corner sharpness with digitar 35mm

[gebseng]

Hi everyone,

(I hope at least some of you are enjoying the holidays already!)

As I mentioned in earlier posts, I am shooting architecture with a Linhof Techno view camera, a Leaf Valeo 22 back (36x48mm sensor) and a Schneider Apo-Digitar 5,6/35mm.
Lately, I realised that rising more than 10 mm with the Digitar 35 gave me a strange corner effect: the image became quite blurry, and there was a notable amount of halo in contrasty areas (a sample image and a 100% crop are attached).
So, I sent the Digitar to Schneider Kreuznach, with a very detailed description of that problem and some example files.
They told me that the copal leaf shutter was misaligned and had to be changed, the service cost € 660.-, which is about 1/3rd of the price of the lens.
After receiving the serviced lens I repeated the test shot and, alas, the corner sharpness and halo had not increased in any way.
Now Schneider says that my lens is in perfect condition after the service, and there's nothing they can do about my problem.
The test shot was done with f8 and a rise of 15mm, which does not exceed the 90mm image circle in any way (i calculated that the upper corners are still within an 84mm circle).
So, do you think that this is just something I have to live with, or is the lens faulty?

thanks for your input,

geb

 

[Christopher]

I would say, from personal experience, that f11 should do the trick. I found that some Schneider lenses just need f11-16 to perform well at the edge of the image circle. (I consider 84mm the edge)

 

[vjbelle]

Am in total agreement with Christopher. That lens cannot be shot at f8 with those kind of movements.... in fact I don't think it can be shot at f8. I have found that 11.3 to 11.66 provides optimal sharpness shifting or not. I don't like using f16 because of diffraction.

Victor

 

[gebseng]

Thanks Christoph and Victor, I appreciate your input!

Actually, I made the test shot with apertures 5.6, 8 and 11, but the halo effect did not really change (I normally shoot with f11 for architecture, since even with a 22mp back things get a little soft at f16)

I am not so much concerned with the corner softness, but do find the halo quite irritating.
Is rising up 15mm in portrait orientation with a 36x48mm sensor just too much? I find that I have to do that quite often.

To be fair, the Rodenstock Digaron 35 does not even allow a movement of more than 10mm anyway, due to its 70mm image circle. So, the Digitar seems to be the only choice for me. Or should I consider a Rodenstock Sironar with its 105mm image circle?

I shot architecture with a 90mm Super-Angulon on 4x5" film through most of the 1990s and 2000s, and a 35mm lens is just the perfect equivalent for that on DB.

best,

geb

 

[vjbelle]

Is rising up 15mm in portrait orientation with a 36x48mm sensor just too much? I find that I have to do that quite often.

One of the times I used my 35XL with movements was on one of my trips to Venice. I was using a P45 and never experienced the halos you mention. I don't remember the exact shift but it could have been near 15mm in portrait orientation. A lot of this is sensor and lens hand shaking..... sometimes it just doesn't work out.

Victor

 

[archivue]

"I have found that 11.3 to 11.66 provides optimal sharpness shifting or not. I don't like using f16 because of diffraction. "
i've found the same !

the 35 xl suffers from field curvature... using a focussing ring, yoy can back a balance between center and corners... but on a screen it's difficult...

maybe you should try to shoot several shots with differents focussing distances...

and remember, that it you are at 10 meters from the building, the roof will be maybe at 15 meters or more...

 

[gebseng]

thanks for all your input!
I am pretty sure this is not sensor or lens hand shaking related, since the camera was mounted on a very heavy tripod, and I also tried this with a p45+ back with exactly the same result.
So, I guess the digitar 35 is only really usable with an 80mm image circle.

best,

geb

 

[Woody Campbell]

thanks for all your input!
I am pretty sure this is not sensor or lens hand shaking related, since the camera was mounted on a very heavy tripod, and I also tried this with a p45+ back with exactly the same result.
So, I guess the digitar 35 is only really usable with an 80mm image circle.

best,

geb

My experience is that the high resolution part of its image circle just covers a 40 x 54 mm sensor (67mm diagonal) at f12, particularly if it's an 80 meg sensor. This is consistent with its MTF curves, which are available on the Alpa website. The MTF curves also show that the lens is optimized for closer distances, rather than infinity, which is also consistent with my experience.

The MTF curves show that f 11 resolution drops off very sharply beyond the inner 75% or so of the image circle and even more so at wider f-stops.

I'm using this lens on an Alpa TC, which has no shifts, so none of this matters to me, but photographers looking to shift with a 40x54 sensor should look for another lens. It's an outstanding lens in terms of accutance, presence, freedom from linear distortion, freedom from flair (I've added a couple of step up rings as a lens shade), bokeh, etc. within the good part of its image circle.

 

[gebseng]

thanks for the insight! Has anybody experience with the 35mm Rodenstock Sironar and its 105mm image circle?

thanks,

geb

 

[wentbackward]

It is only recommended for backs down to 9microns pixel pitch. I think it's max useful resolving power is between 30-40 lp/mm.

 

[archivue]

the 35 sironar digital surfers from focus shift... and when you manage to achieve the best focus possible at the definitive F stop... it isn't better than the 35 XL !

 

[Chris Barrett]

I had the Rodenstock Apo Sironar Digi. It was "ok" on the P45+. Not really adequate for P65+ or better. I'm much happier since switching to the Schneider.

CB

 

[gebseng]

thanks again, this helps a lot!

geb

 

[jeffreytotaro]

I have used the 35XL and now the Alpa version, the 36mm Switar. Both lenses exhibit the type of behavior you showed. Any shifting of a lens will create some softness. Looking back to the film days, shifting a 90mm XL too far resulted in the same sort of softness. The issue is not so much the lens, which in most cases is excellent, but rather the idea that we can now check with a critical eye the sharpness at a high degree of magnification. I don't have the math to back it up, but I would say that checking a 4x5 transparency with an 8x loupe is not the same as checking a P45+ file at 100%, which reveals some of these sharpness fall-off issues. Keep in mind that I have been told by Schneider that this lens does indeed have a field curvature greater than some and that too will result in some corner softness, as someone else mentioned here. I think better to have the image (having a lens that can get there with a large shift) than not to have it all. I sometimes put on the 23HR and crop it if I need a large shift from the 36mm, but then I have a smaller file, so which is better, some fall-off of sharpness or a sharper but smaller file? You've got a great lens, go make pictures with it!

 

[archivue]

" I think better to have the image (having a lens that can get there with a large shift) than not to have it all. "
i second this... yesterday i was shooting a big building, i've push the 35XL to it's extreme... anyway, the top corners was plain blue sky... so i was happy to have a large image even if the corners were not so good... who cares for sharpness in plain blue sky ?

 

[jlm]

Something I've wondered about corner softness in general, especially with wide angles and large image circles

Assuming you set focus at some intermediate distance,on a plane in your image, like a building facade square to the camera: the corners are going to be considerable farther from the lens and therefore wouldn't you expect them to not be in such good focus as the center?
when focus is at infinity, the effect would not be there

 

[archivue]

"when focus is at infinity, the effect would not be there..."

and on the top of it, the 35 XL exhibit a field curvature !

 

[gebseng]

Looking back to the film days, shifting a 90mm XL too far resulted in the same sort of softness. The issue is not so much the lens, which in most cases is excellent, but rather the idea that we can now check with a critical eye the sharpness at a high degree of magnification. I don't have the math to back it up, but I would say that checking a 4x5 transparency with an 8x loupe is not the same as checking a P45+ file at 100%, which reveals some of these sharpness fall-off issues.

This is so true, Jeffrey! I did literally thousands of shots with the 90mm Super Angulon XL until 2009, but very rarely had one scanned bigger than 30 x 40 cm…

But again, my initial question was not only about corner softness, but also about the halo that shows up on contrasty edges in my example. Are you sure this is just part of normal lens degradation close to the edge of the image circle?

geb

 

[etrump]

This is so true, Jeffrey! I did literally thousands of shots with the 90mm Super Angulon XL until 2009, but very rarely had one scanned bigger than 30 x 40 cm…

But again, my initial question was not only about corner softness, but also about the halo that shows up on contrasty edges in my example. Are you sure this is just part of normal lens degradation close to the edge of the image circle?

geb

Just FYI: I shoot IQ180 with the rodenstock HR23 and HR32 and they exhibit very similar behavior wider than f/11. The P65+ had similar effect but it was rarely bad enough to affect a printed image. With the IQ180 it is noticeable under close inspection of larger prints.

 

[jlm]

FYI, from Alpa, re. shimming:

read the second paragraph about focusing the center of the 35 past inf to compensate for curvature of field

http://www.alpa.ch/en/glossary/phot...-the-adjustable-adpaters-for-digital-use.html