Strange focusing

[edumad]

Hi all.

I was doing a side-by-side comparison between the Tamron A2 135mm F2.5 and Pentax Tak bayonet 135mm F2.5, because I need to trim my collection and want to make a decision.

I noticed this strange behaviour while focusing both lenses...

I try to focus on the far distance building with the little tower (upper right corner in these heavy crops).

Tamron Adaptal-2 135mm F2.5, ISO100, F4, 1/320


SMC Tak bayonet 135mm F2.5, ISO100, F4, 1/400


Both show the rooftop (much closer) in focus while the point of focus (circa infinity) is not in focus, neither is the space between (or the space between roof and camera btw).

I checked, I didn't mix shots. I focused using the automatic focus detection, but to my eyes it should be in focus or close. The focus scale does not lie though, the focus was hitting the edge on both. Going back does not improve focus.

Funny thing is, the roof top is in better focus than when I tried to focus the roof (not a bit, noticeable...).

What the heck is going on?


On a related note, my opinion, based on these copies and the few shots taken is that.
The SMC tak bayonette is far less of a dog than many say. But is it noticeable softer than the SMC-M F3.5 at all apertures (at least F3.5-.
The Tak-bayonette is sharper than the Tamron at F2.5, but this seems to change at higher apertures. However the Tak-bayonette showed PF which I did not see for the Tamron under the same conditions. This is probably a result of the tak-bayonete being a budget lens with poor coatings?

[ManualFocus-G]

Looks to me like your adapter is too thick and won't allow infinity focusing

[aoleg]

[edumad]

For the Tamron I used two different adaptors (K and KA). The tak bayonete is K mount, so not adapter.

But this is happening on both lenses...
Hiting infinity or just close I think won't make a difference.

[ManualFocus-G]

Does the camera reach infinity with any lenses?

[edumad]

It does, although I have had less than satisfactory results with many lenses.
For example, just recently the Pentacon 30mm surprised me with its sharpness but at infinity results were poor...

I think I'll grab a couple of lenses and test them out.
Could it be that the camera fell out of alignment?
If this is so, I don't see how this could explain such a big problem with FF, but at lower focal lengths nothing as noticeable occurs.

[ManualFocus-G]

What camera are you shooting with? Sounds like the problem is there unless ALL your lenses are calibrated incorrectly, which is unlikely unless you bought them all from the same person...

[edumad]

Pentax K10D

[xjjohnno]

Hope you haven't forgotten to make sure the focus diopter is set up ok. A minor brainfart on my part a while back had me asking myself similar questions.

[visualopsins]

Atmospheric effects of heat rising from foreground buildings does blur focus at farther distances.

[scsambrook]

Your post seems to say the lenses were turned right back to their infinity stops, and if that was the case then either BOTH lenses have infinity setting problems or there's a problem in the camera's focusing system. Having two lenses with identical faults is a bit unlikely.

The green light is unreliable because it usually comes on BEFORE the peak of focus is reached. It can be on when the image still isn't critically sharp. But if both lenses were set fully back to infinity, that can't be the problem. The next likely culprit is the removeable focusing screen, especially if you've recently had it out for some reason. Check that it's FULLY latched up home. It ought to re-seat correctly automatically, but . . . The final nightmare is that the mirror isn't seating properly, but that's the least likely of all.

I doubt heat-blur is the cause, if the focus were on infinity, I think the near roof would be slightly de-focussed.

A couple of quick shots with the lenses set back to infinity and aimed at the moon should either acquit or convict them!

[thePiRaTE!!]

[edumad]

Ooops.
By lower focal lengths I mean smaller focal distances... sorry.

[edumad]

I'm pretty sure that last time I really did hit infinity, but maybe I was wrong.
I shot this with the SMC tak bayonette focus indicator came on, which was very very close to infinity.
135mm F4 ISO100


And I shot this one, just pushed to infinity.
135mm F4 ISO100

What I observed was that the autofocus detector is actually active within a certain range (while you change the focus). The photo from farther towards infinity (in that range) was better, but still not as good as infinity. Infinity itself was not detected to be in the range of focus.

The Tamron 24-135mm at 135mm gives this while MF
135mm F5.6 ISO100
From the scale it seems to be on infinity


But this lens has more slack to go beyond infinity, this is taken at the limit of the lens past infinity


It isn't supper obvious but the rooftop is slightly sharper here then when infinity is focused.


So the problem seems to relate with the auto focus detection. However super close to infinity should be well past the roof... which makes it very strange that the roof is so in focus. At F4, the roof should not be at or past the hyperfocal point either.

[ManualFocus-G]

Yep, that's why I like these new adapters where you can fine tune the focus confirm

[scsambrook]

Hi Edumad! I also have K10D and gave up bothering with the green light indicator ages ago. It simply isn't reliable. I also have a K100D and find exactly the same with that. The image on the screen is the only reliable judge of focusing accuracy and the promotion of 'focus confirmation indicators' does none of us any good. Whatever the theory, the practice just doesn't work out. [For Pentax,at least - I can't speak for other makes]

The purpose of the green light really seems to be to give you a visible confrimation that the autofocus mechanism has "locked on" correctly when you're using AF lenses. Even then, it's not a truly reliable indicator. If I use my Sigma AF 55-200 zoom at 55mm I get the confirmation light. If I then zoom out to 200mm, the on screen image is often less than perfectly focused and the resulting file is also slightly defocused. So I do focusing at 200mm and then zoom back to compose. That works every time. [Yes, I know in theory some zooms will shift focus going back from the longest setting, but none of mine actually do that]

So, forget the green light, set the eyepiece adjustment correctly, and then trust your own eyes!