[BigMackCam]

To note something again in my started topic: ;-)

I've had 3 (!) K3 III
And two 16-50 PLM.

...

I ve have similar problem with the FA35 2.0 and DFA 50 1.4 wide open on the K3 III... not as prominent as with the new 16-50...but in this way that I don't use other points than center....

EDIT: If your testing method is consistent...

I find that troubling. The D FA50/1.4 is a very well-corrected lens according to reviews, so it shouldn't be displaying optical behaviours that confuse the PDAF system. For me, that points the finger towards the camera more than the lens. Perhaps there's more work required with Pentax' first implementation of this new AF module or the firmware controlling it...

---

Last edited by BigMackCam; 05-19-2022 at 12:02 PM.

[licht96]

Ok..I must be careful with my claims! :-l

I was not testing the DFA 50 1.4 in that way like the 16-50 PLM.

I shot some test shots out of the hand: portrait of my daughter with f 2.0 with some outer AF points....

For the 35 2.0 I took an older paper aiming set (not the one shown in my detailed analysis some sides before).

When I've more time, I will test them in a... lets say: better way.
But frankly speaking: I'm tiered of testing.

When it's not working the usual way, like I do my portrait...

Good side of this: I'm using more and more live view with the smallest AF point.
Or with touch. It's fantastic working, accurate and sharp.

The better face detection helps, too. Not tat 1.4... but 2.0 :-)

[sebberry]

It's the compromises and inconsistencies that are becoming a Pentax hallmark.
Can it not 'just work'? or is that an unreasonable expectation?

[BigMackCam]

Ok..I must be careful with my claims! :-l

I was not testing the DFA 50 1.4 in that way like the 16-50 PLM.

I shot some test shots out of the hand: portrait of my daughter with f 2.0 with some outer AF points....

For the 35 2.0 I took an older paper aiming set (not the one shown in my detailed analysis some sides before).

When I've more time, I will test them in a... lets say: better way.
But frankly speaking: I'm tiered of testing.

When it's not working the usual way, like I do my portrait...

Good side of this: I'm using more and more live view with the smallest AF point.
Or with touch. It's fantastic working, accurate and sharp.

The better face detection helps, too. Not tat 1.4... but 2.0 :-)

Thanks for that clarification. I've updated my previous post accordingly...

You're quite correct, consistent testing is essential if the same conclusions are to be drawn. Nonetheless, what you've noticed with the D FA50/1.4 may be a useful indication. For what it's worth, I was recently using my K-3 and HD DA20-40 - a lens I've not been especially fond of in the past, though it's certainly a decent bit of glass. On this occasion I was happy with focus accuracy using the central focus point (more often that not, I'm a focus-and-recompose guy), but unhappy with the accuracy when using off-centre focus points. So, I went through my usual AF fine adjustment routine, this time checking both centre and a couple of off-centre AF points instead of my usual centre-point-only approach. By the time I'd finished, I was happy with accuracy from all the focus points. Now, with the HD DA20-40, I very much believe field curvature plays a part, and I suspect that a small, acceptable inaccuracy at the centre translates to a larger, less-acceptable error away from the centre. I don't know the characteristics of the new HD DA16-50, nor the D FA50/1.4... but it's something to consider...

As for being tired of testing, I understand completely... but do please bear in mind there are numerous existing K-3III owners here - and there will be many more in future - who would greatly appreciate any effort a fellow member such as yourself might expend, if and when you're able to devote the time to it...

---

Last edited by BigMackCam; 05-19-2022 at 04:12 PM.

[Kobie]

I just ran some testing on this with the 16-50 PLM. Using AF SEL and moving the points away from the middle box it was out of focus at all outer AF points... BUT, switching to full auto and setting the camera up to have a subject at the frames edge so it focused only at the far edges of the frame, it was sharp for some reason. It was only sharp using full Auto. Moving the points manually produced very soft images. Focus from center and recompose had no issues. My conclusion is that there is something going on between the AF location and the lens when the camera itself isn't in full control.

[Euan]

@kobie: That's interesting! Do you think that it may be that in full auto mode the camera is choosing a compromise between several AF points, and maybe that's resulting in sharp images in some cases?

@licht96: Wow, thank you for that update. That's very disappointing that you've tested that many copies and still seen the same issues. And even more disappointing that you've seen it (to a lesser extent) with other lenses.

I'm new to the K3 III, but I see that the last firmware update was fairly recent and it dealt with autofocus stuff. So maybe there's hope that there will be another one that will address this problem.

[Paul the Sunman]

By the way: I mentioned it many times here... I ve have similar problem with the FA35 2.0 and DFA 50 1.4 wide open on the K3 III... not as prominent as with the new 16-50...but in this way that I don't use other points than center....

I have noticed that the K3iii + DFA* 50 f/1.4 combination is more reticent to grab focus at off-centre points (at f/4 in my quick test, maybe 1.5 m distance), and when it does the results are a bit hit-or-miss. I can see this by repeatedly re-focusing using these points; each time, the focus shifts a little, so it is not reliable. I'm sure it is a camera issue.

In situations like this, the disinclination of tech companies to be honest and open with their customers is a blight on the industry. If they acknowledged the issue and gave an indication that they are working on it, we would have more confidence in them, not less.

[Settesz]

Could you check if AF with select point works correctly on older body?

I just ran some testing on this with the 16-50 PLM. Using AF SEL and moving the points away from the middle box it was out of focus at all outer AF points... BUT, switching to full auto and setting the camera up to have a subject at the frames edge so it focused only at the far edges of the frame, it was sharp for some reason. It was only sharp using full Auto. Moving the points manually produced very soft images. Focus from center and recompose had no issues. My conclusion is that there is something going on between the AF location and the lens when the camera itself isn't in full control.

[StigVidar]

New firmware is out. Is it fixed? I have tested a little bit with the lens at 16mm and it is not fixed. Perhaps it is better than before, but I am not sure.

[BigMackCam]

New firmware is out. Is it fixed? I have tested a little bit with the lens at 16mm and it is not fixed. Perhaps it is better than before, but I am not sure.

It's taking Ricoh quite some time to address this I wonder if they're actually aware and actively working on it? It must be costing them one or two sales here and there, of both the K-3III and the HD DA16-50...

[licht96]

Hopefully good news: After a long discussion with the German Ricoh Hotline with no result, I was contacted by a more technical involved expert, who is also a official "Ricoh representative" in the German forum.
He received my documentation (already translated) and agreed, that this is so deep in technical question, that it should be send directly to Japan.
That's it... that's what I wanted from the first day and after three copies of cameras and two of the lens: confirmation that the problem is addressed!
May be there is a result/answer, that there is no chance to solve it due to physics... or the chance to solve it by fine adjustment for some additional points.... or a new lens information that includes the field curvature (or whatever).
It's addressed... and that makes me a little bit calm now.
(by the way I make the "joke", when he will send something to Japan: "Please give a hint, that everyone would enjoy the name of the User Mode in the display, like in previous Pentax cameras" ;-) )

[sebberry]

Hopefully good news: After a long discussion with the German Ricoh Hotline with no result, I was contacted by a more technical involved expert, who is also a official "Ricoh representative" in the German forum.
He received my documentation (already translated) and agreed, that this is so deep in technical question, that it should be send directly to Japan.

Hopefully their discoveries based on your camera are something that they can push out as a firmware update to everyone.

I get the whole field curvature thing but there's really no reason it shouldn't be focusing properly at that point otherwise it completely defeats the purpose of additional focus points.

[StigVidar]

On my camera there are similar focus issues with the DA11-18/2.8 so I think this is more like a general problem with the new af-sensor in K-3 III

[UncleVanya]

On my camera there are similar focus issues with the DA11-18/2.8 so I think this is more like a general problem with the new af-sensor in K-3 III

Did the 11-18 work will with previous bodies for you on off axis focus points? I have a k-3 (original) and KP and I haven’t extensively tested this as I tend to use center point focus and recompose, but I don’t remember any issues with those cameras use of the lens. Likely I’m not using it the same way.

[house]

On my camera there are similar focus issues with the DA11-18/2.8 so I think this is more like a general problem with the new af-sensor in K-3 III

Does the "full auto" solution work for you? If it does the poor af should be a software issue fixable with firmware.

The af user- mode connection is a curious and hard to understand decision.