[bjan]

Frontfokus bei Pentax K-5 und K-r. Wer ist reproduzierbar betroffen? Und wer nicht? - DSLR-Forum

Just to show that we are not fighting alone here ! Cheer up !

[Digidoc]

Since newest camera models (e.g. K5) have the AF fine adjustment function, it is recognized that Autofocus per definition cannot be correct always. We have to appreciate that top models have so many different tools and functions that we more frequently have to abandon the idea that "automatic" is not ideal under all circumstances and that we have to go back to old-fashioned (sic!) handwork.

[simico]

For what it's worth, I just ran a quick series of tests to try and quantify the threshold of tungsten color light at which the front focus manifests itself. ... snip ...
So I tried 4EV again and the focus was fine again. Down to 3EV, the focus was off. I repeated this a few times and have concluded (that at least for this particular combination of K-5 and FA43), at around 3EV there is front focus.

Thanks for the test. Seems like I have round 3-4EV in my living room when the lights are turned on, because the front focus issue always happens with lights on.
Too bad I sold my K-m a few weeks ago, it didn't have any issues at all (no stains, no low light tungsten AF issue, no flash issue, etc). Maybe I should have sticked to it instead of buying the new flagship

[dgaies]

You're welcome. I'll give a try tomorrow once the sun comes back up

Well, I did a quick test and it seems that there is a slight amount of FF at low levels of natural light (<3EV) as compared to the same setup with ample natural light (>5EV) - although it doesn't seem as severe as with artifical light. As others have pointed out, the AF locks very quickly even in very low light as compared to previous bodies. Perhaps in an effort to make low light AF locking faster, too much accuracy (in low-light) was sacrificed in the process. Hopefully a firmware update can allow for a better balance between speed and accuracy in low light situations.

As a side-note, I decided to do a quick test with basically no natural or artricial light (-4EV) using only the AF-assist lamp to focus. The focus was spot on. That said, the target was very close (<3ft, <1m) and it's important to keep in mind that as the target distance goes up, the amount of light the AF-assist light puts on the target focus area drops off quickly.

[Wheatfield]

Thanks TOMMOT you are exactly understanding my " sharp edges" here and there

Is that what you call acting like a troll?

[Tommot1965]

just a thought...has anyone still got a K5 with the original firmware...not 101...be interesting to see if those body's still exhibit the FF..


cheers dgaies for doing the low light daylight test...and yes Id agree the FF is not as severe as artificial low light.. but as you point out, its still there...piss poor effort from pentax really...surely they are supposed to test this sort of thing aint they??!!!

anyway, i aint got nothing else to say about the whole balls up..unless anything new comes to light...all the best guys....

[janneman]

Since newest camera models (e.g. K5) have the AF fine adjustment function, it is recognized that Autofocus per definition cannot be correct always. We have to appreciate that top models have so many different tools and functions that we more frequently have to abandon the idea that "automatic" is not ideal under all circumstances and that we have to go back to old-fashioned (sic!) handwork.

But what AF adjust is about is FF or BF in general or with lenses.

[janneman]

.. but as you point out, its still there...piss poor effort from pentax really...surely they are supposed to test this sort of thing aint they??!!!
)

Did a few more tests, results:
- I have stains
- I have no FF, just a slight BF

Conclusion: the low and artifial light compensation called "+" by Pentax is in fact a thingy on the sensor that looks like a string of pearls.

On a more serious note, I think this is more or less comparable with the stain issue, not everybody seems to suffer from it but a sufficient large number of users do suffre enough for it to be officially called "an issue".

The problem we have with threads like this is that there is a fierce "You fanboy" vs "you basher"debate....
While the fanboys never seem to accept that there is or might be a problem and calle everybody who reports a problem a "basher", many of those who report a problem fail to see that if there is something wrong with a few cameras it need not affect ALL cameras and thus call those who honestly say "no probs here"a fanboy,.
(For the bashers, there are a few more reports of accurate AF out there, even in low light, for instance DPR review is quite happy with the K-5 AF)

By the way, I sent an email to the store about my stains (again), the store is open for half an hour now and still no reply....

[blende8]

I haven't read the whole thread, but just wanted to tell you my results:
For my K5 it is a combination of low light and low contrast.
When it is really dark (EV ca. 1) and the contrast is not simple B/W, the K5 produces a strong front focus, whereas the K20 is focussing correctly!
If it is getting slightly lighter or the contrast is better (B/W testsheet), everything is fine.

[pcarfan]

By the way, I sent an email to the store about my stains (again), the store is open for half an hour now and still no reply....

Hey! Janneman...your theroy may be right....I finally got to check my sensor for the pearls and I hope I did it right (put the lens at f22 and again at f11 and took a shot of a plain sheet of white paper) and there wasn't even a spot on the sensor.

[jani80k]

Just bought a K-5 - same problem. K-x focusses properly, K-5 fails even after many many tries does not get it right. Pictures were taken under low light tungsten circumstances.

[deejjjaaaa]

for instance DPR review is quite happy with the K-5 AF.

Did they actually test AF using something like FA 50/1.4 @ 1.4 under the (dim or not) tungsten light ? I doubt.

[dgaies]

for instance DPR review is quite happy with the K-5 AF)

Did they actually test AF using something like FA 50/1.4 @ 1.4 under the (dim or not) tungsten light ? I doubt.

Well, overall the AF of the K-5 is excellent, a solid improvement over the K-7, which itself was an improvement over the K20D.

I don't know specifically what kinds of tests they (DPR) ran on the AF system, but it seems like the (major) problems come into play when shooting low levels of tungsten light on low contrast targets. I was shooting under these conditions last night and providing the target under the focus point had good contrast, I was having very few issues under these conditions. If I was aiming at a target with less contrast, FF was more likely to occur. But to be honest, in low light and low contrast targets, I've never expected the AF system to be perfect. Please don't misunderstand that statement to mean that I don't think there is room for improvement or that there isn't potentially an issue. I'm not "making excuses" for Pentax, I'm just not ready to call the K-5 "defective" because in the vast majority of my actual shooting thus far I haven't had any issues.

[blende8]

If I was aiming at a target with less contrast, FF was more likely to occur. But to be honest, in low light and low contrast targets, I've never expected the AF system to be perfect.

I agree, but note that the/my K20 is doing much better under these circumstances!
I thought I bought the ultimate low-light machine.
The bottle-neck is the AF, it seems.

[dgaies]

I agree, but note that the/my K20 is doing much better under these circumstances!
I thought I bought the ultimate low-light machine.
The bottle-neck is the AF, it seems.

I agree, it does appear that the AF is the limiting factor in the K-5 being the "ultimate low-light machine". I no longer have any of my previous bodies to compare directly to, but the general consensus seems to be that some of the older bodies (the K20D for example) was more accurate under these conditions. If this is the case, then it certainly suggests that something is wrong and needs to be addressed. I suspect that in an attempt to speed things up, it's possible that the body stops trying to adjust the focus too soon and in low light/contrast situations this leads to a FF condition. I realize that's speculation at this point, but if that is the case it would lead me to believe the problem could be addressed via firmware. If there is a hardware issue, fixing the issue becomes more difficult.