[Digitalis]

Despite the fact that some people in this thread can now only report on the negative aspects of the lens

Do you think the criticism of the AF accuracy of this particular lens is unfair? Do you think it is too much to ask for a brand new lens to be able to be calibrated to focus properly with at least 95% accuracy?

this lens can deliver superb results and that it may be acceptable to live with some uncertainty about focus accuracy

Even in broad daylight this lens will inexplicably mis-focus, typically within the 2m to infinity range.

I would suggest that Digitalis initially had high hopes

I did have high hopes, I really did. I was willing to drop all my baggage over sigma products, and see the ART series as something new.

[JanG]

I feel that at the end of the day, it can only be your call.

Only you know whether you can live with the percentage of bad shots this copy of the lens is delivering.

Despite the fact that some people in this thread can now only report on the negative aspects of the lens, despite having substantially praised the lens for its optical performance potential in the past, there is no denying that outside adverse conditions, this lens can deliver superb results and that it may be acceptable to live with some uncertainty about focus accuracy. Note that the Sigma lens dock may help to address focus issues to some extent, but that an erratic component appears to remain even after calibration, at least for some copies (I am in no position to ascertain that all copies are affected by the erratic focus behavior, nor is anyone else here).

If I were you, I would try to come to an arrangement with the dealer to try another copy of the lens with the fallback option that you could get back your original copy in case it seems to be a perfect copy with the exception of the focusing issue. If you are lucky, then perhaps you get a copy whose focus failures are entirely within the normal range of Pentax PDAF performance (it is not perfect and for technical reasons absolutely cannot be perfect for lenses with an aperture wider than f/2.8 (<- the widest base of the PDAF system, with most focus areas only achieving f/5.6 precision).

For every person who claims my "so and so genuine Pentax lens never misfocusses, there is another person reporting the occasional misfocus result from wide aperture genuine Pentax lens). So don't expect 100% perfection. An 80% success rate is not what one would normally expect, though.

Would be good to be hearing back from you, should you decide to try another copy.

Thanks for your most valuable input, Class A. I have come to more or less the same conclusions and will contact my dealer about that kind of arrangement.

Concerning misfocussing I am still not certain that it is an issue of the lens. I sometimes experience unexplainable misfocussings even with Pentax lenses on my K-3 - though not at a rate of 20%, even 2% might be an exaggeration. Also the rate of misfocussings seemed to be substantially lower on my K20D. It really is a strange phenomenon. E.g. I had bad results when I focussed at wind energy plants on the horizon but got just the closest part of the open field in front of me in focus

All in all I am leaning to keeping the lens as it is fantastically sharp in at least 80% of the cases and I do enjoy its rendering. 80% super sharp photos means: without using the dock. The lens works completely as advertised in 80% of the cases. and that does not sound too bad to me. For critical shots I will always try to take several shots of the subject with refocussing to something at a substantially different distance in-between. This seems to be the trick to avoid series of misfocussed photos, as camera and/or lens seem to remember the achieved focus lock (wrong or right) as long as the lens does not have to readjust substantially.

Well, the next step will be calling the dealer tomorrow.
I will keep you updated.

[JanG]

Just a brief update concerning my PDAF issues:
I just spoke to a service staff of my retailer and they admitted that "some" of the lenses are returned to them because of (unspecified) focussing issues. They did not offer to exchange the lens for another copy but suggested to send in my cam together with the 18-35 and let the combo be checked. Normal return time would be about 2 weeks (wow!!!). This offer would be valid throughout the warranty period of 3 years. The service staff also reported his private experience with a Sigma 18-50 2.8 (don't know the mount) which after purchase had been very sharp except for a band of focal lengths. He had send in the lens with his camera and had got back a perfect combo. So maybe Sigma indeed could do something as was reported here by someone from Australia, IIRC.
For now I will continue my testing for some more time and then decide whether to send my combo in or not.

[Class A]

I just spoke to a service staff of my retailer and they admitted that "some" of the lenses are returned to them because of (unspecified) focussing issues.

This could mean that some copies are working perfectly or that the respective users are either not capable to recognise AF issues or are not bothered enough to complain.

They did not offer to exchange the lens for another copy but suggested to send in my cam together with the 18-35 and let the combo be checked.

I would take them up on the offer to have the lens checked with the camera but I'd try to somehow make it impossible for the service technician to work on the assumption that your case just constitutes the regular need for AF micro-adjustments.

I'd make it clear, when handing in the combo, and through notes attached to the gear, that your problem is not some consistent front-focus or back-focus for some combinations of focal length and distance, but rather a problem of erratic inconsistency.

I'd say only when the service staff is willing to accept the combo based on that premise, you'll stand a chance of seeing an actual improvement. Otherwise, your combo may be just subjected to a regular calibration procedure in which erratic focusing may not manifest itself as a problem at all.

The service staff also reported his private experience with a Sigma 18-50 2.8 (don't know the mount) which after purchase had been very sharp except for a band of focal lengths.

This sounds like his lens needed AF micro-adjustments for certain focal lengths. This can occur with any kind of lens with even primes (fixed focal length lenses) showing a dependency on subject distances and/or chosen aperture.

Some (modern) Sigma lens models allow one to address such consistent front-focus / back-focus issues oneself through the Sigma USB dock. The latter gives one user-friendly access to the AF micro-adjustment data within a lens.

Every modern AF lens (including those from Pentax) contains such corrective AF micro-adjustment values. With Pentax lenses, you can change the data as well but only by manually reading the lens data (through a service menu in the camera), manually manipulating it, and writing it back to the lens (again through a service menu in the camera).

[Class A]

Do you think the criticism of the AF accuracy of this particular lens is unfair?

Systematic focusing errors (i.e., consistent front/back focus) is a PDAF phenomenon that occurs with all lenses and which can be fixed by servicing, or in the case of some Sigma lenses, by oneself.

So we are only talking about erratic focusing which appears to be more frequent with the 18-35/1.8 than with any other lens. I have a hunch that the uniqueness of an f/1.8 zoom has some bearing on the frequency with which erratic focusing can be experienced.

Bear in mind that the PDAF focusing in your camera may only give you accuracy equivalent to f/5.6 unless the f/2.8 areas can engage.

So overall, I sympathise with everyone who has erratic focus issues with the 18-35/1.8 but I don't see how based on singular experiences, some internet reports, and ignoring happy users, it can be concluded that

1. all copies of the lens must have the same issue, and
2. only the lens is to blame, but the camera, which is the ultimate arbiter of when focus has been achieved, is not.

Do you think it is too much to ask for a brand new lens to be able to be calibrated to focus properly with at least 95% accuracy?

It would be good if all brand new lenses focused with that degree of accuracy, but the reality is that regardless of brand, this is not the case.

Earlier in this thread, I pointed to a user who tried five (5) different copies of the DA* 55/1.4. They all exhibited erratic focusing with his camera. Shall we now conclude that all copies have this trait and refer to Pentax as a brand with no QC? Your call.

There are many more examples for shoddy Pentax QC. It happens. We either live with it or somehow make manufacturers understand that good QC can reduce cost rather than increase it.

Even in broad daylight this lens will inexplicably mis-focus, typically within the 2m to infinity range.

Yes, but that does not diminish its optical excellence when focus is achieved.

You can continue to play that one-note song "the AF is not reliable" all day long but you yourself acknowledged to some length how exceptionally well the lens performs in other aspects.

I did have high hopes, I really did. I was willing to drop all my baggage over sigma products, and see the ART series as something new.

The ART series is something new and the optical excellence of its models is unequivocally recognised. It is your prerogative to turn a blind eye to the high praise these Sigma models receive from reviewers and users, but that does not make your opinion the truth.

The fact that there seems to be an unfortunate number of 18-35/1.8 copies that exhibit erratic focusing does not mean that your mom's idea of Sigma as being a shoddy company is still valid.

I said this before: The combination of you owning many Sigma lenses (including the 18-35/1.8) despite the fact that financially you could own any competitor lens instead (as demonstrated by your ownership of very expensive lenses) with your relentless Sigma bashing creates a bit of a logical tension.

[JanG]

This could mean that some copies are working perfectly or that the respective users are either not capable to recognise AF issues or are not bothered enough to complain.


I would take them up on the offer to have the lens checked with the camera but I'd try to somehow make it impossible for the service technician to work on the assumption that your case just constitutes the regular need for AF micro-adjustments.

I'd make it clear, when handing in the combo, and through notes attached to the gear, that your problem is not some consistent front-focus or back-focus for some combinations of focal length and distance, but rather a problem of erratic inconsistency.

I'd say only when the service staff is willing to accept the combo based on that premise, you'll stand a chance of seeing an actual improvement. Otherwise, your combo may be just subjected to a regular calibration procedure in which erratic focusing may not manifest itself as a problem at all.


This sounds like his lens needed AF micro-adjustments for certain focal lengths. This can occur with any kind of lens with even primes (fixed focal length lenses) showing a dependency on subject distances and/or chosen aperture.

Some (modern) Sigma lens models allow one to address such consistent front-focus / back-focus issues oneself through the Sigma USB dock. The latter gives one user-friendly access to the AF micro-adjustment data within a lens.

Every modern AF lens (including those from Pentax) contains such corrective AF micro-adjustment values. With Pentax lenses, you can change the data as well but only by manually reading the lens data (through a service menu in the camera), manually manipulating it, and writing it back to the lens (again through a service menu in the camera).

Thanks again for your very valuable input.
Actually, I have decided to test the the lens under real life conditions asap - something like a gallery opening, a party etc., one of the settings I actually bought the lens for. As it would not make any sense to send in the lens alone I will definitely send it in together with my camera should I feel the necessity after the test run.

What came to my mind yesterday was a strange and maybe stupid idea that was partly based on the impression of less erratic focus problems with my K20D. Could it be that the highly sensitive AF points of the K-3 get misled by reflections in fast lenses, especially if it is as complex as the 18-35 (lots of surfaces)? The AF points go down to -3 EV and might be sensitiv enough to take even very weak reflections as a subject to focus on.
I have observed that if you focus on a subject and the camera misfocusses it will do so again and again - even when actuating focussing for each single photo anew. The only thing to my observation that might get you a properly focussed photo is to let the cam focus on a subject at a substantially different distance and then to refocus on your intended subject. Could it be that the defocussing would resolve the reflections and thus enable the camera lens combo to loose its false focus and to turn to the correct focus distance? - Just speculation.

[MadMathMind]

Systematic focusing errors (i.e., consistent front/back focus) is a PDAF phenomenon that occurs with all lenses and which can be fixed by servicing, or in the case of some Sigma lenses, by oneself.

So we are only talking about erratic focusing which appears to be more frequent with the 18-35/1.8 than with any other lens. I have a hunch that the uniqueness of an f/1.8 zoom has some bearing on the frequency with which erratic focusing can be experienced.

Bear in mind that the PDAF focusing in your camera may only give you accuracy equivalent to f/5.6 unless the f/2.8 areas can engage.

I have shots that miss focus at f/6.3. That should *NEVER* happen. Ever. DoF should cover up the mistakes so it has to miss really badly for this to happen.

I used to love this lens. Now I don't even carry it with me most of the time. It fails me too often. I'll only bring it if I think I can't get the shot any other way (i.e., I need below 31mm FoV). In that way, I can tolerate failure only because my other option is to get nothing.

[UncleVanya]

Class A - you realize that Nikon and Canon also have seen similar issues with their version of the lens, doesn't that fact, 3 manufacturers, 3 focusing systens, inconsistent results across all give you sufficient reason to doubt this lens?

I have no dog in this fight.

[nmcleod1993]

I'm considering getting this lens, has the focus problems been resolved at all? Also if it hasn't is it a lens that I can keep on getting returning until one copy works well?

[Zafar Iqbal]

I'm considering getting this lens, has the focus problems been resolved at all? Also if it hasn't is it a lens that I can keep on getting returning until one copy works well?

Mine focuses better after I calibrated it using the the hub, but it's never good enough for distances longer than few meters. However, I've come to live with it and I often manually focus (except, I forget to do it too often too).

[AntonioS]

I have the same problem with the Sigma 18-35, specially with the K3-II. It focuses when using the live view but misses 50% of the time when using the optical viewfinder.
I tried like crazy to adjust the focus using the Sigma USB dock with no success. This lens has a very broad dispersion (not consistent) focus with the K3-II.

The funny thing is that it works much better with the k5-II.

My bet is that it's caused by a hardware and software combination. Too bad Sigma went mute about this, because the optic quality is astonishing when properly focused.

[AntonioS]

I tried the Sigma dock with FocusTune without much success, but I could tune this same lens in SA mount very well for the Sigma SD1 Merrill.

For now, I'm using it on my K3-II in manual focus (the camera has a Katz focusing screen), but this is completely ridiculous.

[superdave]

Yes this is really sad to have to use it in manual... May be the K-3II successor will be a mirrorless so no more focus issue.

One question about the 18-35, do you have sharp edges if you shot a landscape at 18mm when focusing using LiveView?
On my copy, I have no problem between 24 and 35mm, but between 18 and 22mm, edges are not sharp at all (not in focus I think) when using LiveView. And focusing using the viewfinder is total random...

[AntonioS]

Yes this is really sad to have to use it in manual... May be the K-3II successor will be a mirrorless so no more focus issue.

One question about the 18-35, do you have sharp edges if you shot a landscape at 18mm when focusing using LiveView?
On my copy, I have no problem between 24 and 35mm, but between 18 and 22mm, edges are not sharp at all (not in focus I think) when using LiveView. And focusing using the viewfinder is total random...

For what I remember, yes. And I totally agree about using manual focus. What amazes me is the fact that still today we don't have a convincent explanation about the problem's cause.

[Digitalis]

I have no problem between 24 and 35mm, but between 18 and 22mm, edges are not sharp at all (not in focus I think) when using LiveView.

Your copy might be decentered.

What amazes me is the fact that still today we don't have a convincent explanation about the problem's cause.

I don't think we ever will. Sigma will say it has something to do with Pentax AF and Pentax will say it has something to do with Sigmas electronics.