Canon LTM Canon 35/2.8 Corner Sharpness

Canon M39 M39 screw mount bodies/lenses
S

Sean Reid

Guest
I've just finished the large test of Leica and Voigtlander lenses on the R-D1 and have had a chance to compare my (recently purchased) Canon 35/2.8 (older, all-metal version) with the Leica 35/2 ASPH, Voigtlander 35/1.7 and Voigtlander 35/2.5. Naturally, the lenses differ in terms of contrast and tonality. My Canon, however, is noticeably softer in the corners than the other three, especially on the left side of the frame. Is that a common characteristic of this lens design or is it my example? I'm thinking that I might buy this newer version of the lens if it might do better in the corners: <http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=3862333799&ssPageName=STRK:MEBI:IT>

I know the Canon 35/2 is regarded as being a "better" lens but if I'm not mistaken, it has more contrast than the 35/2.8. If so, it's not really what I'm looking for. Anyone have experience with the older and newer 35/2.8s?

Thanks,

Sean


BTW: For those interested, the lens review article should be announced and linked on Luminous-Landscape.com some time this evening.
 
I look forward to the article Sean. I think I saw your comments somewhere that the 35 Asph Summicron captured more contrast than the sensor in the R-D1 was able to handle, which is what is driving your comments about the Canon 35/2 I assume.
 
Hi Rover,

That's right about the Leica (under very contrasty lighting conditions). For a lens with higher contrast than the classic Canons I'd likely just use a Voigtlander 35/2.5. It's the unusual combination of lower contrast and high resolution across the frame that I'm interested in right now. BTW, the Voigtlander 28/1.9 is a wonderful B&W lens.

Cheers,

Sean
 
If one corner is bad, it sounds like an alignment problem. Does the lens show signs of being serviced?

I believe that the optical formula of the 35mm F2.8 did not change through its production, but the Black Lens is pretty.

In my 5x7's, I do not notice the edges being "too soft", but have not run a scientific check. It is much sharper wide-open, especially at the corners, than the 3.5cm F3.5 Nikkor.

JOE: Have you received results back yet?
 
Brian,

Alignment was my thought. I'm not sure if the lens has been serviced. If this lens formula is normally sharp in the corners (even wide open) I better try another example.

Thanks,

Sean
 
brian,

i have not received the lens yet, the post office is still closed up here, no deliveries till tomorrow.
my p is 'out there' still also.

sean, i look forward to reading your comments on the lens.

i have to admit that i'm not too concerned about soft corners as i feel it will reflect similar attributes of the user ;)

joe
 
I just looked more closely at some flat field tests and the softness is in fact in the corners on both sides, so alignment seems less likely. There's a dramatic difference in the corners between the Voigt 35/2.5 and this example of the 35/2.5

Cheers,

Sean
 
Hi Joe,

It may be that the softness in the corners is not general to the lens. I can say with confidence that the 35/2.8 makes a beautiful Black&White negative/file.

Cheers,

Sean
 
I will look more closely at some of my negatives. My "test shots" were of Autumn Leaves, so if the "plane of Focus" is off (Petzval?, focus is on a slight sphere?) I will find "something" in focus, where a flat object will be soft.
 
Thanks Brian. I do a simple flat field test of every new lens I buy by photographing newspaper stapled to a bulletin board so that it is flat. I mount the camera on a tripod and square it off to the bulletin board, focus and then make a test shot at each major aperture.

Cheers,

Sean
 
Agree with Brian... It seems to me that a flat field is useful primarily for flat subjects; important for enlarger lenses and copy lenses. But not particularly important for photographing a 3-dimensional world. It's possible the softness you're seeing in the corners, Sean, is largely due to curvature of field, and that real-world subjects may be sharper in the corners than a flat-field test would indicate.
 
sounds like a concern i had about the summitar a while ago. a webpage noted it had blurry corners even stopped down at infinity. most people said they didn't experience such problems. maybe it's a sign of amateur lens repair and not of a weakness in lens design.
 
Actually, a flat field test is quite important because it eliminates some variables that otherwise would make it difficult to compare various lenses' degrees of sharpness across their Fields of View. Lenses that don't do well in flat field tests will show the same tendencies with any kind of subject but those tendencies can be hidden or exacerbated by the nature of the subject matter, by DOF, etc.. A flat field test is more demanding of a lens' performance than are many other tests of mixed three-dimensional subjects because there are fewer mitigating variables that can hide weaknesses in the lens.

Lenses that I've tested that excelled with "real world" subjects, (such as the 21, 28 and 35mm Leica ASPH M lenses I recently tested) have also done superbly on flat field tests. For example, the Voigtlander 35/2.5's performance in the "train station" test pictures for the current LL review is consistent with its flat-field results. That said, blurry corners at larger apertures don't necessarily doom a lens but I would prefer that each of my own lenses be sharp across it's FOV.


Cheers,

Sean
 
I certainly agree that a flat field test is absolutely required for a critical lens test. I am not very critical of my lenses at home, especially the old ones designed without the aid of a computer. At work, different story.

Glancing at your results, it looks like the Canon 35 F2.8 had the least amount of "clipping", ie preserved the features at each extreme of intensity.
 
Last edited:
That's right. In practice, a lens like the 35/2.8 expands the effective dynamic range of the R-D1 by compressing the contrast range of the information sent to the sensor. That's why I think there's good reason to consider high-resolution, lower-contrast lenses for digital rangefinders.

Cheers,

Sean
 
all this talk about the 35/2.8 is making the wait for mine even more unbearable than normal.

i can only hope the mail is full tomorrow.

joe
 
I've now had a chance to use the Canon 35/2.8 for a few days and while its not as sharp as the Voigtlander lenses (much less the Leica 35/2 ASPH) in the corners wide open (with most real-world subjects), its still doing very well. I recommend it for the R-D1 and it is going to be my standard 35mm lens on that camera because of its contrast range and overall sharpness (except in the corners at or near wide open). The character of the lens is quite nice.

Cheers,

Sean
 
Last edited by a moderator:
sean,
you have me in a lather here about this lens.
a good thing it showed up today! and my mint canon p showed today also.

this lens is so tiny, i think it's smaller than the cv 35/2.5 but substantially more heft to it.

joe
 
Hi Joe,

It's tiny. If you end up with a good example, hold onto to it because it's a very good lens - all things considered. My 28/2.8 arrived yesterday but it's a real dissapointment - minty but very soft at the edges and corners. I'm sending back and will use a Voigtlander 28/1.9 and then add a Canon 28 when I find a really good example.

Enjoy,

Sean
 
You must log in or register to reply here.
Back
Top Bottom