[Digitalis]

I have a mothballed optics bench at my disposal, when I get it set up again I intend to put this lens through its paces. However from strict visual scrutiny I can tell the DFA*50mm f/1.4 is easily in the top sharpest 50mm lenses I have ever worked with. The problem with ultra-sharp lenses being is that with such high contrast and acuity, focus errors are even more apparent.

[pschlute]

guys, just in case, has anyone tested this lens on focusing chart?

Yes. When I first got it. It needed the same adjustment as most of my other lenses ie the sensor requires the adjustment not the lens

[Paul the Sunman]

Yes. When I first got it. It needed the same adjustment as most of my other lenses ie the sensor requires the adjustment not the lens

I used to routinely test for front or back focus when I got a new lens, but I never found any worth worrying about. So I don't bother anymore unless I detect some issue in normal use.

However, I have the DFA* 50mm on order at the moment. When it arrives, I'll test it carefully (and repeatedly), based on the interesting insights of this thread. Thanks everyone.

[pschlute]

I used to routinely test for front or back focus when I got a new lens, but I never found any worth worrying about

With my K1 all my lenses require +7/+9 to correct backfocus.

[clackers]

With my K1 all my lenses require +7/+9 to correct backfocus.

Your AF module isn't seated perfectly.

I understand that quality control is a Bell curve, that up to ten is within tolerance, that Pentax lets us do this for ourselves (consumer grade Nikon does not) and there will be someone, somewhere who needs -7/-9.

[monochrome]

I don't think anyone has mentioned the recent revelation that central point AF on the K1 is NOT the same as Spot AF. The latter is supposed to be more accurate (or was it sensitive?). Has anyone checked to see if the consistency of AF differs between the two?

SEL-1 is apparently a larger area than SPOT on K-1 PDAF but no one claims it is more accurate. It is more precise.

Supposedly in PDAF the Spot is small enough to OVF identify and focus on an eye without catching the lashes or edge of the nose. Supposedly in CDAF a K-1 one can detect and lock on an eye according to Bruce Banner. Such feature is not documented.

[clackers]

Supposedly in CDAF a K-1 one can detect and lock on an eye according to Bruce Banner. Such feature is not documented.

"Face Detection" itself is documented. That it is de facto 'Eye detection' like on a Sony is an interesting report.

[beholder3]

With my K1 all my lenses require +7/+9 to correct backfocus.

If that is the case, I personally would use develop mode to give it a general +7 adjustment as baseline.
Resurrecting Pentax firmware hacking - Page 9 - PentaxForums.com

[Lev]

Guys, one thing is to correct constant front or back focus, eg aligning lens to the sensor but another is how it behaves after that, how the focus may shift shot to shot. This is more interesting because unlike microadjustments we cannot get in PDAF.

[beholder3]

Guys, one thing is to correct constant front or back focus, eg aligning lens to the sensor but another is how it behaves after that, how the focus may shift shot to shot. This is more interesting because unlike microadjustments we cannot get in PDAF.

The internet is full of fanboy nonsense about the precision and reliability of AF systems. They all struggle and only have a probability to "nail focus".
Here is a comparative test: 7 Autofokus-Systeme im Test - PC Magazin

Even the winner in precision, the CDAF-driven E-M5 gets only 152/200 images in really correct focus (>=95% of possible target focus plane resolution). That is still 25% images with a not-so-great focus.
The Sony A7RII nailed 108/207 images, that is only 50% tack sharp hit rate for static subjects.
The Nikon D500 nailed 9/120 or 8% of the static shots.
The K-1 was reported as having less good results.

Still autofocus "limitations" today certainly are no valid excuse for poor or no images. The above is just statistics for the nerds.

[Lev]

beholder3, good info thanks!


I know this is not directly related to the PDAF auto focus accuracy but aslyfox did remind me to have a look at review of DFA* 50mm and compare. There is a section called "focus shift". I've done this test and the focus point is just frozen in place in all apertures, good news and thanks aslyfox!

[c.a.m]

Here is a comparative test: 7 Autofokus-Systeme im Test - PC Magazin

Thanks for the link.

I read a Google-translated version in English, which might have missed some subtle points of language. Please correct me if I've misunderstood the test report.

The test is interesting in that it included a variety of cameras, including mirrorless and DLSR, from several brands. The basic message is understandable: autofocus is less than perfect and users should expect some mis-focused shots, perhaps as many as 25% of shots. Some cameras have worse AF performance than others.

However, there seems to be at least two important technical flaws. It's not obvious that the test procedure included preliminary steps to calibrate the cameras through AF Fine Adjustment or AF Micro-adjustment, for those cameras that have such a capability (Pentax K-1, for example). If this was done in fact, then the article should have mentioned it and briefly described their calibration procedure. If the cameras were not calibrated, the credibility of the test is questionable.

Secondly, the article doesn't seem to distinguish precision and accuracy, a common omission in this type of test report.

The report reminds me of a series of technical analyses in 2012 by Roger Cicala at Lensrentals, for example: https://www.lensrentals.com/blog/2012/07/autofocus-reality-part-1-center-poi...shot-accuracy/

- Craig

[beholder3]

Most AF tests are poorly documented and that one is not different.
But what you can gain from it - even if none of the cameras would have been calibrated - is two things:

1) results do vary by shot on a regular basis even for static subjects shot from a tripod (calibration changes the target hit rate but has no influence on variation)


2) CDAF is very far from being 100% reliable to find the correct focus plane consistently

[c.a.m]

The following slight divergence from the original theme of this thread is related to camera AF accuracy but not directly related to the D FA * 50/1.4 lens.


Here is another test report dated 2017 produced by Image Engineering; the tests were performed during 2014-2016. It's not clear in the paper whether any of the test data is related to the test reported in PC Magazin above. ColorFoto is mentioned in both articles, and this one acknowledges the "database of ColorFoto."

https://www.image-engineering.de/content/library/conference_papers/2017_03/a...erformance.pdf

Interestingly, the K-3 II is included.

[pschlute]

If that is the case, I personally would use develop mode to give it a general +7 adjustment as baseline.
Resurrecting Pentax firmware hacking - Page 9 - PentaxForums.com

AF/FA needed is within the -10/+10 the camera allows without delving into the service menu. That would set my nerves on edge !

---------- Post added 12-20-18 at 11:07 PM ----------

Guys, one thing is to correct constant front or back focus, eg aligning lens to the sensor but another is how it behaves after that, how the focus may shift shot to shot. This is more interesting because unlike microadjustments we cannot get in PDAF

This is why when we see an interesting subject for a picture, it is best not to take just one picture of it