DPReview Updated Studio Scene With Sony a7RIII and Results Aren’t Impressive

DPReview updated their studio scene with results form the Sony a7RIII and personally, I find the high ISO results less than impressive. I always like to go straight to the more fibrous/3d parts of DPReview’s studio comparison when looking at how a camera performs and while base ISO looks pretty good I wouldn’t use the a7RIII above 6400 ISO. By 12,800 ISO the a7RIII loses too much detail, but there might be a slight issue with DPReviews test if you pixel peep around. At times the a7RII looks better than the a7RIII, while the D850 and GFX almost always look a little better. Maybe the small Sony a7 is a little off axis during testing or maybe this is an optical weakness.

Some sites might try to paint a rosy picture here, but the a7RIII does seem to be performing below the a7RII in some respects and it is very hard to tell if the testing is flawed or the camera. Thankfully the performance claims of the a7RIII do seem to live up to expectation from the videos that we have seen on Youtube, but the a7RIII won’t be replacing the a7S anytime soon.

Join our Sony a7 Owners Group, also follow SonyAddict on Facebook and Twitter.

Sony a7RIII: B&H Photo / Amazon / Adorama
Sony 24-105mm: B&H Photo / Amazon / Adorama

This entry was posted in Sony a7RIII. Bookmark the permalink. Trackbacks are closed, but you can post a comment.
  • Ayoul

    They have changed the lens used for the studio shot between A7RII (55mm) and A7RIII (85mm 1.8). They don’t perform the same, and they even don’t have the same DOF at f/5.6.

    Anyway, to make a proper comparison of the camera rendition of different cameras using different lenses, I strongly advise to not go far away from the center of the picture, where lens differences can start to make a difference very quickly, even closed down.

    This scene is only useful for the noise, not for the details. There is simply to many variables in the equation to judge sharpness with this kind of shot.

    • That’s kind of the point of their comparison though… To let you just the camera across the entire FOV. I don’t really find the noise performance impressive either, but the above also demonstrates the problem with long running “technical” testing. Reid is probably the only person with semi reliable technical testing, but you have to pay for access. The 85mm Should be sharper than the 55mm. Also the hair grab isn’t that far off center… If you look at both brushes they show different amounts of detail too, which is why I think they have an axis error.

      • Ayoul

        The 85mms are traditionnaly a little bit sharper than the 50/55s, but I don’t think that it’s the case with the 85mm FE vs the Zeiss 55mm. And there is still the DOF difference. Where is the map done? probably in the center, on the flat B&W circles. Every three dimensional objects are affected by DOF at f/5.6, 42 Mpx and at this distance. Sony has already stated that the sensor was the same. I highly doubt that hey have changed anything that could make him less sharp with their own lenses (like filter or microlenses).

        I don’t know if details is “the point” of their comparison in the mind of dpreview staff, but I’m sure of one thing : you can’t rely on this kind of shot for sharpness. I see a small improvement in noise at high isos compared to the A7RII, but that’s all I can conclude. There is no way that someone can make a proper comparison of the details rendered by the A7RII vs A7RIII with this shots, which are done with different lenses, and worse, with different FOV and DOF.

        • Part of doing a noise test is seeing how detail is preserved. Especially since almost everyone is processing noise even in their RAW files. So even if you’re not going to judge these tests for sharpness you should judge them for how much detail is preserved as the ISO rises…

          • Ayoul

            Yes, the famous ratio signal/noise, I know that. But it just doesn’t work. Not with different lenses, especially on a short distance with a (not so) flat target. Lens quality, lens fov and dof, lens variance (maybe a faulty lens as you suggest), flat scene with 3d items glued on it. It doesn’t meet the basic requirements to judge this. I know that it’s what it’s supposed to be, or at leat what people think it is, but imho, it just doesn’t work.

  • Kenny Ho

    Looks like Sony ‘overcooked’ the raw files…