[osv]

the imaging resource guys bashed the continuous autofocus pretty hard... it looks like the knob selection for it was removed entirely on the rx1r version... the lens has 9 blades, but there are bokeh complaints, in certain situations.

what i question about the rx1 is how well the lens works for manual focusing... the evf is $450? ouch

with these sony evf's, the more pixels they pack into the sensor, the higher the magnification... as in, a7r evf zooms further in than the a7, which zooms in more than the a7s.

if i had an fa31 for my a7r, i just couldn't see myself trading the whole thing in for an rx1, good as it is.

thx for the feedback on the rx1.

[Digitalis]

the Sony RX1's lens is the sharpest, most uniform full-frame 35mm optic we've ever seen. Period, full stop

Clearly they have never worked with the Leica Summilux-M 35mm f/1.4 ASPH.

Though they are comparing a fixed lens sensor camera to interchangeable ones - and using adapters will also have an effect in image quality. So for a purely technical perspective their testing is flawed. If you wanted to use a lens like the FA*200mm f/4 ED MACRO the sony RX1 wouldn't be your best choice. This is the thing with fixed lens cameras - manufacturers are able to cover up the optical faults of a lens better than cameras that belong to interchangeable lens systems. Even if you are shooting raw on the RX1 I wouldn't be surprised if Sony was altering things behind the scenes.

Good job on the review, all in easy to understand terms and practical comparisons. Thanks

Good to hear that you found it to be useful and easy to understand....as difficult as it is for me I try to keep jargon to a minimum.

[simon6z]

Well people keep saying that if pentax come out a zoom faster than f2.8 it will be damn biggie but see what sigma roll out will mean everything is possible though

[Digitalis]

Pentax forums user Heie Has published his in-depth review of the Sigma 18-35mm f/1.8 ART - he does compare the Sigma 18-35mm f/1.8 to other pentax lenses, including the 31mm f/1.8 ASPH Limited, his results when comparing the sigma lens to the Pentax FA31 are different from my own. In this case the culprit is most likely sample variation. Heie also encountered problems with AF - which I have also experienced. As things stand I'd say the AF on my Sigma 18-35mm f/1.8 ART with PDAF is accurate only 78% of the time*.

*even after calibration I find there is still a need to manually focus the lens slightly so it remains on target.

[Class A]

Pentax forums user Heie Has published his in-depth review of the Sigma 18-35mm f/1.8 ART

Thanks for the pointer.

Stellar job on the review on many accounts, but three issues stick out, one of them being huge.

1. The copy of the FA 31/1.8 he used has issues. Unfortunately, many copies have subpar optical performance due to internal barrel loseness, but a good copy should have performed a lot better.
2. The DA* 20-40 is praised as being sharp, yet in all the shots I looked at, it was clearly lagging behind. I used a mobile phone so maybe I missed something, but there is also an inconsistency in the sharpness findings of the DA* 20-40.
3. The mystery of the AF issues is not explained despite the praise about Sigma's excellent communication. Unless the AF errors were erratic, i.e., not consistent from shot to shot, the use of the USB dock should have addressed the issues. If the AF errors are erratic indeed, what is Sigma's explanation? What is the role of the camera in this?

I remain highly sceptical that the damage this review will do to sales of the lens is actually justifiable.

The lens is described as being unusable with PDAF which I cannot believe to be an accurate assessment in general. Clearly, even Digitalis' close to 80% success rate - while not great - is better than that and LensTip found no issues with the AF at all.

[Digitalis]

he mystery of the AF issues is not explained despite the praise about Sigma's excellent communication. Unless the AF errors were erratic, i.e., not consistent from shot to shot, the use of the USB dock should have addressed the issues. If the AF errors are erratic indeed, what is Sigma's explanation? What is the role of the camera in this?

Many of the AF errors I have experienced are erratic, Heie seemed to have an even worse time with AF than I did. The lens dock can take care of the predictable AF errors, but there are instances of the sigma lens mis-focusing that to me, simply defy explanation*. This is the part that I found to be rather questionable myself, the bias towards the 20-40mm and the negativity towards the FA31, and Heie praising the sigma lens despite a crippling deficiency in terms of autofocus reliability brings the entire review into question. The sigma 18-35mm f/1.8 lens is optically excellent - but there are serious issues that prevent me from recommending it to a casual user.

The DA* 20-40 is praised as being sharp, yet in all the shots I looked at, it was clearly lagging behind. I used a mobile phone so maybe I missed something, but there is also an inconsistency in the sharpness findings of the DA* 20-40.

Also the fact that the resolution test images were over exposed 1/8th of a stop didn't help much either.

The copy of the FA 31/1.8 he used has issues. Unfortunately, many copies have subpar optical performance due to internal barrel loseness, but a good copy should have performed a lot better.

It would have been interesting if I had sent Heie my copy of the FA31. I think his negative opinion of the FA31 would have a sudden drastic reversal.

LensTip found no issues with the AF at all.

I find that suspicious. I doubt they really used the lens in the field all that much, I'm talking spending weeks with just this lens on their camera and nothing else.


* I know how PDAF systems work, and there are situations where you have a perfectly contrasty subject and the lens for some perverse reason chooses to focus behind or in front of it. This isn't just a Pentax DSLR problem I had a student a year ago who had this lens and was using it on a canon 7D and the lens was less than accurate under studio lighting. I have also had students with Nikon DSLRs that have had similar problems with Autofocus with the Sigma 18-35mm f/1.8 ART.

[Class A]

...Heie praising the sigma lens despite a crippling deficiency in terms of autofocus reliability brings the entire review into question.

I think that is a drastic view, but I have to admit that I wondered myself as to why -- if the AF issues are as serious as they are presented -- the review wasn't shortened to a brief "too expensive for a manual focus and/or CDAF lens".

For many, adequate PDAF performance is of paramount importance, as testified by the number of commenters that already said they won't consider this lens anymore. I wondered why so much effort was spent on describing other aspects of the lens if there is such a showstopper.

The lens dock can take care of the predictable AF errors, but there are instances of the sigma lens mis-focusing that to me, simply defy explanation*.

You are hitting a major issue:

There is simply no good explanation as to why the lens should be causing erratic PDAF.

We agree that lenses can easily cause consistent PDAF errors and that the latter can be addressed by AF-adjustments (in particular with the Sigma dock that allows much more precise correction than the typically in-camera adjustment).

We also seem to agree, however, that there is not good explanation available for explaining why the AF sometimes is spot on, BF, or FF, given a consistent focus target, and a fixed subject-distance and focal-length pair.

I suspect that even if the subject-distance and focal length is kept constant, the focus target is not. In other words, my best stab at explaining the observed issues is to assume that varying real world scenes are able to make the camera's AF system unexpected choices. Note that I'm talking about the camera's AF system.

We know that AF areas are large and that it is not entirely predictable how the phase alignment performed by the camera's AF system will end up preferring one part of the scene as a focus target over another. There are certain basic rules as in nearer and high-contrast details being preferred, but in the end the mechanism is not entirely predictable.

We also know that the problem becomes harder with wide-angle lenses and that large apertures (such as f/1.8) further aggravate the issue. Obtaining consistent focus with real world scenes at 18mm and f/1.8 is harder than it is at 31mm and f/1.8 or at 18mm and f/4.

Unless an explanation can be found as to how a lens should be able to erratically mess up a camera's AF system -- it is, after all, the camera that makes the decision about where to put the focus -- or someone can empirically demonstrate with statistical significance that the Sigma 18-35/1.8 performs worse regarding AF than other lenses with identical parameters, I am assuming that it is wrong the blame the lens. I can certainly understand why users can come to the conclusion that the lens has issues, but I remain unconvinced that such a conclusion can be justified.

One potential explanation for erratic AF behaviour would be shaky electronics that fail to consistently deliver the correct AF adjustment values to the camera due to unstable power supply, or similar. However, such a cause seems highly unlikely.

[MadMathMind]

I started a new thread to track focus issues: https://www.pentaxforums.com/forums/10-pentax-slr-lens-discussion/277499-sigm...es-thread.html

[Digitalis]

One potential explanation for erratic AF behaviour would be shaky electronics that fail to consistently deliver the correct AF adjustment values to the camera due to unstable power supply, or similar. However, such a cause seems highly unlikely.

Very unlikely but we shouldn't rule that out. Sigmas HSM is a stable, well tested technology. It hasn't had the troubled history that Pentax SDM has.

or someone can empirically demonstrate with statistical significance that the Sigma 18-35/1.8 performs worse regarding AF than other lenses with identical parameters

The problem is that no lens from any other manufacturer has identical parameters.