120 Rollfilm, Flatness, Backs, Lenses, and Results (Bronica GS1)

[yessammassey]

TL; DR? Skip six paragraphs (including the list) to get to the meat of the issue. I have a terrible record of buying 'bad' wide angle lenses. It kicked off with Nikon SLRs. I went through 5 used manual focus Nikkors between 24 & 28mm before just kind of giving up. They all had either weird centering flaws or severe incongruities in performance between distant and close focus. Sometimes 1 or 2 out of 4 corners would never sharpen up when stopped down, or there would just be a 'smearing' of detail from edge to corner when focused past 5m or so. My last try with Nikon was with a 24mm 2.8 non-Ai purchased from a local camera store. It was sharper wide open in two of the corners than it was in the center!

So what was I to make of my experience with these lenses being so different from all of the positive reports I was seeing online? After 5 lenses, 3 bodies, all sorts of film stocks and scanning methods, all producing the same results with these lenses, I could only conclude that most of the used lenses being sold on the market these days must be worn out from decades of use. They've been knocked around and their helicoids actuated tens of thousands of times. As a result, the optical elements lose their factory alignment. The copies that people are reviewing online were better cared-for, and/or their quality was at the high end of the sample variation curve. I guess I shouldn't been surprised that the older lenses that people are actually selling at this point (in stores or on the auction site) are the duds that they don't want to keep for themselves. Otherwise they hold onto them for the most part. The Sekor-C 90mm that came with the RB67 I bought from the estate of a deceased photographer is truly perfect. And of course he held onto it until the end.

But it took me a while to realize that I wasn't going to be getting much good stuff buying from resellers. Here's the whole sorry story told by a list of lenses that were purchased from the auction site - Camera Hell. These all came after I gave up on the Nikkors.

• 28mm 2.8 Minolta late MD (from a charity thrift store's auction site account) - super dusty but optically centered, Kept.

• 24mm 2.8 Rokkor - decentered, Returned
• 24mm 28 Rokkor - decentered, Returned
• 24mm 2.8 Rokkor - apparently 'repaired' at some point. No infinity. Returned, fixed, sent back to me. Decentered, Kept (I couldn't bring myself to ask for a refund after the guy went through all the trouble of fixing infinity focus)
• 65mm 4 Mamiya-Sekor K/L - decentered floating element. Used for tripod landscapes at f/22 & hyperfocal distance for ages so I didn't realize until long after the return window had closed, but unusable at f/8 & wider, especially focused close. Now going back I can see the unsharp 1/3rd of the frame on all the f/22 landscape slides. Kept, only because I can't conscientiously sell it.
• 80mm 2.8 Arsat-C - perfect, Kept
• 180mm 2.8 Carl Zeiss Jena - perfect, Kept
• 55-90mm 4.5-6.7 for Fuji GA645Zi - weird soft spot along left side, Returned.
• 65mm 8 Fujinon SW for Fuji RF - decentered, Returned
• 65mm 8 Fujinon SW for Fuji RF - left side unsharp, middle moderately sharp, right side sharp out to the edge. WTF? Returned.
• 100mm 3.5 Fujinon S for Fuji RF - left side sharp out to the edge, middle sharp, extreme right side unsharp. WTF? Kept (wasn't as bad as 65mm)

... That's an inordinate amount of background information to get to my question, but you might notice that getting down to the end of the list, there are more medium format lenses. You might also notice that the issue I call decentering (basically uneven corner performance) wasn't always the issue with the MF lenses. Two out of the three Fuji lenses seemed to have some different kind of issue.

I've heard whispers of a thing called 'film flatness'. It seems to be a subject of some contention. Some maintain that it's an overblown issue. Either way, it's something that seems to be much more of a concern with 120 film than 135.

Meat:
I wonder, would film flatness issues cause the kind of linear, across-the-frame gradation of sharpness that you see (very faintly) in a photo like this? It's the left and right edges of a scan from the 100mm 3.5 Fuji lens. It's so hard to display with a scan on the web, and the 'sharpening' caused by jpeg downsampling destroys any sense of the difference in sharpness between the left edge and the middle (therefore I've cut it out) but try to notice the difference in sharpness between the tree on the left and the tree on the right, or the difference between the highlights on the left and the one lit window in the building on the right, or the difference between the two bottom corners. This was taken with a Fujinon SW 100mm 3.5 at f/16 w/ a cable release on a heavy Silk tripod and pan head. Fuji GL690. (click to view full size)

I'm relating all of this because I just acquired a Bronica GS-1 with 50mm 4.5 lens. I bought the kit from a reseller (shame on me at this point), but it seems like it was someone's working rig before it was sold. Like the RB67 I bought locally, it came with a body, lens, and a bunch of accessories, but not all the accessories one would expect. As if someone found it all together in a box and just took it to the camera seller. So the prospects that it would be decent seemed high enough for me to give it a shot.

With just about every used lens that I've returned, there was some kind of mark or scratch. Sometimes fairly obvious, and sometimes very minuscule, but I could always kind of piece together in my head why the lens might have issues, based on the external signs of damage. (Not that external signs of damage would necessarily always indicate optical issues... I have some beat-up-looking lenses that work great.) With the 50mm on the GS-1, it looks about a flawless as I have ever seen a lens of its age. Even better than the 90mm on the RB67. So I had a good feeling about the test roll I shot and developed yesterday.

But the results, while not conclusive, gave me pause. Please, have a closer look. This was taken at f/5.6, handheld.


(click to view full size)

I'm not worried about the soft corners. That comes with wide-angle MF SLR lenses of this age, especially when not even clicked down a full stop. What concerns me is the difference in detail reproduction between the areas on the left and right side (A & B). Kind of reminiscent of the Fuji 100mm, eh? Only it's the left side that's worse in this case. Just barely, though. Both sides are pretty soft, which indicates to me that (hopefully) both sides will sharpen up in a fairly symmetrical way as the lens is stopped down. Still though, I don't think you could say that both sides are equally unsharp.

I had a message exchange with a tech who had worked on a few of the Fuji interchangeable-lens RFs, and he took some pains to assure me that the old Fuji MF RF lenses were so solid, simple, and heavily-built that it would be very difficult to cause one to become decentered without the application of extreme force. So I'm left to wonder if the side-to-side differences are perhaps not related to the configuration of glass elements at all. Maybe the issue is happening at the film plane. Maybe the film isn't as flat on one side as it is on the other.

Here's something I noticed about the negatives from the Bronica test roll.

See how the top and bottom edges of the exposed area are bowed outward? The bowing isn't symmetrical. neither edge forms a parabola. This is from a strip of Delta 100 sandwiched between two pieces of ANR glass, resting on a perfectly flat light pad, checked with a two-axis spirit level. Scans from a flatbed show the same thing.

Is it more likely that my film loading technique is to blame, rather than any kind of optical defects, when it comes to the issues shown in the above scans? Or maybe there are issues with the film pressure plate? I'm looking for any reason to not just return this lens & camera. I think someday I might rack up too many returns, and people will just stop selling to me. But especially since I've seen what 'good', 'correctly aligned' lenses can do, I'm just not satisfied with these sorts of results.

PS: Hate scanning? Don't believe it's a legitimate way of determining anything about your gear? I'll be glad to mail you a negative. I'm serious! And if you paypal me $100 (my time is valuable... although you might not believe that to be the case based on this novel of a post I just wrote), I'll send you a print! I'm serious!

 

[NJH]

You got me curious now to look at my SQ-B and ETRSi negatives and transparencies for this 'bulging'. Have to say I haven't noted any obvious de-centering on the 4 SQ and 4 ETR lenses I have although none of them except the 65mm PS are particularly memorable at F5.6 either, that lens is just stunning. On the film loading is it possible to get it wrong? On my Bronicas you can't really get it wrong if loading the insert away from the back, then install insert into back. Its never looked to me though like the paper is held particularly flat, Rollei may well have had an advantage with this on their 6000 series cameras as the film doesn't have to make 2 such hard abrupt turns in and out of the gate.

I have to say I very rarely have ever shot at these in-between F stops, nearly always either wide or near wide open or stopped right down to F11 or more.

 

[yessammassey]

Loading the GS-1 is a relative cinch. Seems hard to truly mess up, but if I shoot it again I'll try holding tension on the manual winding knob as I slip the back on. I'm running out of cheap b&w film, though.

I'd be curious to know what you see, looking at the SQ & ETR frames. The Fuji 6x9 frame where I am seeing a similar issue have straight edges, so it could be that the uneven edges on frames from the GS-1 are unrelated to the issue. Or maybe the lens is the culprit there.

How much stock should I put in the idea that the field of focus might shift on one side as a lens is stored down? I shot a fence wide open at 4.5, same lens, same roll of Delta 100, and the difference between the left & right sides was negligible.

 

[MattKing]

Not all medium format cameras and backs have a completely unobstructed light path to the very edges of the film gate. In some cases there is material near but not actually on the same plane as those edges of the gate that serve to mask extraneous flare - sort of a window outside the window. If that is the case with your GS-1, the bowing doesn't mean anything particular.

 

[Dan Fromm]

Hmm. So you're shooting handheld near wide open and getting poor image quality, especially off-axis. Why am I not surprised? Especially since your "a" and "b" areas seem to be on opposite sides of the fuzzy plane of best focus,

If you want to find out what's happening, shoot from tripod, use a cable release, shoot at apertures at which the lens covers the format and be very careful about where the plane of best focus is.

Your bad luck with lenses is staggering. How do you test?

 

[moto-uno]

Of late I've had the pleasure (?) of taking apart a number of Fuji Gl690 and Bronica Etr lens . I'll agree with the tech
that you spoke to . The chance of any of these elements being decentered strikes me as borderline impossible . The reason
I suggest this is because every one I've worked on has had the lens fit very securely in it's carrier ( they didn't come loose
from their carrier without noticeable pressure on the lens element ) and the threaded retaining ring pretty much guarantees
locking them into the housing . Every one of these lens has been around the block after this many years . I also tend to be
rather smooth with my film winding and releasing of the winder , it made an instant change in negative spacing with my GL690.
Regards,Peter

 

[yessammassey]

Hmm. So you're shooting handheld near wide open and getting poor image quality, especially off-axis. Why am I not surprised? Especially since your "a" and "b" areas seem to be on opposite sides of the fuzzy plane of best focus,

If you want to find out what's happening, shoot from tripod, use a cable release, shoot at apertures at which the lens covers the format and be very careful about where the plane of best focus is.

Your bad luck with lenses is staggering. How do you test?

Yeah, the lens seemed to be in such nice condition that I didn't seriously think I would find any real issues. I just decided to burn a couple of frames to get an idea of how bad the corners were at wide aperture. The key issue here, though, isn't off-axis image quality per se. It's the difference in image quality between opposing areas of the off-axis parts of the frame. If you look closely at the crop of that tree photo, you should see a large palmetto at the base of a tree near the letter B. Compare that to the trees and grass located about the same distance from the camera, to the right of the letter A. These two areas are at similar distances and angles relative to the camera, but area A is fuzzy, while area be maintains some contrast. This is possible evidence of a lens defect, in my estimation.

I'm not very good about doing rigorous tests until I've already gone out and shot a roll. Then if I notice something that's actually affecting my results, I take it back to a wall of shame to find out if it's me or the camera. For example, here's a shot from the 65mm Mamiya K/L, a lens I bought specifically for landscapes because of it's reputation for sharpness. This is the image is again sort of a throwaway 'test' shot I took at the end of a hike, using f/5.6 because I didn't want to mount then camera back onto the tripod. (I was even so careless as to not properly adjust the GND filter.)


The difference in sharpness between the left and right bottom corners (as almost imperceptible as it may be on a scaled-down Flickr jpeg) prompted me to take it to a wall. Used tape measure and spirit level to ensure that the lens was orthogonal to the wall (with maybe <1cm margin of error), and made sure to set the floating element to the focus distance (I could hear an unusual sound of metal rubbing on metal as I moved it). I shot about half a dozen different views of the wall from different distances, re-setting the the camera each time. They all came back with the same result: In-focus on the right, out of focus on the left. As seen here. (Again, almost imperceptible at this resolution, but check the bottom left vs. bottom right corners.)
RB67 brick wall test 65

I've shot a lot of photos of this wall over the years. It isn't an ideal testing environment, but by comparing results from different lenses I am able to get enough of an idea. I like to think that I'm not too picky about my lenses unless they display a fault that I notice during routine use. I used to have a lens align test chart, but it was destroyed and I haven't invested in another. It was more useful for telephotos, anyway.

...Anyway, back to the 50mm in question. Here are a few other images from the same roll:

This was taken at either f/6.3 or f/7.1:

For reference, here's a similar scene, taken on the Mamiya-Sekor 90mm (one of the 'perfect' lenses) at either f/8 or f/9. Similar scene, but a poorer-quality scan, done with a flatbed.

And finally, here's another one from the 50mm. Same roll as all the others. This was taken wide open at f/4.5. Interestingly, it doesn't seem to show any difference in detail reproduction or sharpness between the left and right sides, unlike the tree photo that I'm so concerned about. Just look at the corrugated metal on the roof; the blurriness increases towards the edges, but in equal measures on the right and left sides. There's no apparent focus shift. Weird. Or good, I guess.

 

[yessammassey]

Of late I've had the pleasure (?) of taking apart a number of Fuji Gl690 and Bronica Etr lens . I'll agree with the tech
that you spoke to . The chance of any of these elements being decentered strikes me as borderline impossible . The reason
I suggest this is because every one I've worked on has had the lens fit very securely in it's carrier ( they didn't come loose
from their carrier without noticeable pressure on the lens element ) and the threaded retaining ring pretty much guarantees
locking them into the housing . Every one of these lens has been around the block after this many years . I also tend to be
rather smooth with my film winding and releasing of the winder , it made an instant change in negative spacing with my GL690.
Regards,Peter

Yes, I try to be smooth with the winder on my GL690. As a Kiev 60 user, I've been well-trained in the careful operation of a film advance lever, as there are consequences for being slapdash with it!

But, since you agree that there is little chance for decentering with the Fuji RF lenses, do you have any ideas about cause of the blur on the right side of first image in my first post (the original post in the thread)?

 

[moto-uno]

That's a good question , does the lens lock securely to the body , with no perceptible movements ?
On your wall test shot , on my monitor , it appears as though the left hand side is overexposed and
not as contrasty (sp) as the right hand side . Peter

 

[NJH]

I did this sort of test in my back garden a few years back with a Leica M8, I then decided it was essentially pointless as nobody with any sense takes architectural type pictures at in-between apertures such as F2.8 to F4 on 35mm or F5.6ish on MF. Like you have seen I had the same observation with a brand new Zeiss 50 Planar M, checked the distances etc. to confirm I wasn't going crazy but one side was slightly further away. At the time after talking about it with the shop I just accepted that its just a side effect of field curvature and lack of Depth of Field at these apertures unless one can get the geometry of the test absolutely perfect your always going to see differences.

 

[NJH]

Sorry forgot the lights in the first top right image looks like coma or flare, or even for the bottom right bit the flare may be from a smudge getting lit up.

 

[yessammassey]

Sorry forgot the lights in the first top right image looks like coma or flare, or even for the bottom right bit the flare may be from a smudge getting lit up.

There's no play; I've checked it on two different 690's. But, that lens does have a pretty good amount of haze and even a few spots of fungus. I'd never thought that it would do anything except lower contrast... But come to think of it, both of the Fuji 65s I returned had scuffs or haze on interior glass as well, though not as bad as the 100 used to take that first photo.

The RB67 65mm (the subject of the brick wall test) and the GS-1 50mm are both about as clean as any used lenses I've ever come across, though. And I think (80% sure) that the floating corrective element in the 65 is the cause of its woes.

That just leaves the 50. Incidentally, it's the one I'm least convinced about. There may be no problem at all, or that tree image could indicate that the left side will always be a little behind the right.

I guess it's time to load up and shoot one more roll with the GS-1, just to make sure.

 

[moto-uno]

Hope it works out for you , as by most reviews over the years those Bronica GS lens were well received . Peter

 

[yessammassey]

Shot a second roll, this time taking extra care to keep tension on the film as I wound the leader onto the spool, and keeping tension as I set the insert in the back.

Not a huge difference in the symmetry of the top and bottom frame edges compared to last time.

I did notice that the pressure plate on the back of the film insert is a tilted a little bit farther outwards on the left side vs the right. I have no idea if it was intentionally manufactured to be that way or not. Both of my Fuji RF's have the same sort of thing going on with their pressure plates, so maybe it is applying even pressure across the frame when the back is closed and the pressure plate is... pressing against the film. Hard to say.

Still working through the roll, but it's looking more and more like the tree photo that sparked this thread was a bit of a fluke. This was also taken at f/5.6 - another real-life conditions test - and the amount of blur and distortion at the left and right edges seems pretty equal between both sides.

 

[btaylor]

I agree with Dan about the tripod. For me, MF slr's and motion blur are a problem. I use a tripod whenever possible, and prefer to pre-release the mirror. I wouldn't trust any lens test without a tripod and mirror up before exposure. I have had a GS1 for years and have found the lenses to be some of the sharpest I have ever used. Curiously, much of the GS1 gear on the used market is nearly mint.

 

[ruilourosa]

have you tryed to print vs scan? bronica lenses are a treat... no problem whatsoever and i have 6 of them for gs-1

flat bed scanners are not the test bench you should use to take any conclusions...

 

[NJH]

People here always say that but honestly putting the negs in the enlarger only makes everything better rather than magically fixing all issues as apugers seem to assume. Thing is though its much faster to pop a negative into the enlarger, lift the head up as far as possible on the column and eye ball or use a grain focuser to inspect the projected image on the base board. I don't know why people don't do that first, perhaps its too obvious?

 

[Neil Grant]

....there's a whole lot of stuff here that I don't have time to read really carefully. I'd just make a few points though:

1. aligning cameras squarely to test targets is extremely difficult.
2. the cameras themselves may not be exactly 'true' anyway.
3. film flatness, especially with roll films, may not be perfect.
4. reflex mirror action may 'suck' film out of 'flat'

 

[yessammassey]

Here is another shot from the same roll. It seems that mobile Flickr won't generate a bbcode sharing link, so you can click, or just take my word for it.

I went over this roll, looking at differences in actutance between objects near the left vs near the right edges. While results were generally acceptable to good for uniformity of focus across the frame for the shots at F/4.5 and f/11, the image linked above (taking aperture somewhere around 6.3 to 8) shows a perceptible shift. Check the two biggest rocks on the left and right sides. The rock on the left side is OOF, while the one on the right is still rather sharp. These rocks appear to be just about equally distant from the camera. It's subtle, but fits with what I've already observed, going back to the original b&w picture of the trees.

The point about using an enlarger's projection rather than a scanner to judge this sort of thing is well-taken. Unfortunately, my enlarger only holds 35mm film. (The intention is to get something bigger someday, but as it stands I'm in a scan & print workflow for medium format). It is definitely not as convenient as using an enlarger; for one thing, if something looks out of sorts, I have to scan it again in a different orientation, on a different section of glass, to confirm that whatever it is is part of the film and not an issue with the scan.

 

[Chris Livsey]

....there's a whole lot of stuff here that I don't have time to read really carefully. I'd just make a few points though:

1. aligning cameras squarely to test targets is extremely difficult.
2. the cameras themselves may not be exactly 'true' anyway.
3. film flatness, especially with roll films, may not be perfect.
4. reflex mirror action may 'suck' film out of 'flat'

I would add
5. Very, very, few lenses commercially available approach anything like a flat field of focus, the 100mm f3.5 Zeiss Hasselblad is a notable exception and is not "perfect".

 

[Sirius Glass]

People here always say that but honestly putting the negs in the enlarger only makes everything better rather than magically fixing all issues as apugers seem to assume. Thing is though its much faster to pop a negative into the enlarger, lift the head up as far as possible on the column and eye ball or use a grain focuser to inspect the projected image on the base board. I don't know why people don't do that first, perhaps its too obvious?

Some people would rather become testinestas that directly get the answer.

 

[moto-uno]

^ "testinestas" , Can't find that word in the dictionary , and hesitate to speculate on it's meaning . Peter

 

[Sirius Glass]

It means someone who would rather do endless testing rather than actually doing anything. When HIE was discontinued I was able to buy two 36 exposure rolls. I asked on APUG how to expose the film for best results. Several self proclaimed testing experts said that I should choose exactly one scene, set the camera on a tripod and expose it from f/1.4 though f/22 and 1/1000 second though 10 seconds, develop the film and then I would know everything I would need to know about the film. Of course following those dunderheads advice would mean that I would never get to use the last of the HIE file to photograph anything. So I started using the term testinestas to describe anyone who preaches the extensive senseless testing with no standards nor any concept how to do scientific testing, just test test test and then more testing. Some Zonestas are of the testinestas ilk, just keep on testing and never going out to use the camera, but boy do they know everything about the film except how to actually use it.

 

[NJH]

Very true, its not just the testers though its those who can't but help make everything difficult for themselves.

 

[yessammassey]

My issue is that I've got until Monday to decide if this thing stays or goes back. Whatever decision I make, I'll have to live with after Monday. No looking back after that.