Reader Comment: Zoom Lens Sample Variation especially the Sony 24-70mm f/2.8 GM II
Related: coin-operated product review, Panasonic, reader comment, Sony, Sony FE, Sony FE 24-70mm f/2.8 GM II, Sony Lenses
Alex C writes:
I was most interested in your findings on the new Sony 24-70mm f/2.8 GM II zoom. I have the Panasonic 24-70m f/2.8 S Pro which is just as bad if not worse on the right side! I wonder if they come from the same factory? Surely neither Sony nor Panasonic make their own glassware and there can't be that many companies making such things?
It's a shame because at the wide end in particular the zoom does remarkable things in some parts of the frame; it did well in comparison with the Sigma Art 28/1.4 on the S1R, although the Sigma is hobbled by egregious focus shift (which I cannot do anything about because, unless I'm being a dunce and missing something, you can't focus stopped down even in full manual on the Panasonic S series!)
[DIGLLOYD: what I remember is that the Panasonic S1R had no focus-stopped-down feature and this led to focus shift issues with Leica SL lenses]At 40-70mm, the right side of the frame is pretty much unacceptable even at f/8, although stopping down does help. That's an expensive lens and it's basically unusable for anything other than the most casual snapping.
There's obviously a trend in the camera industry of designing lenses they can't actually make properly!
Although I just wonder if the same factory is to blame here... if I remember rightly you tested a very badly skewed 50/1.4 S Pro, also likely from the same place.
[DIGLLOYD: see Pines by Creek]
DIGLLOYD: I would love to test a good sample of the Sony FE 24-70mm f/2.8 GM II, because it’s obvious that the lens design is superb. It would be a 'killer' lens to have if you can find that good sample.
I don’t know anything about the factories, but it could be that it is contracted out to the same 3rd-party facility. But I don’t know for sure.
Quality control is an industry-wide issue. I am loathe to say Sony is better or worse, but I am a little shocked at this point about the 24-70mm GM II. A rollout PR disaster if you ask me, because there is no way any pro should have to deal with the godawful results I am seeing, let alone in an about $2300 lens. But coin-operated reviewers will deliver just what the marketing department needs, so I’m probably just pissing into the wind.
Based on the past 13 years and what I see across all brands, I think quality has on average fallen behind. Complexity of lenses has increased, with no commensurate improvement in build tolerances. And with some lenses, it seems highly probable that the lens design is far too sensitive to minor deviations.
The results with both brand-new samples of the Sony FE 24-70mm f/2.8 GM II both at 50mm and at 65mm show the most severe problems I can remember— a textbook example of “bad sample”. When f/8 is no better and maybe worse than f/2.8, that’s shockingly bad, dead on arrival. The wide end has problems too, but the wider focal lengths overcome the issues with stopping down by f/5.6-f/8, which is not very satisfying for an f/2.8 pro lens.
I cannot say how many Sony 24-70mm f/2.8 GM II lenses have problems, but I will hazard this guess: if 1 in 10 samples performs reasonably well across the zoom range without obvious issues, you should consider yourself lucky. I would not accept anything better than 1 in 5 good samples as credible, because an associate with two other brand-new samples found that one of them was grossly poor. The other was not obviously bad, so possibly it is that 1 in 5, but I did not personally test those two.
It looks like Sony is designing lenses that cannot be reliably built to deliver anything close to what the computed (not measured) fantasy MTF charts claims. Such charts are pure marketing hokum.
That’s self evident with the Sony FE 50mm f/1.2 GM (with its jaw dropping MTF chart), but at least it shows no obvious badness. But this new and highly promising Sony 24-70mm f/2.8 GM II is both superb (in places), and awful. Sadly, the awful part overwhelms anything else. I expect at least superb to acceptable range, and it does not rise to that.
The only credible vendor (on MTF) out there is Zeiss, making real MTF charts from real lenses using the Zeiss K8 MTF tester. I was there and saw how it is done. Indeed certain MTF chart I’ve published are from my own personal samples of Zeiss Milvus/Zeiss Otus lenses.
It is my understanding that the reason Zeiss Otus lenses cost so much (about twice the Milvus line) was a very high rejection rate in order to guarantee the performance that was claimed in the (measured from real lens) MTF chart. It is improbable that Sony or Fujifilm or etc would ever consider such an approach, and I think the proof of that is self evident in the pricing.
