rogue.guineapig wrote in post #18711003
OK so I did a little sleuthing because Tony is probably busy with Pre-Photokina stuff, and I had some questions.
He made all the files available, so I downloaded all of them, opened them up in ACR and PS.
Observations:- On EOS-R image 7369 I started seeing banding pop up with a mere +1.00 on the exposure sliders in ACR.
- In EOS-R image 7382 it's also visible with a decent little bump to shadows and a half-stop exposure bump. Neither of these seem like real hard pushes to me.
- With Tony's brick walls and textures I think it helps mask the issue but if you look at the solid dark gray posters it seems more apparent. I'm using a Retina 5K screen here for reference.
But then there was a question of glass:- The Canon systems shot with the 70-200 f/2.8L II IS
- The Sony A73 shot with the Sony 70-200 2.8 GM OSS
- The Sony A7Riii shot with what photoshop file data calls the "DT 70-200 F2.8 SAM" which I believe is the Sigma 70-200?
In all cases Tony shot wide open, which (to me) would make the glass more obvious and less "transparent" (softness, fall off, etc) in the test shots that are supposed to highlight the sensor, not the glass.
The A7iii looked just abysmal in that test so I did a bit of thinking. Some of the comments said, "yeah it looks BAD but not like SENSOR-BAD but like GLASS-BAD" which I agreed with.
I seemed to recall reading something rather unfavorable about that Sony Lens, and did a little looking to find:
An Update and Comparison of the Sony FE 70-200mm f/2.8 GM OSS
The LensRental tests used ONE body (A7R2) for the Sony, Nikon and Canon 70-200 pro-level 2.8 zooms, and they did not find the Sony to be overly great (a bummer, because that lens is **** expensive).
Tony's results seem to confirm that not only are the Canon 70-200 and the "DT 70-200 F2.8 SAM" lenses really great wide open, but that the Sony 70-200 lens is certainly NOT...and that's consistent with the Lens Rentals tests.
So Tony seems to have accidentally made a good case about glass... and also banding with the EOS-R.
Personally I'd love to see all these cameras mounted and adapted to a single piece of glass for these tests just to control all variables save for the sensor.
I realize adapters aren't perfect but in this case I think they'd be fine since there's no AF speed needed or quirky lighting going on.
Anyway. Enough geekery.