3 of 15 people found the following review helpful
Generally great camera.,
Verified Purchase(What is this?)
This review is from: Sony a7 Full Frame Interchangeable Lens Camera Body Only - Black (24.3MP) 3 inch LCD (Electronics)
From the professional reviews I read online, this camera meets most of my expectations (especially consider Steve Huff). Poor corner performance with even 35mm rangefinder lenses, and poor auto white balance (over sensitive to green, it seems), let the camera down somewhat.
Tracked by 2 customers
Sort: Oldest first | Newest first
Showing 1-9 of 9 posts in this discussion
Initial post: 10 Jan 2014 12:59:54 GMT
The 2 issues with third party lenses are related to focal length and back projection distance, and how this impacts on the sensor. The relatively short back projection distance of Leica M mount lenses can give rise to imaging issues at the edges of the frame when using lenses wider than 35mm in focal length. This is due to the angle of incidence of the light rays at the edges of the sensor and which hit it at quite an oblique angle. I've used my 15mm Voigtlander Super Wide Heliar and the drop off in image quality is very noticeable. But this was so, even using it on my M6 film camera. What the A7 is showing me is just how sharp this lens is in the centre, which has surprised me. Film wasn't revealing it at is best.
However, my intention is to use my Leica R reflex lenses and other brand reflex lenses and as these have a much longer back projection distance I am not anticipating many issues. I have a 24mm Elmarit and I am really looking to see how this performs. But I certainly won't waste my money on anything wider. From 35mm and longer, there shouldn't be an issue using reflex lenses on an A7. If edge performance at optimum opertures of around f5.6 when the centre is sharp are more likely to be down to the cheap adapters that many, as me, will be using. Tolerances in manufacture are more likely to be an issue with most of these cheap items.
Posted on 4 Feb 2014 13:30:31 GMT
Brian Hrodgar says:
In your laconic review you claimed '...poor corner performance with even 35mm rangefinder lenses...' Just to say that in my limited experience I´ve had excellent corners with a Leica Summarit 35mm and likewise a Contax 28mm F2.8 G. Having deuteranopia red/green colour blindness I can´t comment on the AWB.
In reply to an earlier post on 2 Apr 2014 21:34:10 BDT
It is noticeable that wide lenses for these cameras have huge rear elements. This is a clue that they are nearly 'telecentric' on the sensor side. Looking in the back of the lens, the deeper-inset the aperture appears to be, the better. SLR lenses will have mirror-clearance by definition : it's really the rangefinder lenses that could be disappointing.
Thinking of the magenta cast to the corners, it's only a matter of time before people come up with plug-ins to correct it - maybe they already have ?
In reply to an earlier post on 3 Apr 2014 00:07:23 BDT
Last edited by the author on 3 Apr 2014 00:14:03 BDT
Whilst true for a number of w/a lenses, they don't necessarily have huge rear elements. Indeed, my Leica rangefinder mount Russian f6/28mm Orion-15, a copy of an older Zeiss Biogon design is a tiny lens. This is an inherently slow lens, but at f8 I am truly surprised at its extremely good performance across the frame. I'd expected my Kiron f2/28mm in Minolta MD mount to be an excellent performer and beat it given its reputation with film, and being a retro design, but it is only when it reaches f8, that it begins to show what it can do, but it is just fractionally behind the Orion at all apertures, and which at its open aperture of f6 is noticeably sharper than the Kiron at f5.6. Verdict: in normal shooting, the Orion by far.
Leica R mount f2.8/35mm. Unexpectedly poor wide open, noticeable improvement central definition and sharpness at f4, edges lagging behind, but at f5.6 wonderful central sharpness and central definition with edges only fractionally behind, but really only noticeable at 100% on screen.
Zeiss f1.5/50mm Sonnar for older 1960 vintage Contax IIIa r/f camera. Truly super definition and sharpness in centre, but over quite a restrictive area, about 1/3rd of the frame. !/3rd either side very poor and never catch up. A disappointing result for a lens that performs superbly with film. Best used on APS-S sensors.
Leica f1.5/50mm Summarit. An old rangefinder design from the early 1950s'. Very poor throughout producing a strange purple circular haze in the centre of the image as it is stopped down from f8 onwards becoming more clearly delineated the smaller the aperture.
Minolta f1.7/50 MD. Very good central sharpness even at wide open at f1.7, edges slightly behind, but by f4 excellent centre with very good edges, and at f5.6 truly excellent across the whole frame.
Canon EF f1.8/50 Mk II, the current model. Central definition matches the Minolta throughout the range, but betters it for edge sharpness which reaches very good by f2.8, and gradually improves to optimum around f5.6, bettering the Minolta throughout. Verdict: A difficult choice between the Canon and Minolta. I finally chose the Minolta for its overall image quality which had a more natural feel to reproduction, whereas the Canon looked a little clinical.
Leica f4/90mm Elmar. An old rangefinder design. Very nice imagery as opposed to absolute sharpness. Surprisingly tactile images. Don't worry about its slight drop in overall sharpness, this performs admirably.
Leica R f2.8/135mm Elmarit. Superb definition throughout, little change in visual performance from f2.8 to f8. I suspect the sharpness of this lens is limited by diffraction. Heavy lens, and doesn't balance too well on the A7 body. Better on a tripod or monopod, but be very careful as the lens does no have its own tripod mount, so get an adapter that incorporates one, as the A7 body will not be able to take the strain.
Olympus f3.5/f4.5 35-70 zoom. Tested at 35/50/70mm settings and produced very good performance throughout for a zoom, no doubt as it is not beng stretched at 2x ratio.
Canon EF f3.5/f4.5 28 to 105mm MkII (specifically the one made in Japan, not Taiwan). This is from Canon's mid-range falling between the cheap entry level zooms and the L series optics. Performance was very surprising overall and is my "go to choice" as a walk around lens when I want versatility as opposed to fast aperture. In many senses, its performance is so close the the Minolta MD50mm that in a blind test I think it would be difficult to separate them at f5.6. This lens is that good.
So there you have it, a miscellany of lenses, some work some don't. The problem is you won't know until you try it, so making a blind purchase a bit of a gamble.
In reply to an earlier post on 22 Apr 2014 07:41:00 BDT
Last edited by the author on 22 Apr 2014 08:07:31 BDT
Sorry, I didn't make myself sufficiently clear.
"The problem is you won't know until you try it, so making a blind purchase a bit of a gamble."
This is precisely why I said "wide lenses for these cameras have huge rear elements".
The emphasis should have been on "for these cameras".
The lenses that have been specifically made for these cameras have large rear elements.
One way to tell, just by looking at any lens, whether it will work well with this, is by the size of the rear element.
Of course, there's no guarantee, but it's a useful rule-of-thumb.
How well does the theory match your experience ? I can't afford any of these nice things !
The issue is that light hits the corners of the sensor at a shallow angle.
The sensor is designed for a specific angle, so a fully tele-centric lens would probably be overcorrected.
As you say, retro-focus, or inverse-telephoto is the way to go.
You say "strange purple circular haze in the centre"
but purple edges are a more common problem, even on a smaller sensor
In reply to an earlier post on 22 Apr 2014 13:00:53 BDT
Now I follow the point you were making.
I haven't acquired any of the (expensive) Sony lenses for my A7, so I can't comment about how large their rear lens element is. It was Olympus back in the mid-2000's who pioneered properly computed telecentric lenses for their digital 4/3rds sensor. The idea was much poo pooed at the time by Canon and Nikon who were still relying on their film lenses. But the results from the Olympus 14-54mm Pro lens certainly proved their point. And Sony did the same with their R1 bridge camera of 2005, which has an APS-C sensor and a superb Zeiss lens of which the rear element is only 2mm (yes, 2mm) from the sensor. Dpreview reviewed the R1, and if you read this, it is still on their site, you will understand why I sold my Olympus E500 with the Pro lens and bought the R1.
You ask does the theory match my experience? Whilst I understand why telecentric lenses will have large diameter rear elements, I believe there is more going on with optical designs than we may appreciate, and I would have to say that the jury is out (with the A7) on this in my experience. For example, the Orion-15 28mm r/f lens is a small lens in that both front and rear elements are small, but it performs extremely well in my A7. However, the older (1960) Zeiss 50mm Sonnar has a rear element larger than the Orion, but despite this and its longer focal length it is not providing the same overall coverage that I got with film.
Thanks for the links. The purple fringing shown is certainly by far the worst I've seen. Having a snow scene will certainly exaggerate the effect as it is difficult to get pure white snow as the blue sky is reflected by it, which our eyes/brain overlook as we interpret snow to be white. I do wonder if this has a lot to do with Sony cramming 24 megapixels on its APS-C sensor? I still have, and use, my 16 megapixel 5N, and do not experience anywhere near these results, and I do have the same Heliar lens. However, using film lenses on APS-C sensors does hide how poorly they can image at the edges, and which the same lens used on my A7 will readily show up. Much of what I commented on above with my A7, would go unnoticed on my 5N, apart from the very strange circular magenta caste exhibited by the old f1.5/50mm Summarit and which is dependent upon aperture settings below f8. This is really odd and I can't find an explanation for it. Using the 5N means I can fit the Sonnar and make use of its superb central definition and overall imaging properties.
I suspect a lot of purchasers of the A7 will have had in mind also using legacy 35mm camera lenses, but unlike with film, the results can't always be guaranteed on the A7. As a generality, I'd say caution will need to be exercised with r/f wide angles, and possibly 50mm, but from 90mm up should be OK. With slr lenses, they should be more compatible due to their retro-focus designs.
In reply to an earlier post on 30 Apr 2014 20:23:20 BDT
Last edited by the author on 30 Apr 2014 20:28:33 BDT
I note your Orion is an f6 max lens - that might make things easier.
When you say "aperture settings below f8", I presume it is wider-open apertures that are the problem.
Someone should compile a list of which lenses show problems.
would be an obvious place, but it's very quiet there.
There is software called CornerFix that should help.
I expect DxO will also come up with a correction if demand is sufficient. I can't see them profiling many old rangefinder lenses though.
Seems they conclude images will always need manual tweaking.
In reply to an earlier post on 30 Apr 2014 22:44:05 BDT
There are potentially 3 problems with the edges and corners, and all but one of them is correctable to a greater or lesser degree in imaging software programmes. Most now have generic corrections specifically for digital images.
The correctable problems are colour fringing and vignetting, the programme I use handles this. The one that cannot be corrected, and is the one most most likely to arise when using lenses designed in the days of film on a digital sensor, is a softening (blurring of the image) at the edges and corners and can be expected to be worse with rangefinder lenses because the light rays reach the edges of the sensor at a more oblique angle. However, as noted with my Orion-15, it surprised me by being usable, and this is only partly due to its smaller apertures. I simply wasn't expecting it to work at all. The only comparable focal length I have in an slr lens is the Kiron f2/28mm. If the argument that retro designs, because they have a physically longer flange to sensor distance, should perform better, this expectation was not realised, so much so that at f2 to f4, the lens is basically unusable, just passes muster at f5.6, and OK at f8. The only advantage of the Kiron is having the f2 aperture to aid focusing. So, in answer to your question, wider apertures are likely to be the major concern except, in my tests, of medium tele and longer focal length lenses.
What my tests have revealed for me is that assumptions can't be made beforehand as to whether one type of lens and/or focal length will actually work. I'll add to the list of lenses tested as and when I get an opportunity, then perhaps I will look to making the results more generally available. Even then, it can only be a tiny cross section of lenses available. It won't for example, include from Nikon or Pentax as I don't own any, although I have 3 from Leitz still to try, and a few from Fuji, Praktica, Tamron, and Zeiss Jena still to check out.
In reply to an earlier post on 7 Aug 2014 03:47:39 BDT
CornerFix now also seems to be at
BH PhotoVideo suggest the A7R corners (High Resolution model) are worse : A7S (High Sensitivity model) is better - see 'Vignetting Examples'
‹ Previous 1 Next ›