[Class A]

However the people asked about it seem to fall into one or the other camp and every review seems to have problems - that is informative.

Every review?

How about this ephotozine review?
It is not for the K-mount version but armchair experts have made the argument that the AF issues must be systemic because they exist on other mounts as well.

As for people falling into two camps: That means it is possible to get a copy that works fine. If you are not willing to do what it takes to get such a copy, that's fine with me, but please don't become one of those that can only see one side of the story.

[MadMathMind]

The 31 doesn't miss though and just renders better.

The first part is totally false. I get a *lot* of misses with my FA31. I shoot it south of f/2.8 almost exclusively and I get a ton of out of focus shots. Put it up to f/3.2 or so and there's no issue. That's what I see with my Sigma, for the most part. I also have issues with the FA77 as well, missing focus at wide apertures, so it's not just my copy. The K-5IIs really is hit and miss (quite literally) below f/2.8.

But the second part is true. It does a better job with people, especially. The Sigma is awesome for landscapes and things, though.

[UncleVanya]

Every review?

How about this ephotozine review?
It is not for the K-mount version but armchair experts have made the argument that the AF issues must be systemic because they exist on other mounts as well.

As for people falling into two camps: That means it is possible to get a copy that works fine. If you are not willing to do what it takes to get such a copy, that's fine with me, but please don't become one of those that can only see one side of the story.

Statistically I'd say nearly every review and I would also say that if the performance is so variable that 3 sample copies can fail here - that's not a good track record. I also will say I read the review linked above and I have little faith in that review; it provides very little in the way of discussion of focus and doesn't really seem that in depth compared to many others I have read.

I'm not in the camp of the pixies who ignore all evidence to the contrary but I'm also still holding out hope. When a publication or reviewer I respect comes forward with an exhaustive side by side comparison of this lens and another fast one like the FA 31 f/1.8 and shows that the problems are inherent not just problems of the lens itself then I will be convinced. But when person after person tries it and fails with it; some with initial success and hope, many with the AF dock - it just becomes an Emperor's new Clothes situation. It is hard to bother with trying to find a good sample in a sea of bad ones. If that's indeed what is going on. People have tried multiple lenses and bodies and had consistently bad results and others have had no problems at all. This does not sound like the way I want to spend my time - trying to second guess if the sample I have is any good.

Look - if it works for you - publish a review - make it thick with photos and do some A/B comparisons against other fast lenses and be brutally honest - set it up with multiple subjects and test targets and be methodical and prove that your copy works. Then OFFER it to PF for a follow up review and allow them access to your good sample. THIS is how to shut up people who say the lens is flawed.

[nicoprod]

The first part is totally false. I get a *lot* of misses with my FA31. I shoot it south of f/2.8 almost exclusively and I get a ton of out of focus shots. Put it up to f/3.2 or so and there's no issue. That's what I see with my Sigma, for the most part. I also have issues with the FA77 as well, missing focus at wide apertures, so it's not just my copy. The K-5IIs really is hit and miss (quite literally) below f/2.8.

But the second part is true. It does a better job with people, especially. The Sigma is awesome for landscapes and things, though.

My copy of the 31 almost never misses, and I use it 1.8 to 2.2 almost exclusively. Since it gets better passed f3.2, you may have a less than optimal copy, or is it you K5IIs?

---

Last edited by nicoprod; 08-04-2015 at 08:05 PM.

[Digitalis]

The first part is totally false. I get a *lot* of misses with my FA31. I shoot it south of f/2.8 almost exclusively and I get a ton of out of focus shots. Put it up to f/3.2 or so and there's no issue. That's what I see with my Sigma, for the most part. I also have issues with the FA77 as well, missing focus at wide apertures, so it's not just my copy. The K-5IIs really is hit and miss (quite literally) below f/2.8.

Strange, I use a K5IIs with plenty of Fast autofocus glass and rarely have issues with focus at apertures faster than f/2.8. Though I spend about an hour, on average, painstakingly calibrating each lens to the camera...thankfully I only need to do that once a year or so....or every time I get a new camera body.

P.S I probably should have mentioned this, the action I have been taking against the Australian distributor has been in a logjam, the distributor maintains that the lens was within tolerances when they tested it. I took the issue up with sigma Japan - I sent them copies of my optical test bench data, and their techs were horrified. Sigma confirmed that the lens passed all their tests at the production line, so somewhere along the line it could have been dropped upon delivery (one of the corners of the box was damaged, so that is a very real possibility). Hopefully Sigma will communicate to CRK over this issue and get them to back down. It has been a drawn out affair of corporate stubbornness on CRK's part.

[drypenn]

The first part is totally false. I get a *lot* of misses with my FA31. I shoot it south of f/2.8 almost exclusively and I get a ton of out of focus shots. Put it up to f/3.2 or so and there's no issue. That's what I see with my Sigma, for the most part. I also have issues with the FA77 as well, missing focus at wide apertures, so it's not just my copy. The K-5IIs really is hit and miss (quite literally) below f/2.8.

I have both the 77 and the 31 and a couple of digital bodies. The 31 and the 77 has an exemplary AF accuracy with the K3 and the K30, though the 31 and the KX is a hit and miss affair, (but more on the hit side). Curiously the KX is spot on with the 77.

[Class A]

I get a *lot* of misses with my FA31. I shoot it south of f/2.8 almost exclusively and I get a ton of out of focus shots. Put it up to f/3.2 or so and there's no issue. That's what I see with my Sigma, for the most part.

The best PDAF accuracy provided by an Pentax DSLR is based on an AF area with a baseline of f/2.8.
It therefore makes sense that results below f/2.8 will be "hit and miss" to some extent. I'm not sure whether "a ton" should be out of focus, but most certainly one must not expect perfect results for any aperture bigger than f/2.8 whether the lens is a Pentax or a Sigma.

The resolution accuracy is just not good enough to consistently achieve accurate focus at f/1.8 or even f/1.4.

Anyone who has tried to manually focus an f/1.4 lens wide open with the stock screen (that won't discriminate mis-focus below f/2.8) can testify to the impossibility to ensure that every shot will be spot on. With a lot of practice one can come close or even hit the mark by understanding where to place the focus within the blind range, but identifying the blind range and then judging where to place the focus within it based on subject distance takes time and there will always be outliers.

Note that the f/2.8 resolution is only achieved in one direction (horizontally), i.e., there are no f/2.8 precision cross-type AF areas (neither for the K-5 II nor for the K-3). That means if the horizontal AF areas fail to lock on to some detail, the AF precision falls back to the standard f/5.6. And any AF areas outside the centre have f/5.6 resolution to begin with. This fact alone can explain a lot of variability regarding AF results, depending on whether the shooter only uses the centre AF area and/or achieves a horizontal AF lock or not.

Many photographers do not seem to realise that achieving appropriate focus is the sole responsibility of the camera. The lens plays a purely passive role and the camera alone decides how to hone in on the correct focus and when to give up (or incorrectly declare that focus has been achieved). This is clearly demonstrated by the fact that when using CDAF, the Sigma 18-35/1.8 has no AF problems whatsoever for anyone. Only PDAF represents a problem for some. Now even though the lens has a passive role, it must play its part, of course. For instance, if spherical aberrations are not controlled, the lens may exhibit a focus shift behaviour. Or if the lens reports a wrong AF gearing to the camera, the camera will consistently overshoot and may give up achieving focus based on the fact that a regular amount of attempts were not enough.

It would be absolutely great if someone who has severe AF problems with their Sigma 18-35/1.8 could enter a dialogue with Sigma Japan to find out if perhaps two different versions of lens EEPROMS are used, one containing incorrect gearing information and the other with a correct one. As far as I know, this has never happened yet so we are left with a few reports that may be the result of unrealistic expectations, blaming the lens rather than the camera, or may be justified due to some problem with the lens copy (EEPROM-based or optically induced). Due to the existence of happy users (and the relatively low complaint rate), I'd say it is pretty clear that the Sigma 18-35/1.8 does not have a systemic problem.

[Class A]

Strange, I use a K5IIs with plenty of Fast autofocus glass and rarely have issues with focus at apertures faster than f/2.8.

Please see my previous post why even the best current Pentax AF systems cannot be expected to achieve perfect focus for apertures bigger than f/2.8. Your use of the term "rarely" seems to confirm that but given the variability with respect to AF area choice and subject matter (that may not excite the horizontal f/2.8 AF areas), it is not "strange" at all that some shooters like MadMathMind are underwhelmed by AF accuracy, Pentax lens or not.

Hopefully Sigma will communicate to CRK over this issue and get them to back down. It has been a drawn out affair of corporate stubbornness on CRK's part.

I'd say there is a very low chance of anything happening on this basis. By going to a different repairer, you let CRK off the hook and they have no unfinished business, as far as they are concerned. Not fighting a continued battle with CRK was your right and may have been the best choice given the circumstances. I expressed earlier how much I despise service centres that will declare a lens "within specifications" even though it has obvious problems. I don't see Sigma Japan acting on the basis of your test bench results either, as it would be reasonable to assume for them that your case is an isolated one. Sigma Japan has been quoted as stating that there are no known problems with the Sigma 18-35/1.8 so on what basis should they start an investigation?

[Class A]

But when person after person tries it and fails with it; some with initial success and hope, many with the AF dock - it just becomes an Emperor's new Clothes situation.

Except, that it is clearly not the case that "person after person tries it and fails with it". You are ignoring the positive examples, you are ignoring the rather low levels of complaints, and you are extrapolating a few bad reports to a systemic problem. DPReview has received three optically compromised copies of the DA* 55/1.4 in a row. A PF user has tried five DA* 55/1.4 copies and could not find one that did not behave erratically in terms of AF. Does this mean the DA* 55/1.4 is always de-centred and it would take an inordinate amount of time to get a copy that does not lead to erratic AF results? Your call.

If you do not want to spend the time to get a good copy of a lens, that's fine. However, it is another step to claim that a certain lens model requires more effort in this regard than another.

Look - if it works for you - publish a review - make it thick with photos and do some A/B comparisons against other fast lenses and be brutally honest - set it up with multiple subjects and test targets and be methodical and prove that your copy works.

Et voilà, we have reversal of the burden of proof.

Please note that the reviewer of the PF review of the Sigma 18-35/1.8 only contacted Sigma USA once and published the damning review without consulting a second time with Sigma USA. Sigma USA is a distributor only anyhow and at no point the reviewer contacted Sigma Japan. This is very unfortunate, I think.

While the number of non-performing copies the reviewer and Adam received is most unfortunate, we do not know whether there is a simple explanation that links the origin of those copies with their performance. All theories of the lens having a systemic problems are based on circumstantial evidence. No one has given a cogent account as to why special caution needs to taken when acquiring a copy of the Sigma 18-35/1.8. But sadly, now some people are requesting a rigorous proof as to why they need not be concerned.

N.B.: Why do I care at all?
In all likelihood I won't get a copy of the Sigma 18-35/1.8 myself as I have set my mind to an FF camera and am not in interested in APS-C only lenses.

However, I do care about what glass is available for Pentax cameras. The commercial insignificance of the K-mount already lead Tamron, Zeiss, and Cosina to cease production of K-mount versions of their glass. Even Sigma, with their strong support of the K-mount, does not offer all their lenses in K-mount. A most likely undeserved bad reputation for the Sigma 18-35/1.8 may cause so much damage to K-mount sales of the lens that Sigma may one day decide to not bother with K-mount anymore either.

While the new Pentax FF lenses appear to be great performers optically, they are overpriced, AFAIC. Losing Sigma as an option would mean a big blow in terms of availability of lenses that I regard to be optically excellent and very competitively priced.

Finally, I have an adversity to injustice.

[Digitalis]

I'd say there is a very low chance of anything happening on this basis. By going to a different repairer, you let CRK off the hook and they have no unfinished business

Considering how long this has taken to get this far i'm inclined to agree.However a violation of Australian consumer law is no quibbling matter, goods not meeting manufacturer standards and distributors selling products as if they were, is a problem.

I have had very positive dealings with Sigma Japan before, and not just over K mount lenses. Some F and EF mount lenses I have bought new* had experienced issues and Sigma Japan has been prompt with dealing and accommodating my requests for calibration or repairs, even though the official Australian distributor CRK was not. This isn't the first time CRK has tried to sweep a lens issue under the rug.

*Even some older second hand lenses that needed to be re-chipped for full functionality, Sigma took care of that.

[howieb101]

This thread seems to have de-generated a bit dependent on select peoples point of view.

From what I can see, there is enough evidence to suggest that AF could be an issue with this particular lens. I would suggest only buying if you know the retailer has a good return policy. Why risk several hundreds of $$$ on something with known AF issues (even if this is a select subset of users). There is nothing worse than having read thread after thread, ignoring feedback and then not being satisfied.

This thread wouldn't exist if the AF wasn't an issue at least for some users.

I personally bought the K3 having read all the comments about mirror flop. I had to make my own assessment of this risk.

Any buyer of the Sigma 18-35 will have to make their own assessment of risk as people do when purchasing the SDM DA* lenses.

[Digitalis]

Any buyer of the Sigma 18-35 will have to make their own assessment of risk as people do when purchasing the SDM DA* lenses.

That is the bottom line.

[rawr]

My 2c (again) ... all I know is that my Sigma 35 f1.4 Art on a K-3 has similar issues to those reported with the 18-35. It's also the only AF lens I have (out of a few dozen) that displays such dumbness. Furthermore my issues with the 35 f1.4 even match those of folks using Nikon ...

So IMHO something does seem afoot with some of these new fast Sigma's.

As has been noted elsewhere, PDAF issues like those reported here are a very good advertisement for mirrorless.

[Digitalis]

As has been noted elsewhere, PDAF issues like those reported here are a very good advertisement for mirrorless.

I know this is de-railing the origiinal topic but: Using fast glass Designed for SLRs like the sigma 35mm f/1.4 adds bulk and kind of, ruins the ergonomics of small ILC cameras.

[rawr]

small ILC cameras

The horse has bolted. FF ILC's have left 'small' behind, now it's merely 'smaller'. Look at bodies like the A7 II, A7RII etc, and those big new FE Zeiss and other lenses. Hardly small

'Small' ILC only really applies nowadays to APS-C and micro 4/3 ILC, but even there some of the lenses (zooms especially) are huge and un-gainly.