35mm equivalent shootout: A7 vs NEX-7 vs EOS-M

Discussion in 'Other Systems' started by Jefenator, Dec 30, 2013.

  1. Jefenator

    Jefenator TalkEmount Top Veteran

    Nov 23, 2012
    Oregon, USA
    Real Name:
    The three contestants are:
    • Sony A7 with Leitz Summicron-R 1:2/35 (Made In Canada, early '70s vintage)
    • Sony NEX-7 w/ SEL24
    • Canon EOS-M w/ EF-M 22mm pancake lens

    Here's some full test images:

    A7 w/ Summicron-R 35 @ f/5.6 35 @ f/8 by Jeff Addicott, on Flickr

    NEX-7 w/ SEL24 @f/8 by Jeff Addicott, on Flickr

    EOS-M w/ 22mm @ f/5.6 by Jeff Addicott, on Flickr

    With some 100% crops from the mid upper left:

    35-shootout-upper-centers by Jeff Addicott, on Flickr

    Is it just me or is the NEX-7 / SEL24 combination getting beaten up a bit at the pixel level?

    Well, when you get an unexpected result, best thing is to repeat the test. Here's the upper right corner from a different series of shots (focused specifically on that corner):

    35-shootout-corner by Jeff Addicott, on Flickr

    I must say, I really did not expect the A7 with a 40-year-old lens to be leading so comfortably, so repeatedly! Here's a center crop from a test scene at about 5':

    35-shootout-center by Jeff Addicott, on Flickr

    And now the part of the comparison where I sort of expected a bloodbath: the BOKEH test!

    A7 w/ Summicron-R 35 @ f/2 by Jeff Addicott, on Flickr

    NEX-7 w/ SEL24 @f/1.8 by Jeff Addicott, on Flickr

    EOS-M w/ 22mm @ f/2 by Jeff Addicott, on Flickr

    A7 w/ Summicron-R 35 @ f/4 by Jeff Addicott, on Flickr

    I repositioned the EOS-M to compensate for the wider lens & smaller sensor. Even with a slight unfair advantage, it trails the pack for BOKEH potential. The full-frame rig looks about the same, stopped down two stops!

    So there you have it! A few extra notes:

    • The Summicron-R is pretty rock solid for the most part, but I did observe a tiny bit of smearing in the extreme outer corners at f/8. The color response seems noticeably different. And finally, be warned: it is really freaking heavy!
    • The EOS-M is very nice optically and very compact - I routinely throw it in a jacket pocket. This comes at the cost of very slow AF performance and a very awkward interface. Much more frustrating to actually work with than either of the Sonys IME.
    • Like Like x 15
  2. Bugleone

    Bugleone TalkEmount Veteran

    Aug 21, 2011

    What DOES appear to come out of this is the poor realative value of the NEX7/24mm...this is still nearly a THOUSAND POUNDS here is the UK !!!!....the canon is clearly a much better image maker and can be bought in any shopping centre for about £300 or sometimes less!....

    ........Even the A7/summicron is about a £3000 rig!!!.....so, very nearly as good for ONE TENTH the price!!!!!!

    I have many interests and can't justify three grand for a camera,....especially one that does NOT demonstrate absolute superior image quality. I don't especially like canon's business ethics in the past and have not bought any canon products since 1984, however, I dislike Sony even more and my NEX is the only sony item in my house......

    ..........Think I might take alook at the canon.

    EDIT:It would be interesting to see what the canon could do while wearing the summicron.....
  3. radimere

    radimere TalkEmount Regular

    Apr 2, 2013
    Poor SEL24 sample?

    Interesting comparison! Thanks for taking time out for this. :)

    I wonder how much of the SEL24's softness is due to CA? Also, the closeup shot seems awfully soft for ƒ/4. Not doubting your methods, just doubting the quality of your copy. Mine is tack-sharp at ƒ/4.
  4. Armanius

    Armanius TalkEmount Regular

    Aug 8, 2011
    Houston Texas USA
    Interesting comparison. The Cron looks really good. I liked the Zeiss the least.
  5. Jefenator

    Jefenator TalkEmount Top Veteran

    Nov 23, 2012
    Oregon, USA
    Real Name:
    ISO comparison

    I should probably include this, since it was done with the 35mm equivalent lenses.

    Here's the ISO test scene

    A7-ISO-TEST-2 by Jeff Addicott, on Flickr

    And here's some 100% crops at various ISO settings

    A7 NEX-7 EOS-M ISO Comparison by Jeff Addicott, on Flickr

    The A7 is definitely cleaner, with a higher usable range. Both APS-C cameras have some noise even at ISO 100. I guess that's part of the full-frame advantage!

    The Sonys tend to get banding towards their maximum ISOs, which comes out very ugly if you try and push the file any farther. The Canon files seem more resistant to that, all the way up to 12800.
  6. lenshoarder

    lenshoarder TalkEmount Veteran

    Feb 7, 2012
    Well, at least in that first series of crops, it's hard to tell. They were taken at different times with different lighting. The thing that strikes me most about the NEX 7 shot is that the tree is much brighter than the other two shots. You took that one first didn't you? It looks like that was taken first, the sun moved, the A7 shot was next, the sun moved and the EOS M shot was last. You can really tell looking at the leaf. It's brightly lit in the NEX 7 shot, somewhat lit in the A7 shot and completely dark in the EOS M shot.
  7. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    As for the Zeiss - in the first bunch of crops, the NEX-7 had by far the most difficult lighting situation, so it's not close to being a fair comparison. In the second bunch, the Zeiss clearly looks less sharp than the others due to the high amount of CA in this worst case scenario - yep, CA are the weakest point of the Sonnar. In the third bunch, the Zeiss got to be slightly misfocused - I shot more than 30.000 frames with my Zeiss and not one of them was that bad. As for the bokeh - looks to me like the Zeiss has the best bokeh, but it's VERY hard to tell with this scene.

    Also, keep in mind that the 24mm Sonnar, while Zeiss claims it got a floating lens group, is far better in mid to close-focus situations, not necessarily the best when focused to infinity. This is an advantage for many, as the lens will without doubt be used quite a bit for portraits, street photography and other similar work. For me, for example, it's a disadvantage, shooting mainly landscapes with it. But even then, even wide open (like 90% of my landscapes are shot), it's able to render the color and shadow transitions better than any other lens I've used to date. It's not all just about sharpness, you know?
  8. radimere

    radimere TalkEmount Regular

    Apr 2, 2013
    Agreed. Not rationalizing my purchase by any means, but the Sonnar is my favorite lens, even next to Nikkors, and it's because of the tonal smoothness, the contrast, the lack of color cast, and the short DOF falloff. The short MFD, compact build and balanced weight are great, too. The Sigma 30 f/2.8 might be a tad sharper (not sure) at f/2.8, but it just churns out sterile images compared to the SEL24 (sterile is not necessarily bad—it's great for architecture, interiors and cars).

    Anyway, as long as a lens is not an absolute dog, downsampling to ~16MP and applying sharpening negates most sharpness differences between lenses.
  9. Jefenator

    Jefenator TalkEmount Top Veteran

    Nov 23, 2012
    Oregon, USA
    Real Name:
    Those first shots were taken as close together chronologically as possible. (1:26, 1:28 and 1:38 for the NEX-7, A7 & EOS-M respectively.) I think much of the variation bight be from the auto exposure and WB responses of the different cameras. (Though I'll be the fist to admit: getting consistent lighting for outdoor testing is a pain in the butt.)

    I tried time and time again to make the Zeiss 24mm appear as sharp as the others at the pixel level, and could never quite do it. A previous comparison between the NEX-7 and EOS-M yielded similar results.

    I think that both the A7 and the EOS-M have an unfair advantage with larger pixels. It's not just the SEL24 that has been suffering by comparison - the A7 seems to be easily running circles around the NEX-7 with every A/B I've done so far. (Summicron 50 vs Sigma 30 and Nikkor 24 vs Sigma 19.) I hate to say it but this could just be the nature of full-frame vs APS-C. I'm not quite done testing, though…

    I am by no means trashing the SEL24 - I love the overall character, AF performance kicks major butt over the EOS-M and in terms of sharpness, it definitively outclassed the slower legacy 24mm lenses I compared it against on the NEX-7.
  10. Poki

    Poki TalkEmount Hall of Famer

    Aug 30, 2011
    Auto exposure and auto WB might not be a good idea for a comparison to start with. As for pixel size - sure, it's a disadvantage for the 7, and while it should easily outperform the Canon when downsampled, I think compared to the A7, it's pretty fair. It shows the small but still noticeable advantage a full frame camera has.

    Maybe a test with fixed settings and less extreme contrasty situations might be worth a try. Comparing a lens with high CA to lenses with less CA will always make the lens with more CA look the worst in such conditions, no matter how much better it might be in pretty much all other categories (not saying it is - just pretending).
    • Like Like x 1
  11. radimere

    radimere TalkEmount Regular

    Apr 2, 2013
    Photographically speaking, such scenes are undesirable to shoot, anyway. :) So they rightfully belong only to academic testing, with few exceptions.

    Sensors with larger photosite pitches (e.g., the A7) will always be more forgiving of lens resolution at the pixel level than one with a smaller ones (NEX-7), all things being equal. It'd be interesting to see a comparison between the Summicron and SEL24 on the same body. There was one from way back comparing the SEL to a 'Lux, both mounted on a NEX-7, and I think for a cheaper enthusiast lens, the SEL acquitted itself quite well:


    That said, try flicking your eyes back and forth between the A7/Summicron and NEX-7/SEL24 image in the OP—the latter has better contrast, and perhaps acutance. If you don't pixel-peep, that translates to better perceived sharpness across the scene. You can bump up the clarity/local contrast in post, but it's difficult to finesse without making the final image look artificial, especially where portraits are concerned.

    In short, I prefer a lens that "renders" [I hesitate to use that term because cork-sniffers love it] well to one that just reproduces sharp edges. :)
    • Like Like x 1
  12. Amamba

    Amamba TalkEmount All-Pro

    Apr 13, 2013
    SE MI
    EOS M + 22 is a very good and cheap combo. I actually like Canon sensor, while it's behind Sony in DR, it's still a solid performer and I think their Jpeg engine is better.

    But EOS M is just not a compelling camera for me. I did look at it when it came out, and was disappointed. There's no built in flash, so it's pretty useless for a portrait shooter (it's a must for fill in lighting, and if I have to carry the external flash with me at all times, it defeats the idea of a portable mirrorless). It doesn't have an EVF nor any means of attaching one. It doesn't have an articulating screen so you have to keep the camera in front of your face to frame and focus. Try doing this with a large tele lens. Speaking of lenses, it only has three native lenses (AFAIK) so you need to get an adapter and use Canon DSLR lenses (or legacy) for anything else. The AF is slow, it was very slow when the camera was first released, and is better now after a firmware update but I don't think it's on par with Nex 6 yet.

    Finally the price... It's only cheap because it was a marketing failure and went on fire sale, at least in the US.
    Nex 7 seems to be a very demanding body. I believe a more fair comparison for EOS M would be with Nex 6 and Sony 35/1.8 or Sigma 30/2.8. I think the results would be much closer.

    I actually like Canon, and if they ever decide to make a wholehearted attempt at a mirrorless, I'd be very tempted to go back to them. But EOS M to me is a half a$$ed camera system, and the EOS M2 isn't any better. They seem to be afraid to create a really good product that would steal from their DSLR sales.

    Sent from my Nexus 7 using Tapatalk
    • Like Like x 1
  13. Jefenator

    Jefenator TalkEmount Top Veteran

    Nov 23, 2012
    Oregon, USA
    Real Name:
    The EOS-M has been fun to have and has bagged me some nice keeper shots by virtue of being so small & handy.

    I must concur, I would never pay the full intro street price and it would not be my main shooter.