Not surprised by your test of the Sony 28-70mm, but...
...it would have been more fair and interesting to test it on the A7, not the A7r. Sony doesn't appear to recommend the lens with the higher-resolution A7r.
Also, it's not accurate to refer to the 24-70mm as a "Zeiss." I am sure the people at Zeiss cringe every time a reviewer does that. This lens is a Zino - "Zeiss In Name Only" - it carries a badge licensed by Sony from the Zeiss licensing department. I own quite a few Carl Zeiss lenses for my Hasselblads, and the 24-70 is by performance as much a Zeiss lens as the 1989 Chrysler LeBaron "Maserati" was a Maserati. When I saw the range of quality coming out of the first batch, it was generating reports nothing like a real Zeiss-made/Zeiss-QC'd lens. It would be interesting to know if the lens you tested came from that first, troubled batch.
You tests pretty much confirm my actual shooting experience with the lens, although since I use Lightroom 5.4 for post-processing and apply the profile, I don't see the CA, distortion or vignetting a "real" lens test might reveal. The lens is absolutely brilliant at 28mm. If the 24-70's sweet spot is 35mm, that also confirms the experience of those on the Sony NEX forum - the lens is weak at 24mm, which is what one is supposedly paying the extra $900 for. Oh, and the Zeiss logo, too.
I found your transmission data at 70mm interesting. Earlier this week, I shot the Boston Marathon with both the 28-70 and an old Nikon 70-210 f4-f5.6 AF lens. The Sony at 70mm was a full stop faster at the same aperture setting as the old Nikkor at 70mm. Even accounting for optical glass and computer formulation changes in the 20 years between the design of the two lenses, that surprised me.
| Read all replies for this comment