Update 2019-09-14: the 2nd sample is notably better across the field. The first sample and many samples out there might well vary widely. High performance lens designs require tolerances that can be difficult to achieve in manufacture—there are many fine lens designs which are never built because it would be too hard (expensive) to reliably built to the designed performance level. Sigma’s quality control needs to be improved, and it raises the issue of whether an advanced design will retain its performance over time after minor bumps and such.
I take a first look at the Sigma 35mm f/1.2 DG DN Art. The results are ridiculously bad, so one might assume that I received a bad sample. Except that a trusted reader reports the same behavior with his sample on his Sony A7R III.
I capture my thoughts on the performance of the sample I’ve got in my overview page, after shooting a variety of outdoor images and then some indoors under controlled conditions:
Hopefully Sigma just shipped a bad batch of lenses. It is highly unlikely that it is my sample alone, based on what this trusted reader tells me about his sample—too consistent.
The MTF page looks at the fantasy MTF chart versus what the lens actually delivers in reality:
I have also had two reader reports of focusing accuracy problems with the Sigma 35mm f/1.2 DG DN Art. One of those reports is on the Sony A7R III (focus inaccuracy) and the other on the Panasonic S1R with a failure to reach infinity focus with AF.
I have confirmed focusing accuracy issues with my sample also, no surprise given the strongly veiled wide-open performance. So that’s three different lenses on two camera platforms and three different cameras, all with focusing problems.
Hold off on the Sigma 35mm f/1.2 DG DN Art! I have requested a replacement sample as I see no point in proceeding at this time, and of course the Sony A7R IV will be a far more demanding camera than the Sony A7R III (60MP vs 42MP).
I’m a reader from China and have been subscribing to your website for the past 3 years. I really appreciate your strict assessments on camera bodies and lenses and benefit a lot from your articles.
I bought Sigma’s new 35/1.2 Art and 45/2.8 C about a month ago, both with L-mount as I have a Panasonic S1R. The claimed MTF of 35/1.2 Art suggests outstanding performance. I had very high expectations indeed. With a few quick test shots, it soon turns out that the Sigma 35mm f/1.2 DG DN Art is a total disappointment. At f/1.2, obvious aberrations almost poisoned the entire half outer part of the image, not to say the sharpness. The performance doesn’t reach acceptable level until f/2.8, which is exactly what you have inspected. A lens of this kind should be designed to use at f/1.2 with an acceptable performance. Why don’t buy a 35/2 if the lens is not usable until f/2.8?
As for the Sigma 45mm f/2.8 DG DN, I bought it as a fun lens, so I don’t have high expectations on its optical performance.
BTW my Fujifilm GFX100 shows unstable exposure measurements. I haven’t looking into this issue, but it feels like that the aperture gets stuck and fails to open while measuring.
DIGLLOYD: my findings on the Sigma 35mm f/1.2 DG DN Art are now “thirded” by Ding—another reader found the same thing, so did I, and so did Ding. Nevertheless, I have another sample on the way to compare the two to see if there is much if any variation. My view is that the disappointing performance from f/1.2 through f/2 is by design—not at all to my liking and no match for Canon’s gorgeous Canon RF 50mm f/1.2L and Canon RF 85mm f/1.2L (presumably with 24mm and 35mm offerings to come).
The Sigma 45mm f/2.8 DG DN Contemporary is a sleeper lens that might have potential when used appropriately—I’ll be looking into two sides of that question: rendering style at f/2.8 and f/4, and technical performance stopped down. It is very compact and nice on a Sony body, so it might actually be very good for landscape, focused properly accounting for focus shift.
Fujifilm GFX100 when I tested it was very consistent for metering, so that point may be about a particular lens. Or perhaps something else is going on that I did not observe, based on different settings or some such.