Using a digital camera to test a large format lens

[cptrios]

Well, first of all, I'm not totally sure where this post belongs, but since the main focus is a large format accessory, this seems right! Also, I'm still feeling my way around my first steps in LF photography, so this is a highly-uninformed question.

A little while back I picked up a very nice Fujinon 150mm f/5.6, having previously only used an Angulon 90/6.8 with a busted shutter on my B&J Press camera. I was immediately dismayed by the softness in scans at high magnification. I've since been able to attribute that problem to back-focusing caused by the ground glass being a bit too deep - an issue that I have yet to remedy (though the extent to which I've wound up overcompensating has been a real eye-opener re: 4x5 depth of field).

Somewhere in there, I decided to kludge up a holder to mount my NEX-7 on the camera for the purpose of trying lenses out. I was very pleasantly surprised by how well the Fujinon did:

A 100% crop from the upper-right quadrant of the digital image, which was roughly in the center of the image circle. LR sharpening turned down to zero (which always depresses me). This one is at either f/8 or f/11, I think.

This made me breathe a big sigh of relief with regards to the quality of this copy of the lens. But the simple form of my question is: should it have? More specifically: if the lens is capable of resolving that much detail (at least in the center) on a 24mp APS-C sensor, is it reasonable to expect that it's at least putting that level of detail on the film? Could a theoretical film with the same lp/mm as the sensor resolve the same detail? Or is there some function of a digital sensor that's giving me a false sense of optimism?

 

[Alan Edward Klein]

Well, a digital sensor has no intermediate film and scanning effects that would reduce the resolution of the final results, including your ground glass issue. As an aside, Clyde Butcher uses medium format lenses made for film cameras with his Sony digital cameras and prints 60"+ wide photos. I would assume, large format lenses would be similarly capable.

 

[ic-racer]

Fujinon 150mm is a nice lens. I have one and it is a beautiful lens for film photography.

 

[btaylor]

You’ve proved the lens has great resolving power which is not surprising. Now you have to properly adjust the groundglass and film plane to be identical. Then you will get the sharp pictures you expect.

 

[cptrios]

Thanks for the replies. I'm getting a mild "I was right" vibe, which is nice! I actually think I'll end up getting rid of the lens at some point soon, since I'd rather have something wider (and with deeper DoF), but it definitely seems nice. And I bought a lot on eBay recently with some film holders and two orphaned ground glass backs, so hopefully one of them is sound and can be attached to my camera! If mine's error were in the other direction, I'd try shimming it, but I haven't the foggiest how I'd move it forward without sanding the whole thing down. I'm sure that would end poorly!

 

[BobD]

Digital cameras have built-in sharpening and aberration-correcting algorithms, you know.

 

[btaylor]

Thanks for the replies. I'm getting a mild "I was right" vibe, which is nice! I actually think I'll end up getting rid of the lens at some point soon, since I'd rather have something wider (and with deeper DoF), but it definitely seems nice. And I bought a lot on eBay recently with some film holders and two orphaned ground glass backs, so hopefully one of them is sound and can be attached to my camera! If mine's error were in the other direction, I'd try shimming it, but I haven't the foggiest how I'd move it forward without sanding the whole thing down. I'm sure that would end poorly!

There have been many discussions over at the Large Format Photography Forum on ground glass adjustment, it may be helpful to poke around there for some techniques.

 

[MattKing]

I can see using it to test for damage or obvious defect.
But I'm not sure that using a small sensor to image a small portion of the lens' image is going to tell you much. You basically need something like a flat piece of film to check whether the light images out to the corners the way you want it to.

 

[cptrios]

Yeah I wouldn’t expect the same performance toward the borders (especially since the angle of light will do weird things to the sensor the farther out it gets). I did take some shots with full vertical shift, and they also looked good, but that’s still likely only around the 25%-from-bottom mark.

Still, it doesn’t need to do anywhere near that well, considering the resolution of any film beyond CMS 20. By my calculations a full array from the Nex-7 sensor in the 4x5 frame would be over 700mp…just a smidge of overkill there!

Also, I appreciate nobody judging me for pixel-peeping. I have dreams of huge prints.

 

[Craig]

Also, I appreciate nobody judging me for pixel-peeping. I have dreams of huge prints.

I think the biggest killer of large, sharp prints isn't the resolution of the lens itself, it's vibration. Either in the camera when taking the photos or when printing. A tiny bit of vibration will wipe out your sharpness at big enlargements.

 

[DREW WILEY]

Like many beforehand with the same notion, using a digital camera to rate large format lens performance is a mismatch, and potentially going to give you misleading results. For one thing, view camera lenses are engineered with a large image circle to accomodate movement at both head-on and strongly tangential angles of light. Digital sensors are generally not, and you've only capturing the center of the image circle at best anyway. I don't get the point. Your digital sensor is what - about one square inch of photon capture overall, while 4x5 film is nearly twenty square inches. That like comparing a dinosaur beside an ant, yet expecting to interpret the world of that dinosaur through the eyes of that little ant. Doesn't make sense. Your Fujinon 150 is plenty competent to do the job well, and pixel-peeping won't tell you anything additional which is actually useful. If you can't get sheet film shots which are far far far more detailed than what you get with a tiny digital sensor, then the lens itself is certainly not to blame. With such a massive difference in capture surface area, it would have to be on hell of a fungus-ridden funky old lens to come up the same. But there are many other variables involved; and you're only as good as the weakest link.

 

[Neil Poulsen]

Well, a digital sensor has no intermediate film and scanning effects that would reduce the resolution of the final results, including your ground glass issue. . .

That depends on the lens being tested and its age. Color film renders different ranges of color on different layers. Given a range of colors that correspond to a particular layer in film, modern lenses are optimized to make that range of colors sharpest on that layer of film. Consequently, modern lenses have differing planes of focus for differing ranges of colors. The overall effect is to make images printed from color film sharper.

On the other hand, digital sensors render all colors in the same plane of focus. As a result, a digital sensor may not be the best device to check the sharpness of a lens.

 

[Mal Paso]

That depends on the lens being tested and its age. Color film renders different ranges of color on different layers. Given a range of colors that correspond to a particular layer in film, modern lenses are optimized to make that range of colors sharpest on that layer of film. Consequently, modern lenses have differing planes of focus for differing ranges of colors. The overall effect is to make images printed from color film sharper.

On the other hand, digital sensors render all colors in the same plane of focus. As a result, a digital sensor may not be the best device to check the sharpness of a lens.

Huh? Where does this come from?

 

[Light Capture]

Well, first of all, I'm not totally sure where this post belongs, but since the main focus is a large format accessory, this seems right! Also, I'm still feeling my way around my first steps in LF photography, so this is a highly-uninformed question.

A little while back I picked up a very nice Fujinon 150mm f/5.6, having previously only used an Angulon 90/6.8 with a busted shutter on my B&J Press camera. I was immediately dismayed by the softness in scans at high magnification. I've since been able to attribute that problem to back-focusing caused by the ground glass being a bit too deep - an issue that I have yet to remedy (though the extent to which I've wound up overcompensating has been a real eye-opener re: 4x5 depth of field).

Somewhere in there, I decided to kludge up a holder to mount my NEX-7 on the camera for the purpose of trying lenses out. I was very pleasantly surprised by how well the Fujinon did:

A 100% crop from the upper-right quadrant of the digital image, which was roughly in the center of the image circle. LR sharpening turned down to zero (which always depresses me). This one is at either f/8 or f/11, I think.

This made me breathe a big sigh of relief with regards to the quality of this copy of the lens. But the simple form of my question is: should it have? More specifically: if the lens is capable of resolving that much detail (at least in the center) on a 24mp APS-C sensor, is it reasonable to expect that it's at least putting that level of detail on the film? Could a theoretical film with the same lp/mm as the sensor resolve the same detail? Or is there some function of a digital sensor that's giving me a false sense of optimism?

Matches my experience with some random tests (can't claim it's a rule, limited testing was done). I was surprised how well it looked. Used Nikon Z7 for testing (apparently it's one of the cameras that work well with all kinds of glass).
Earlier digital backs don't really deliver good results with older lenses. Newer backs and cameras are producing excellent results with lenses I tested.
I had some lenses that struggled to resolve 6mpix but when newer cameras appeared those lenses performed way better. My conclusion was that there's much more to the performance than blaming the lens. There was a good portion of the results that could be attributed to newer sensors.

On the other hand, great results achieved in live view with digital camera don't translate to good results on film.
Film holders might be off for their distance to film. Ground glass is also off by a bit. Film always bulges to some extent.

The only way of translating these results exactly to film is measuring dimensions of the film holder (distance to film), matching that with groundglass distance and using vacuum film holders.
Also, if 4x or 8x loupe is used, 4x or 8x magnification will be reasonably focused.

With all that being said, excellent results can be achieved with good technique. Focusing wide open (some LF lenses will open even further than their minimum f-stop) and stopping down with hyperfocal distance and calculation in mind will mitigate these. If you're only stopping 1 or 2 stops down then above considerations become critical.

 

[DREW WILEY]

There are all kinds of differences which need to not only be realistically weighed against each other first, but also optimized in order to be realistic. Some of the relevant issues were pointed out in the previous post. You need to have a true film plane, ideally measured with a depth micrometer and based on an actual vacuum or adhesive holder retaining the film truly flat. Your front and rear standards need to be in perfect parallel alignment, which is hard to do with any wooden camera. Furthermore, view camera lenses need to be assessed at tangential angles as well as head-on angles, which most digital cameras were never designed for unless they're specifically classified as digital backs for view camera usage. Then there are indeed those issues of how the respective wavelengths of light land in the layer, or on a surface. Have fun, but don't assume what you think you have discovered has general relevance. If it works for you, fine; but that doesn't necessarily mean it will be a suitable approach for others with more finicky standards.

 

[M Carter]

Digital cameras have built-in sharpening and aberration-correcting algorithms, you know.

In their in-cmaera JPEG rendering, yes, depending upon the camera; but anyone doing serious testing would be using camera raw files.

 

[BobD]

... but anyone doing serious testing would be using camera raw files.

Which is why I mentioned it.

 

[M Carter]

Which is why I mentioned it.

Funny, I quoted the entire comment. must have missed that, hiding between "digital" and "you know" somewhere?

 

[wiltw]

Consider that the digital SLR with a good lens (Canon EF 70-200mm f/4 USM L IS) can resolve easily resolve 2100 Line-pairs per Picture Width...which equates to 2222 line-pairs per inch or 87 line-pairs per millimeter.
According to large format lens manufacturer documentation, such lenses only provide 5-10 (maybe 20, in rare cases) line-pairs per millimeter. It is not 'inferior' lens optical performance, but the fact that the total lines of resolution have to cover FOUR times the amount of film (in one direction), 92mm short dimension of film frame vs. 24mm!

OTOH, if you used 75mm dSLR lens vs 75mm large format lens to photograph same object at same camera position, that object would be identically sized on focal plane regardless of the format size (while you normally would use something about 1.5x 'normal' FL to make the object the same 'fraction of the frame size'. The large format image would be fewer line-pairs used to capture that object!