Poll, R mount and 3ed party lenses

KevinRA wrote:

PAntunes wrote:

KevinRA wrote:

+1 to the OP on lesser reliability on DSLRs on normal phase detect AF (e.g. no live view) - - I had no luck with 5DSr and IV with Sigma wide angle fast primes - and pre then even with cross point AF points with other fast lenses - less consistent even after calibration. Never eve had an issue with a Canon lens.

Plenty of reviewers cite issue - especially Bryan at the digital picture - so whilst not proven - there is plenty of reliable reports out there backing the OPs point.

The review of the sigma art 85 1.4 on the-digital-picture:
After capturing over 600 tripod-based images of various subjects strictly for the purpose of testing AF, each capture starting in an out-of-focus condition, I found that this lens focuses consistently accurately.

https://www.the-digital-picture.com/Reviews/Sigma-50mm-f-1.4-DG-HSM-Art-Lens.aspx

"The longer I focus tested this lens, the less sure I was about its focus accuracy. I formally tested close, far and distances in between using Canon's best AF system in the EOS 5D Mark III and 1D X along with the more affordable AF system in the EOS 60D. Focus targets ranged from the Datacolor SpyderLensCal to pineapples and from flowers to people.

What I learned from the many hours (actually extended into days) spent shooting and analyzing thousands of images is that the 50 Art lens' AF cannot be completely counted on. ...sometimes, more images are out of focus than I am comfortable with.

A second copy of this lens shows similar inconsistent performance. Both this camera and the lens (via the dock) can be focus calibrated, but you cannot calibrate for inconsistency."

https://www.the-digital-picture.com/Reviews/Sigma-24mm-f-1.4-DG-HSM-Art-Lens.aspx

I shot hundreds of tripod-based f/1.4 focus accuracy test photos using a 5D Mark III, each from a slightly de-focused lens, using various AF points, both indoors and out, and saw near-ideal results from all tests. But sometimes, typically in a specific test scenario, the results were all over the place, showing AF inconsistency. I have not determined what leads to the inconsistent AF results, but they happen enough that you need to be aware of this issue.

Sitting on the ground with my elbows resting on my knees (a very stable 3-point position), I focused a mid-upper-left AF point on the sign and recomposed slightly. Over half of the results were unusably miss-focused.

https://www.the-digital-picture.com/Reviews/Canon-EF-24mm-f-1.4-L-II-USM-Lens-Review.aspx

Auto focus, driven by Ring USM (Ultrasonic Motor), is fast, quiet and accurate (mostly) - even in low light conditions. I say "mostly" because I am getting an occasional front-focused shot that was not my fault (tested on a 1Ds Mark III body).

https://www.the-digital-picture.com/Reviews/Canon-EF-16-35mm-f-2.8L-III-USM-Lens.aspx

AF accuracy cannot be tested in every scenario possible, but ... too often this testing takes on a very time consuming life of its own. But, this is nearly always with a third party lens. On the EOS-1D X Mark II, the 16-35 f/2.8L III's focus accuracy performance has been nearly flawless. Both copies of this lens focus very accurately with the 5Ds R in the lab. But, for some reason, both lenses gave me a modestly reduced AF hit rate on two other 5Ds R bodies. Although focusing was inconsistent, AF Microadjustment (to the front) made one of the two lenses focus very well with only infrequent misses. Strangely, I have not yet been able to completely dial in the other camera yet. Results from the 7D II were not as consistent as from the 1D X II.

He probably hasn't found more problems with Canon lenses because he does af tests mostly to third party lenses. The sigma lenses are just as consistent as OEM lenses. And we're talking L series OEM lenses...