[tibbitts]

Surprisingly, manipulating the front of the 16-45 barrel from one extreme to the other, even at maximum extension, doesn't seem to produce visible misalignment in output - at least not at 6mp, which is what I last tested mine with. I should repeat the test now with more mp, but maybe someone else already has?

Paul

[glasbak]

Corner performance portret mode pentax 16-45 versus tamron 17-50 all at F1:4.0:

At 16mm decentered by gravity.
At 16mm rubber band centered.
At 17mm Tamron

At 28mm decentered by gravity .
At 28mm rubber band centered .
At 28mm Tamron

At 45mm decentered by gravity.
At 45mm rubber band centered.
At 50mm Tamron

Tests at 50 times focal length.

[tibbitts]

Corner performance portret mode pentax 16-45 versus tamron 17-50 all at F1:4.0:

At 16mm decentered by gravity.
At 16mm rubber band centered.
At 17mm Tamron

At 28mm decentered by gravity .
At 28mm rubber band centered .
At 28mm Tamron

At 45mm decentered by gravity.
At 45mm rubber band centered.
At 50mm Tamron

Tests at 50 times focal length.

Thanks for posting that! I should redo my tests now that I have more than 6mp available - maybe they would more clearly show a problem. It is hard to believe that all that play wouldn't have any consequences in performance.

When you say rubber-band centered, how did you accomplish that?

Paul

[glasbak]

When you say rubber-band centered, how did you accomplish that?

Paul

Just a wide elastic band around the barrel of the lens, focus/zoom becomes difficult, and don't even try autofocus.
This centers the wobbly front group for these tripod shots.

This problem with the disappointing corner performance is widely known, the dpreview test of this lens mentioned it also.
And even on my 6Mpixel *istD it shows on portret mode shots.
For landscape shots the decentering of the downward pulled front group is less harmful.

[tibbitts]

Just a wide elastic band around the barrel of the lens, focus/zoom becomes difficult, and don't even try autofocus.
This centers the wobbly front group for these tripod shots.

This problem with the disappointing corner performance is widely known, the dpreview test of this lens mentioned it also.
And even on my 6Mpixel *istD it shows on portret mode shots.
For landscape shots the decentering of the downward pulled front group is less harmful.

I've now retested with 16mp and a much older and more used/worn 16-45, and you're absolutely correct. Whether it's the higher resolution or 5+ years of wear/tear or a combination, the wobble in my 16-45 now has a significant/visible effect on performance.

Paul