[Wheatfield]

Interesting question, Wheatfield. What are you getting at?

I'm puzzling away (instead of doing something constructive) trying to see the connection between WB and AF.

Please understand I'm not debating - simply asking.

Thanks!

The root problem is that the AF sensor is not colour agnostic.
An easy way to work around this is to tie an AF bias to the white balance setting.
Anyway, I believe this is about the only way to fix something like this, unless there is a way to make a colour blind sensor that will work in an AF matrix.

[Class A]

after having been reading a bit in older threads I have found the existance of the tungsten light AF with fast lensen for many years. (also in other brands) Why should we expect Pentax to solve the issue all of a sudden with a firmware update ?

We can expect that because every Pentax camera with a "+" in the "SAFOX" AF designation should be agnostic to the colour temperature. So yes, the problem existed for the K20D and older, but shouldn't exist for the K-7 and K-5.

The root problem is that the AF sensor is not colour agnostic.
An easy way to work around this is to tie an AF bias to the white balance setting.

The K-r apparently uses the WB setting to correct for the focus shift occurring under Tungsten light. By nature, this has to remain imprecise and will fail for incorrect WB settings and/or incorrect AWB. The K-5, however, has a special colour temperature sensor that should help to compensate precisely.

It is well-known that the K-5 needs an update for low-light AF. It gives up trying to find a precise lock too early in conditions that are just not dark enough to make the AF assist light come on. If the latter happens, everything is fine again.

Whether or not the FF issues reported here have something to do with the above problem, can be corrected via a firmware update, or point to a hardware defect, I don't know. I do know, however, that the K-7 and K-5 should show less FF under Tungsten lighting than say a K-x or K20D.

[filorp]

my k5 has the same issue... -10 daylight +5 tungsteen... more or less for each lens. There is a color sensor embedded into AF sensor whitch doesnt seams to work properly or it doesn't work at all.... because software of the camera doesnt know haw to use an information from this sensor.... the Pentax technician told me that i dont need accurate AF during the day because of small aperture being used.... so i need just to correct the af for the tungsteen light.... it was funny wasn't it?

[JamieP]

I've had similar issues to what the OP describes. Originally I thought I was having a problem accurately setting my focus adjustment so I created this thread:

https://www.pentaxforums.com/forums/pentax-dslr-discussion/124389-struggling-...justments.html

With more testing and shooting I've become convinced that the focus is varying greatly in different lighting conditions. Since I hadn't heard other users complaining about this, I was wondering if it was an issue with my camera. Ultimately I decided to exchange it since it had a stain in the center of frame, so I'm hoping my focus problems also go away with a new body. It arrives on Monday. We'll see...

[JamieP]

I'm curious how well AWB is working for users that are having focus issues. I was really looking forward to improved AWB coming from the K10D, but haven't found it to be very accurate so far. I'm wondering if the issues are related- i.e. if a faulty wavelength sensor could affect both.

[Todd Adamson]

I have spent the last hour trying to reproduce a front focus issue with my K-5 and FA77 in low, tungsten light. I can't do it. Focus was off in less than 10% of shots, and then it was slightly back-focused. I shot at f/1.8 and f/2.0, and I tried with the camera's WB set to Auto as well as tungsten. Seems to be nailing focus under all of these scenarios. Here is a representative shot:



Regarding AWB on this camera, I will just say this: I am going to start being a stickler for setting custom WB or at a minimum shooting gray card shots for everything I do.

[dreamfoto]

Why aren't there any samples of Front Focus problems in this thread? I didn't noice any misbehavior from the Auto Focus so far.

Should also mention I use mostly my manual lenses...

[simico]

This has been mentioned in at least two other topics. My post in one of them.

According to Pentax the next firmware will contain some fix for AF under tungsten. I really don't understand all the fanboys trying to dismiss real issues... oh wait, must be because of their blind fanboyism.

[Todd Adamson]

I really don't understand all the fanboys trying to dismiss real issues... oh wait, must be because of their blind fanboyism.

Not sure if you're talking to me, but I'm most definitely not a fanboy. I'm a Nikon shooter professionally; I got a Pentax system for fun, and am exploring whether it might also serve as a worthy backup if I need to be without my D3.

I'm also not trying to dismiss this issue. I was a bit concerned when I read this thread, which is why I went to test it right away. All I'm saying is that I can't personally reproduce the FF problem with my gear under the conditions mentioned by the OP.

[cpk]

not knowing any better, I did my focus adjustment for my lenses at the end of a long hall which required the use of tungsten light. I did this for both a K20D and a K-7; and I will do it again when I get my K-5. In light of the current discussion, this is probably the best approach for me, although it may not suit others. Most of my lens-wide-open photography is done under tungsten light; outdoors in daylight I am generally stopped down two or three stops where the depth of field should cover autofocus errors.

I am assuming that the difference in autofocus under daylight and tungsten light results from the different wavelengths of the source lights. If so, the problem may not be so easily solved.

[Fer]

Where are all the photos with the supposed problems?
With all due respect to bjan and Pentaxeros all talk and no photos means nothing.
I haven't encountered the kind of focus errors with my K-5 with my fast lenses and if there were, it could be for a number of reasons, user error being one of them.

The problem is not "supposed" but very real, and it does not just exist in tungsten, but also in candlelight. It appears that it is light with an orange/yellowish hue that fools the AF system.

Here are some self-explanatory images from the tests I ran. I'd encourage every K-5 owner to do this very easy test: no light except one candle, and a contrasty object next to it (in this case the K-x manual because it was at hand). The funny thing is that I've tested my K-x with the DA70 in this same scene and it nails focus almost every time.

In my experience, the K-5 is able to focus much darker objects than the booklet cover in this scene, but only if it activates the AF assist lamp. The problem is, in this scene, the camera doesn't think it needs the AF light to achieve focus, and indeed the AF confirmation comes very quickly. Unfortunately, it misses pretty much every time. I wouldn't be surprised if the problem was solved by manually forcing the AF assist lamp to work, but as far as I know there is no way to do this.

Unlike what Pentaxeros mentioned above, I'm not convinced this can be solved by simply dialing -10 AF microadjustment. I think -7 or -8 generally looks best (but not in focus, as shown by the images focused with LV), further than that it seems to get worse. Oh, and the AF microadjustment seems to mess with AF accuracy in daylight and at longer distances (as I found out this morning with my DA18-55)

All in all, extremely annoying, and something that Pentax needs to fix soon.

These were all taken with K-5 and DA70 or DA35 at the largest aperture allowed by each lens:

[dgaies]

The problem is not "supposed" but very real, and it does not just exist in tungsten, but also in candlelight. It appears that it is light with an orange/yellowish hue that fools the AF system.

Have you considered that candlelight, which had a color temp of 1800-2000K, might be outside the range for which the AF system was designed to work? It's certainly outside the range where AWB works.

[apisto]

Fer,

Do you have a K-7, and if so, does it respond the same way in this focusing test?

[Horst-Helmut]

Have you considered that candlelight, which had a color temp of 1800-2000K, might be outside the range for which the AF system was designed to work? It's certainly outside the range where AWB works.

I said this in another thread.

I was shooting at a snooker club and missed all but one shot because of this, the tables have compact fluoros. Deleted 50 or so shots. Same thing happened in studio in a moody film noir shoot, Hensel heads with halogen modeling lights, everything was front focused, even at 50mm, f6.7-11! Can't show the pictures because the model won't allow to put them online.

In the snooker club I tried to fiddle with the focus adjustment and it didn't help at all. I really don't think the AF should work that bad. That's completely unacceptable from a camera with a price tag of 1200e.

If I can't trust the gear to do what I'm telling it, then I think it really is time to change system. K-5 was the last chance I'm giving to Pentax. If these issues (flash overexposure, unreliable focusing under artificial light) are not being solved in the next 1-2 months it's byebye Pentax. If that is even possible through firmware.

Kind of shame, I reallyreallyreally like Pentax lenses and the ergonomics of the bodies but I need a camera that I can trust. I don't have time to look the focus after every shot I take especially in the clubs and gigs etc. where I usually have one chance to get the picture.

[Pentaxeros]

El problema es solo con lentes con enfoque de tornillo, y solo en bajas condiciones de luz de tungsteno.

Con lentes SDM o HSM, esto no ocurre, con SDM o HSM, el enfoque es perfecto en cualquier tipo de iluminación.

Pentax, corregirá esto mediante firmware en breve.

Saludos...