Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
12-23-2014, 07:12 AM   #76
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 10,928
QuoteOriginally posted by Digitalis Quote
Nope - Sigma hasn't made any statements regarding focus accuracy and this particular lens.
Perhaps because there is no need for any statements?

A lot of manufacturers are not making any statements regarding focus accuracy, even though some of their copies need AF adjustments.

I note that Heie never followed up with the result of the planned further investigation. Given that he praised Sigma for their excellent communication, I must say I find it disappointing that we never heard about the results of trying to shed further light on the matter. What the review documents, still leaves all sorts of scenarios open, from a lens design that does not work with Pentax AF (highly unlikely) to not properly using the lens dock (I'd say unlikely too, but then we just don't know) and extreme misfortune with respect to the quality of the test samples.

12-23-2014, 08:43 AM   #77
Site Supporter
Site Supporter
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 10,158
QuoteOriginally posted by Class A Quote
Perhaps because there is no need for any statements?
Or perhaps Sigma have been sticking their fingers in their ears over the issue, and make conciliatory noises when someone tries to speak to them directly.. after all APS-C isn't a professional format, and Pentax hardly has the kind of market share to tilt the balance of sales as far as they are concerned. The strange thing is that complaints about this lens are rather sparse - I get the feeling this lens isn't all that common. I have never heard any comments on AF accuracy with this lens from Sigma camera users (And last time I looked Sigma SLRs have categorically worse AF than Pentax) As far as direct evidence goes I know I'm only one out of three other working professional photographers in my city who own one and they both have had issues with it.

Last edited by Digitalis; 12-23-2014 at 08:55 AM.
12-23-2014, 07:18 PM   #78
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 10,928
QuoteOriginally posted by Digitalis Quote
The strange thing is that complaints about this lens are rather sparse - I get the feeling this lens isn't all that common.
Well, that's one explanation. There is an (obvious) alternative explanation.

Now that Heie posted more about his AF experiences, I'm inclined to believe that there is some issue with the lens in K-mount.

Whether there is a systematic K-mount problem or they messed up a batch of lenses that were fitted with K-mount is impossible to say from my vantage point.
12-25-2014, 02:33 PM - 1 Like   #79
Junior Member




Join Date: Oct 2013
Location: Canberra, Australia
Posts: 31
A possible cause of AF misbehavior in Sigma 18-35/f1.8

I haven't visited PF for awhile and unfortunately haven't seen this thread before. I've been using a Sigma 18-35 (first with K30 and K-01, and now with a newly acquired K3) since August 2014. About 6 weeks ago, I started a thread in DPReview forums describing some simple experiments aimed at finding the cause(s) of the focusing problems:
Sigma 18-35/f1.8 - a possible cause of AF misbehavior: Pentax SLR Talk Forum: Digital Photography Review . Below is the starting post of that thread, which I believe is relevant to the discussion here. A report on using this lens with K3 will follow shortly.

*****
Inspired by the excellent in-depth review on Pentax Forums (Sigma 18-35mm F1.8 "Art" Review Posted - Previews and Reviews | PentaxForums.com) and the subsequent discussion here (Sigma 18-35 F1.8 "Art" In-Depth Review Posted: Pentax SLR Talk Forum: Digital Photography Review), and being a (partially) happy owner of a copy of the Sigma 18-35/1.8 lens that is consistently mis-focusing in PDAF, just like the lenses described in the reviews, I decided to try to explore the problem a bit further.
As the first step, I tried to quantify the frequency of correct and incorrect focusing by taking multiple photos of my "test object" (an electrical pole some 20 meters away) in series of 10 shots. To make sure that each focusing is independent, I was focusing to an object much closer (in ~0.5m) between the shots. At first, only 1-3 shots in each set of 10 were in focus, but as I repeated the sets the frequency of properly shots increased. It was intriguing, and I searched for similar observations posted by other users of that lens. In one post (Sigma 18-35 F1.8 "Art" In-Depth Review Posted: Pentax SLR Talk Forum: Digital Photography Review) a similar experience is reported by a person using the Nikon version of this lens. Following his findings, I gave the lens a more intense quick refocusing "training". The result was impressive: the frequency of correctly focused shots increased from 10-30% to 80-90%.
Next day, the lens behaved a bit worse at first, but went to "almost normal" again after brief "training". I've been using the lens daily for 5 days since the first "training", and getting about 90% of correct focusing. It may be a bit early to say that the problem is overcome, but I'm reasonably confident using this lens with PDAF now.
So the focusing problem appears to be of primarily mechanical nature, not the firmware issue as I initially thought. But then there remains a mystery of why this lens was always focusing correctly in CDAF, but not PDAF. I think the answer is in the difference in the focusing algorithm that defines the lens barrel motion. In CDAF, the camera appears to scan the focus range around the previous setting for the point of maximum contrast. There is a noticeable barrel shift every tine the focus button is pressed, no matter how many times you re-focus, even if the previous focus setting is perfectly accurate. On the other hand, in PDAF (single AF) the barrel shifts noticeably when you focus the first time, much less the second time, and shows no detectable motion after that. It appears that the focusing gear is a bit "sticky" and requires either substantial motion such as CDAF scanning, or "training" by repeated re-focusing.
Of course, the mechanical nature of the misfocusing problem does not exclude the existence of the possible firmware bugs, but their role is likely to be secondary.
It would be great if the other owners of misbehaving copies of the Sigma 18-35 tested their lenses using this or similar procedures. I look forward to their reports.

12-25-2014, 03:37 PM - 1 Like   #80
Junior Member




Join Date: Oct 2013
Location: Canberra, Australia
Posts: 31
Mechanical problem or not?

QuoteOriginally posted by Class A Quote
No one seems to acknowledge that even the PF copies of the lens worked perfectly with CDAF. In other words, the lens will focus correctly, if it receives the right signals from the camera. This, with a high probability, rules out mechanical issues.
An important step is missing here. The lens will focus correctly if it receives the right signals from the camera and executes them accurately. The latter may not always be the case.

QuoteOriginally posted by Class A Quote
It is conceivable that the camera would stress the lens differently when using PDAF compared to CDAF, but this does not make much sense in engineering terms and I'm rather sure that any answer to this puzzle will have nothing to do with the lens motor or its associated mechanics.
I think this does make sense. The movements of the lens focusing mechanism that I've seen in the experiments with PDAF and CDAF described in the post above differ. In case of CDAF, the system always scans a substantial range in the distance scale (from farther to closer), then goes to the farther point again, and from there to the point of maximum contrast. What is important is that the same scanning is repeated each time the AF button is pressed, even if the object is already perfectly in focus. The movement in the case of PDAF is different. It is substantial if the staring focusing point is way off, but repeated pressing of the AF button causes very small movement, or no detectable (audible) movement at all.

If the mechanical response of the lens was perfect (lens motion is executed exactly as directed by the camera), then both CDAF and PDAF would give reproducible results, which would be always accurate in the case of CDAF, and either accurate or not (depending on the AF adjustment) but still reproducible in case of PDAF. However, if the focusing motor has a hysteresis, the pattern would be totally different. The CDAF focusing algorithm circumvents the hysteresis by approaching the set point in a motion that mimics the focusing scan. The PDAF focusing algorithm, apparently, does not do that. If the focusing motor has a hysteresis of a few steps, it would cause the lack of reproducibility in PDAF, just as being observed.

Variability in focusing behavior between the copies of this lens can be explained by variations in the magnitude of the motor hysteresis, which, in turn, can be caused by mechanical variation between the copies of the motor, or even something as simple as variations in lubrication (which I suspect is the case).
12-25-2014, 07:49 PM   #81
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 10,928
First, thanks a lot for sharing your observations!

They may be the first step towards an explanation of what is going on.
It would be most appropriate for Sigma USA to engage more actively, but in the absence of any communication with them, we'll have to investigate ourselves. I hope you may be able to share your observations with Sigma Japan who may be unaware of issues and/or may get on the right path to address the issue.

If your observations apply, it may be the case that reproducing the issue with the well-used copies available to engineers is not possible and that it would be necessary for them to obtain a particular brand new copy from a certain batch to actually reproduce it. I realise that a good QC department would obtain a brand new lens from a certain serial range anyhow without such a suggestion, but perhaps Sigma Japan is just unaware of the issue or has classified it incorrectly (as a "shallow DOF" challenge and/or user error).

QuoteOriginally posted by sharik Quote
An important step is missing here. The lens will focus correctly if it receives the right signals from the camera and executes them accurately. The latter may not always be the case.
That's why I was careful to not use absolute terms and considered the different movement patterns between CDAF and PDAF in the next paragraph.

QuoteOriginally posted by sharik Quote
If the mechanical response of the lens was perfect (lens motion is executed exactly as directed by the camera), then both CDAF and PDAF would give reproducible results, which would be always accurate in the case of CDAF, and either accurate or not (depending on the AF adjustment) but still reproducible in case of PDAF. However, if the focusing motor has a hysteresis, the pattern would be totally different.
You may be on to something, but I find the ultimate cause you are offering as an explanation unlikely to apply. I'm not an expert, but I believe the precision required to accurately focus makes it highly unlikely to allow for any kind of hysteresis without messing up the focus result all the time (even when using CDAF).

Perhaps a more promising explanation could be based on different closed-loop feedback exit criteria for PDAF vs CDAF.

The combination of
  1. the lens motor sometimes missing a couple of steps (apparently depending on how "well oiled" the lens is due to regular use), and
  2. CDAF having a lower tolerance against remaining focus errors
would explain the behaviour observed.

The PDAF algorithm does not need a closed-loop feedback approach as much as a CDAF algorithm because in principle one initial reading and then actuating the lens once would be sufficient. In practice, there is of course a lot more feedback involved, but still it is reasonable to assume that the PDAF algorithm will have laxer exit criteria as it may assume that a correct actuation of the commands will result in correct focus.

CDAF, on the other hand, in particular if implemented in a naive way, cannot predict the correct focus position in the same manner and hence may make more attempts to get the focus right using the closed-loop feedback principle.

If that hypothesis were true then it would take a "cold"/"sticky" lens slightly more time to obtain CDAF focus compared to a "well-oiled"/"90% hit accuracy with PDAF" lens. The timing differences would be very small but might be measurable without using professional equipment.

Last edited by Class A; 12-25-2014 at 08:36 PM.
12-27-2014, 05:53 PM   #82
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 10,928
I just read this report by a Sigma 18-35/1.8 user:
"I've had this lens glued to my K-3 for weeks now. I think the lens is nothing short of fantastic, and I have had no AF problem with it at all. One other Pentax user I know has the lens and she too reports no AF issues with it."
So it very much appears that the lens design as such is sound and there is just an unfortunate number of subpar copies out there.

In other words, there does not appear to be a need to "stay away from the lens", if one has access to a store with a good return policy.
02-04-2015, 01:15 PM   #83
Site Supporter
Site Supporter




Join Date: Feb 2008
Location: Emeryville, CA
Posts: 1,119
For those interested in the follow-up by PF
Sigma 18-35mm Autofocus: A Second Look - Previews and Reviews | PentaxForums.com

02-06-2015, 09:22 AM   #84
Forum Member




Join Date: Feb 2011
Location: Vantaa, Finland
Posts: 83
My copy seems to focus quite realiably, except for 18-20mm range + infinity (or near infinity). 0% success rate there. It focuses on around 2m mark instead of infinity, and it varies a lot. In all other cases it works very well (maybe 95% success rate), even in low light. Now that I know this, it is easy to avoid, but lens of this price shouldn't have these problems.

Paired with K-3.
02-23-2015, 08:34 AM   #85
New Member




Join Date: Aug 2013
Posts: 24
May be I am lucky(so far). I do not have AF issues with this lens on my K-30. I shot mostly at 35mm and 1.8-2.0 in low light conditions and indoors. I had fantastic results. I didn't shoot much in the range below 35mm. I read users saying AF issues are more likely in high contrast conditions. To be frank I found AF on this sigma better than few Pentax lenses I own. I simply love this lens.
Not sure if it is related - I have not done a firmware upgrade on my K-30 for more than year. I suggest apprehensive buyers to rent the lens and decide your self.
03-13-2015, 09:06 PM - 2 Likes   #86
Veteran Member
MadMathMind's Avatar

Join Date: Dec 2013
Location: Houston, TX
Photos: Gallery
Posts: 1,717
Original Poster
So I've worked with my lens and the dock. I've used Lens Align and another technique I picked up somewhere on this forum: first focus in PDAF (viewfinder). Then switch to live view and attempt to focus. If the focus ring doesn't move, you're in business. Fiddle around until you get what you like.

My keeper rate is over 98% now. I've got very few issues pinpointing exactly what I want. An occasional shot goes errant here or there, but I've had that happen with both my FA77 and FA50 as well. From the photos I've shot in the past few days, I think only 2 of the Sigma went awry while one each of the FA77 and FA50 were misfocused--and the Sigma shots outnumber the FA77 and FA50 about 3:1 or more.

Part of my success has also been my increased understanding of how autofocus works. I now look for good focus points--turns out that giving the lens "something to focus" on has helped a lot. (I suspect that has not been the issue for Adam and Heie--I'm sure they know this!) I also have better understanding of DoF and think a bit more carefully about what would be a better focus point to center the DoF so that if the lens does misfocus a bit, I'll still get what I want.
03-15-2015, 05:42 PM   #87
New Member




Join Date: Mar 2015
Posts: 3
How are you focusing?

Given the conflicting reports on this topic I'm interested to hear how everyone is focusing. I was very disappointed with the performance of this lens while focusing with half-shutter but switching to the AF-C back focus button technique (thanks to Wired: https://www.pentaxforums.com/forums/38-photographic-technique/236970-perfect-...everytime.html) made such a difference that I'm really curious to see if the technique is the answer here.
03-15-2015, 07:47 PM   #88
Veteran Member
MadMathMind's Avatar

Join Date: Dec 2013
Location: Houston, TX
Photos: Gallery
Posts: 1,717
Original Poster
QuoteOriginally posted by aaronhbridges Quote
Given the conflicting reports on this topic I'm interested to hear how everyone is focusing. I was very disappointed with the performance of this lens while focusing with half-shutter but switching to the AF-C back focus button technique (thanks to Wired: https://www.pentaxforums.com/forums/38-photographic-technique/236970-perfect-...everytime.html) made such a difference that I'm really curious to see if the technique is the answer here.
Back button for me. I have used the AFS-C trick in the past but found it unnecessary. I use AF-S and click the button multiple times until I can't hear the lens move (hard with the Sigma, since it's so quiet, but you get the idea). I try to limit recomposition and use manual selection of the focus point. The rest is what I said above.
04-19-2015, 01:46 PM   #89
Veteran Member
MadMathMind's Avatar

Join Date: Dec 2013
Location: Houston, TX
Photos: Gallery
Posts: 1,717
Original Poster
My keeper rate had been great...until today. It was all over the place. I got maybe 30% in focus. Thing I don't get is that it would focus in one shot but if I hit the AF button, it would refocus and totally miss. Not sure what was going on today. I had no predictability at all. Looks like it's back to the dock again, see what's going on.

My FA77 missed a few times too. I'm wondering if something is up with my AF module. Perhaps it needs a cleaning.
06-06-2015, 03:56 AM   #90
New Member




Join Date: Oct 2010
Posts: 7
I`m happy user of Sigma 17-50 f/2.8 but I was thinking to update my lens collection and I was planning to buy Sigma 18-35 f/1.8. After reading all this threads about AF issue I will stick with 17-50. I hope new FW for 18-35 resolves all AF problems. Anyway IQ from this lens is tempting.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
af, aperture, art, art focus issues, f/1.8 art focus, ff, focus, issue, issues, k-5, k-mount, lens, lenses, pentax, pentax lens, shots, sigma, sigma 18-35 f/1.8, slr lens, thread
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
The Sigma 18-35 f/1.8 Art club scratchpaddy Lens Clubs 112 11-29-2018 05:05 AM
Sigma 18-35 1.8 art vs HD DA 20-40mm F2.8 - 4 Limited DC WR jrcastillo Pentax SLR Lens Discussion 111 09-22-2015 03:02 PM
Sigma ART 18-35 1.8 and K-30 jrcastillo Pentax K-30 & K-50 36 07-28-2014 09:42 AM
Sigma 18-35 f/1.8 vs the Pentax DA 16-50 f/2.8 someguy42 Pentax SLR Lens Discussion 9 12-06-2013 12:02 AM



All times are GMT -7. The time now is 09:12 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top