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

[ColSebastianMoran]

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

Yes, have studied that terrific test article many times.

Key point to me: most lenses have designs with tradeoffs and are optimized for certain magnification ratios. They do less well at other ratios. The 55 f/3.5 I've understood as a terrific performer, very sharp, ever since I started in photography 50+ years ago. But, compared to the f/2.8 and to modern lenses, it's very good for closeups and flowers, but not as good at 1x.

I believe this is also the answer to Chris Crawford's good question; enlarging lenses are designed for enlarging, not 1x copying.

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?

You two probably know more than I do about lens design. I have been given the notion that a symmetrical lens will be insensitive to magnification ratio. Hard to find a lens with perfectly symmetrical design, but one is the Tominon 50mm f/4.5 lens for the Polaroid MP copy system. I've tested this lens with very good results from 1:3 to 1x to 3:1. Doesn't much matter since we don't see symmetrical designs in current lenses, but you think the theory has merit? That symmetrical lenses will be insensitive to magnification, will work at a wide range of ratios?

 

[Chriscrawfordphoto]

Yes, have studied that terrific test article many times.

Key point to me: most lenses have designs with tradeoffs and are optimized for certain magnification ratios. They do less well at other ratios. The 55 f/3.5 I've understood as a terrific performer, very sharp, ever since I started in photography 50+ years ago. But, compared to the f/2.8 and to modern lenses, it's very good for closeups and flowers, but not as good at 1x.

I believe this is also the answer to Chris Crawford's good question; enlarging lenses are designed for enlarging, not 1x copying.


You two probably know more than I do about lens design. I have been given the notion that a symmetrical lens will be insensitive to magnification ratio. Hard to find a lens with perfectly symmetrical design, but one is the Tominon 50mm f/4.5 lens for the Polaroid MP copy system. I've tested this lens with very good results from 1:3 to 1x to 3:1. Doesn't much matter since we don't see symmetrical designs in current lenses, but you think the theory has merit? That symmetrical lenses will be insensitive to magnification, will work at a wide range of ratios?

I don't know a lot about lens design either, unfortunately. Lenses are definitely made for certain magnifications, though. I learned this the hard way recently. I shoot with Olympus Micro Four Thirds and I have their 60mm f2.8 Macro, which is capable of 1x magnification at the closest end of the focus range. I have always used it for two things: photographing bugs and flowers, and photographing artwork for artists who need photos for their websites, for making prints, and for publication. The lens is excellent for both of those things, and for photographing flat art, it gives far better edge sharpness than a regular lens like the Olympus 12-40mm f2.8 or the Olympus 45mm f1.8. Those lenses are very sharp for normal subjects but not for flat art.

Recently, I was photographing some sunflowers with it, and the closeups looked great, but then I backed up and did one from about 100 feet away, to see a wide view of the field of sunflowers. The images were sharp in the center and very soft away from the center! NOT a good landscape lens!

 

[Jhausler]

Really appreciate the work here and the great conversation.

I've decided to update my scanning kit from a BEOON stand, which isn't particularly level, to a kit from NegativeSupply. I've been looking for a lens to buy and adapt to a SonyA7 (mk1) and have a few questions.

You placed the 105 f/2.8 Micro-Nikkor AF-D in Tier 2, yet the 105 f/2.8 Micro-Nikkor AF (non Digital) is in Tier 5. I understand these to be the same optically, and only different in the ability to send distance information to a Nikon camera. Should their output vary that differently?

I only ask because while chatting with AJ, co-founder at NegativeSupply, he recommended the NIkkor 105/2.8 or the f/4 with the PK-13 extension, so i've been eyeing those up online.

Another question.. if i plan to duplicate 35mm and 120 film, will a 55m lens be any better or worse, or too close/far to the film/light source for duplicating? Seems there are so many choices in lenses.

Last one... any difference between the Nikon Micro-Nikkor-P 55mm f/3.5 and the version without the P?

 

[Chriscrawfordphoto]

Really appreciate the work here and the great conversation.

I've decided to update my scanning kit from a BEOON stand, which isn't particularly level, to a kit from NegativeSupply. I've been looking for a lens to buy and adapt to a SonyA7 (mk1) and have a few questions.

You placed the 105 f/2.8 Micro-Nikkor AF-D in Tier 2, yet the 105 f/2.8 Micro-Nikkor AF (non Digital) is in Tier 5. I understand these to be the same optically, and only different in the ability to send distance information to a Nikon camera. Should their output vary that differently?

I only ask because while chatting with AJ, co-founder at NegativeSupply, he recommended the NIkkor 105/2.8 or the f/4 with the PK-13 extension, so i've been eyeing those up online.

Another question.. if i plan to duplicate 35mm and 120 film, will a 55m lens be any better or worse, or too close/far to the film/light source for duplicating? Seems there are so many choices in lenses.

Last one... any difference between the Nikon Micro-Nikkor-P 55mm f/3.5 and the version without the P?

There can be a lot of sample variation in lenses. Two of the same model can have one of them super sharp and the other really crappy. With modern lenses especially, there is ZERO quality control at the factory. NONE. 80 to 90% of the lenses I look at that are in current production are decentered. Many of them so severely that the lens is completely unusable. This wasn't a problem back in the 80s and 90s when I was younger. If you bought lenses from good companies like Canon, Nikon, and Olympus...they were good. Period. Sadly, that is no longer true.

 

[sojournerphoto]

Thank you for running these tests. Really interesting and useful.

 

[JoeV]

There can be a lot of sample variation in lenses. Two of the same model can have one of them super sharp and the other really crappy. With modern lenses especially, there is ZERO quality control at the factory. NONE. 80 to 90% of the lenses I look at that are in current production are decentered. Many of them so severely that the lens is completely unusable. This wasn't a problem back in the 80s and 90s when I was younger. If you bought lenses from good companies like Canon, Nikon, and Olympus...they were good. Period. Sadly, that is no longer true.

Chris, do you have a tutorial on how to test a lens for being decentered?

 

[Noll]

The sharper the better, especially with emulsions ISO200 and slower. I'm using the hi-res mode on an Olympus EM1mk2 + Zuiko 30 mm f/3.5 macro to make 50 MP scans of ultra-fine grained Velvia 100 slides. The 30mm macro is a phenomenally sharp budget macro lens, and even with the hi-res composite there barely seems to be enough resolution in 50 MP to resolve the finest details. I bracket the exposures and combine in post to retain shadow and highlight detail.

 

[peterm1]

A year or two back I bought a Pentax M42 based bellows outfit with slide holder to scan a large number of old slides I have from the 1980s/90's. The manual for the outfit recommends use of the M42 Super Takumar 55mm f1.8 in part because the outfit is set up for this specific lens (markings on the bellows slide indicate how much extension is needed etc. I found its results to be fine for this purpose. I have a few more slides to scan and I may consider trying other lenses - for example I have the 50mm f4 Macro lens in Super Takumar. With a bellows outfit, as the bellows takes care of the extension needed to provide the required magnification no extra extension is needed on the lens itself but this macro lens is reported to have good flatness of field so there may be benefits in that regard out towards the edge of the slides. I doubt that there would be much if any other sharpness benefit though as in event both lenses will be sued at f8 or maybe f11. But I will do some test shots and report back.

 

[skahde]

When I started camera-scanning my slides, lens-quality never bothered me and I never cared much about. I took what I had since 1990 and it seemed to work: Micro-Nikkor Ais 2.8/55mm. There is a reason I never considered selling it.

 

[ColSebastianMoran]

First, sorry to be slow in replying.

You placed the 105 f/2.8 Micro-Nikkor AF-D in Tier 2, yet the 105 f/2.8 Micro-Nikkor AF (non Digital) is in Tier 5. I understand these to be the same optically, and only different in the ability to send distance information to a Nikon camera. Should their output vary that differently?

I only ask because while chatting with AJ, co-founder at NegativeSupply, he recommended the NIkkor 105/2.8 or the f/4 with the PK-13 extension, so i've been eyeing those up online.

Yikes, the dangers of publishing.

I have only tested one 105 f/2.8 Micro-Nikkor and that's the AF-D version.
I placed it in group 2, pretty good, based on manual focus results on Sony bodies.
Testing the same lens on my D7200, the results are poor, that's the Group 5 entry, and I cannot explain the poor result. Other lenses do well on the D7200.

Another question.. if i plan to duplicate 35mm and 120 film, will a 55m lens be any better or worse, or too close/far to the film/light source for duplicating? Seems there are so many choices in lenses.

A 100mm lens will give good working distances for 35mm to FF, and about about 18" from film for 120. Fine if that works in your setup.

Last one... any difference between the Nikon Micro-Nikkor-P 55mm f/3.5 and the version without the P?

I've got all of them and don't like any for 1x for cam-scan of 35mm. Bjørn Rørslett, a real expert, has comments on all the versions, lots of changes over the years. However, I believe his reviews are for shooting larger subjects, e.g. flowers. I don't like these lenses at 1x, but might be fine for, say cam-scanning of 120.

You're using an A7 body, I think that's an excellent choice for camera scanning; has the features you want and not the ones you don't need, the focus peaking is super for manual focusing. For your body, I would highly recommend:
- 70 or 105 Sigma Macro ART (gives you AF if you want it)
- 55 f/2.8 Micro Nikkor AI or AIS
- 50, 70, or 105 Sigma EX DG Macro (their lenses for AF Nikon or Canon bodies)
- or the more exotics mentioned in my lists

 

[Godfrey]

When I started camera-scanning my slides, lens-quality never bothered me and I never cared much about. I took what I had since 1990 and it seemed to work: Micro-Nikkor Ais 2.8/55mm. There is a reason I never considered selling it.

Similarly, although I have several appropriate lenses, I just happened to accrete them over the years as I worked through various other projects and didn't buy them specifically for this purpose.

My pre-AI Micro-Nikkor 55mm plus M tube (for 1:1 magnification) was purchased about maybe 21 years ago from an Ebay seller for $60 in "fine glass, good operation, rough cosmetic" condition. It was obviously used professionally for flat art copy work for several decades when I bought it, and is still just fine for use, although at this point (57 years old) it really should go out for a CLA as the old lubricant on the focus helicoid is pretty dried out.

It worked fine on the Sony A7 when I had that, and on the Leica CL, and now on the Leica M10-M/M10-R, and on all the other cameras in between ... Including on my 1961 Nikon F plain prism. 😀

G

 

[hap]

Does have anyone with experience using a 60mm 2.8 macro Leica R for scanning?

 

[sojournerphoto]

First, sorry to be slow in replying.


Yikes, the dangers of publishing.

I have only tested one 105 f/2.8 Micro-Nikkor and that's the AF-D version.
I placed it in group 2, pretty good, based on manual focus results on Sony bodies.
Testing the same lens on my D7200, the results are poor, that's the Group 5 entry, and I cannot explain the poor result. Other lenses do well on the D7200.


A 100mm lens will give good working distances for 35mm to FF, and about about 18" from film for 120. Fine if that works in your setup.


I've got all of them and don't like any for 1x for cam-scan of 35mm. Bjørn Rørslett, a real expert, has comments on all the versions, lots of changes over the years. However, I believe his reviews are for shooting larger subjects, e.g. flowers. I don't like these lenses at 1x, but might be fine for, say cam-scanning of 120.

You're using an A7 body, I think that's an excellent choice for camera scanning; has the features you want and not the ones you don't need, the focus peaking is super for manual focusing. For your body, I would highly recommend:
- 70 or 105 Sigma Macro ART (gives you AF if you want it)
- 55 f/2.8 Micro Nikkor AI or AIS
- 50, 70, or 105 Sigma EX DG Macro (their lenses for AF Nikon or Canon bodies)
- or the more exotics mentioned in my lists

I’m using the Sigma 105mm macro art on my Panasonic S1r. It seems to work pretty well, even if I use multi shot high res mode at 1:1 on 35mm film. At f4 the grain in the extreme corners is slightly soft at 100%, but it’s easy to stop down to 5.6.

I usually use manual focus on the grain with 20x live view.

The S1r has the advantage that, although multi shot is only 8 frames, it does the combination in camera and avoids a computer step🙂

 

[ColSebastianMoran]

I’m using the Sigma 105mm macro art on my Panasonic S1r. It seems to work pretty well

I have than lens in Sony and I think it's Tops. Resolution and clean images matching the five or six best lenses I've tested and used.

 

[Godfrey]

Does have anyone with experience using a 60mm 2.8 macro Leica R for scanning?

Yes, it's one of my most used lenses. As I said up thread, it has a little bit of field curvature so when scanning flat negatives and such, it's best to stop it down to f/8-f/11.

I tend to use it more than the Micro-Nikkor 55mm because the Leica cameras include the lens profile for the Macro-Elmarit 60 mm which does improve its output.
i've also found it to be an excellent general purpose "long normal" for more common photography, and the infinity to 1:2 macro range with standard mount is very handy.

G

 

[Grim Tuesday]

@ColSebastianMoran have you tested any of Nikon's modern macro lenses? I have a 55mm f2.8 AI that I have always been very satisfied with and just upgraded my camera body from APSC D5500 to full frame Z7. Reading the literature about modern Nikon lenses is very interesting. To summarize:

- The famous 55mm 2.8 was the first to use an automatically cammed close range correction to reduce spherical aberration at close or infinity focus
- In the 60mm AF-D, some optical tradeoffs (a built-in teleconverter) were made to make the lens natively go to 1:1 without extension tubes and support the autofocus screw drive. This is probably why it is slightly worse than the 55mm 2.8. It's also the oldest lens Nikon has posted MTF data on, and compared to newer lenses it doesn't look amazing.
- In the 60mm G (current production for F-mount) tradeoffs were made to allow the lens to be internally focusing. The MTFs posted on this lens, however, do look awesome
- There's some marketing hype about the bokeh in the 50mm MC Z-mount lens. Usually quality of bokeh comes from adding spherical aberration and at the expense of resolution. MTFs support this compared to the 60mm G, with them being slightly worse. Of course, it's hard to know if the MTFs are for infinity focus or something close up.

To distill these ramblings into a question: You mentioned you currently lean towards recommending a modern autofocus lens from the camera manufacturer for most film scanning purposes these days. For Nikon Z, would you recommend the 60mm AF-S G-ED with FTZ adapter (not the old 60mm) or the brand new Z-mount 50mm MC? Would I notice a difference on the 46mp sensor between one of these and my good old 55mm f2.8? I'm trying to get a setup good enough that I can feel comfortable selling my Coolscan 9000 🙂

 

[sojournerphoto]

@ColSebastianMoran have you tested any of Nikon's modern macro lenses? I have a 55mm f2.8 AI that I have always been very satisfied with and just upgraded my camera body from APSC D5500 to full frame Z7. Reading the literature about modern Nikon lenses is very interesting. To summarize:

- The famous 55mm 2.8 was the first to use an automatically cammed close range correction to reduce spherical aberration at close or infinity focus
- In the 60mm AF-D, some optical tradeoffs (a built-in teleconverter) were made to make the lens natively go to 1:1 without extension tubes and support the autofocus screw drive. This is probably why it is slightly worse than the 55mm 2.8. It's also the oldest lens Nikon has posted MTF data on, and compared to newer lenses it doesn't look amazing.
- In the 60mm G (current production for F-mount) tradeoffs were made to allow the lens to be internally focusing. The MTFs posted on this lens, however, do look awesome
- There's some marketing hype about the bokeh in the 50mm MC Z-mount lens. Usually quality of bokeh comes from adding spherical aberration and at the expense of resolution. MTFs support this compared to the 60mm G, with them being slightly worse. Of course, it's hard to know if the MTFs are for infinity focus or something close up.

To distill these ramblings into a question: You mentioned you currently lean towards recommending a modern autofocus lens from the camera manufacturer for most film scanning purposes these days. For Nikon Z, would you recommend the 60mm AF-S G-ED with FTZ adapter (not the old 60mm) or the brand new Z-mount 50mm MC? Would I notice a difference on the 46mp sensor between one of these and my good old 55mm f2.8? I'm trying to get a setup good enough that I can feel comfortable selling my Coolscan 9000 🙂

Whilst I can’t comment on the Nikon lenses, I’ll add in that I am happy with the results I am getting from my Panasonic S1r, at 47Mp, and Sigma 105mm Art macro in L mount. I previously used a Nikon 9000.

A couple things I thought about before putting this set up together:

- the Nikon scans full colour at each pixel, so (all else being equal) you need more resolution in a Bayer sensor to match the Nikon - about double or c.43Mp.

- field flatness is important. The Sigma seems pretty good. I scan at f5.6 or f8. Alignment also matters. I’ve read that the Laowa 2x apo-macro has some field curvature, tempting though it is to try it (not sure I really want to stitch for 35mm though). This might also mean that an excellent lens optimised for film will underperform in this application on a digital sensor, given the cover glass stack tends to introduce field curvature as a minimum.

- the Panasonic does in camera multi shot high res. This increases resolution by giving full colour at each pixel and the intermediate positions (probably) don’t increase absolute resolution but do reduce aliasing. I very rarely use this in 35mm, but it does improve (shrink) grain in some slow films. I went down the high res scanning route in part because I wanted to be able to scan down to the grain - aliasing often makes it look bigger and more intrusive than it really is and can spoil a print. Plus, I am interested in how the lens image is recorded in film as part of my process.

Mike

 

[Grim Tuesday]

Perhaps this is a question for a different thread, but what is the current thinking vis-a-vis manual vs auto focus macro lenses? A few years back, last time I was thinking about selling my Coolscan I remember there was broad agreement that manually focusing at the beginning of a session was the way to go. Now, I sometimes see people suggesting autofocusing on each frame, to account for individual curvature of each frame. Unfortunately I have no autofocus macro lenses so cannot test myself. Where do others land on this?

 

[Godfrey]

Hmmm ... Just thinking out loud ... The Hasselblad CFVII 50c back has about the best dynamic range of anything I've used. To fill its 33x44 mm sensor format on the long edge with a 35mm frame requires 1.2:1 magnification. I wonder if I can get close to that using a pair of 32mm extension tubes behind the Planar 80mm lens on the 500CM body? And I wonder how well the Zeiss Planar 80 at such near-range magnifications?

I know the Makro-Planar 120mm has superb flatness and beautiful rendering, but to get to 1.2:1 magnification takes at least 3 32mm extension tubes plus its macro focusing helicoid, which is a hugely clumsy thing to manage.

All just for fun, I may play with some setups later today. 😀

G

 

[skahde]

Hmmm ... Just thinking out loud ... The Hasselblad CFVII 50c back has about the best dynamic range of anything I've used. To fill its 33x44 mm sensor format on the long edge with a 35mm frame requires 1.2:1 magnification. I wonder if I can get close to that using a pair of 32mm extension tubes behind the Planar 80mm lens on the 500CM body? And I wonder how well the Zeiss Planar 80 at such near-range magnifications.

At such magnification and without doing any calculations I would eyeball that a Nikkor 2.8/55 mm or any other 35mm-Macro lens would fill the sensor just fine and try to adapt that. Just to put another option on the table.