommend simple, straight-forward setups, modern lenses, that don't need adapters.
- 70 or 105 Sigma Macro ART in Sony and Canon mounts
- Manufacturers' macro lenses (that focus to 1x)
- 7 Artisans 60 f/2.8 Macro version II (better than ver 1) for most APS mounts

The Micro-Nikkor 55mm f/3.5 does a perfectly fine job with most format negs for prints up to 11x17 inch. It has less geometric distortion compared to the Macro-Elmarit-R 60mm but not, perhaps, the same contrast and color fidelity. Trade-offs ... Rendering the scans in post has larger influences on your final output than these two lenses do.

A modest trick that I used for a long time:

If you use a high-quality APS-C body to copy 35mm negs and slides instead of a FF body, you are working with magnification ratios in the 1:2 range rather than 1:1 range most of the time. Most 35mm SLR macro lenses perform slightly better in this range than in the 1:1 range, and using a smaller format copy camera allows easier focusing and better DoF; the lower magnification potentially also means a little less tendency to camera movement. A good 24 mpixel sensor camera (I used the Leica CL digital body for this extensively when I had it) used at low ISO settings nets a high resolution scan this way that is every bit as good as a FF sensor with the same pixel resolution.

I have to be much more careful with the setup and precise with my focusing now that I've got only 35mm FF bodies to work with. I almost kept the Leica CL body around specifically for this use, but eh? I wanted the money it represented for some other things I was involved in. Trade-offs... 😀

I'd use the 907x with the Makro-Planar 120mm lens for more scanning of color negs due to the 907x sensor's superb dynamic range and separation of tones and colors except for that fact that this camera/lens combination is huge and bulky when set up for the magnification required by 35mm FF format (approx 1.2:1)... it takes the lens at maximum extension plus several extension tubes or a big bellows extension which makes it very susceptible to camera shake unless I really lock everything down very carefully. A wireless remote release then produces the very best exposures.

G

 

[Sonnar Brian]

At F8, the 55/3.5 Micro-Nikkor was top of class for corner sharpness, of 32 "normal" lenses tested in 1976.

Compare with the Konica 55/3.5,

 

[Chriscrawfordphoto]

It is interesting how poorly the enlarging lenses tested. They're actually made specifically for photographing film (onto paper). I wonder if the poor performance was because you were doing 1x magnification instead of the much higher magnifications usually used for enlarging? Maybe they're optimized for higher magnification?

 

[ColSebastianMoran]

The Micro-Nikkor 55mm f/3.5 does a perfectly fine job with most format negs for prints up to 11x17 inch. It has less geometric distortion compared to the Macro-Elmarit-R 60mm but not, perhaps, the same contrast and color fidelity. Trade-offs ... Rendering the scans in post has larger influences on your final output than these two lenses do.

A modest trick that I used for a long time:

If you use a high-quality APS-C body to copy 35mm negs and slides instead of a FF body, you are working with magnification ratios in the 1:2 range rather than 1:1 range most of the time. ... snip ...

Godfrey, you make a good point, as always!

Here is the 55 f/3.5 Micro-Nikkor, LEFT at 1:1.5 for 35mm to APS, RIGHT at 1x for 35mm to FF body. Both at f/8.