24-70mm f2.8 - Tamron on my Camera, Canon in the closet

[bobkeenan]

I posted an earlier posting about putting the Canon in the closet.....maybe I did that too early?



I really like the VC capability of my Tamron 24-70. But my percentage of shots that are in focus dropped a bit. So I did a MA cal, using Reikan's FoCal on the lens with my 5DMkIII and got some results that seemed odd to me.
So I went and did a bunch of calibration with the Focal and Lens Align systems with both my Canon and Tamron 24-70's on both my 5DMkII and III.
Here is the long story of my results:
http://www.bobkeenanphoto.com/.../autofocus-af-microadjustment-ma-epiphany/
The short story is that the Tamron MA shifts a fair amount compared to the Canon.
Any thoughts on this? Did I get a bum Tamron?

[bobkeenan]

I went out and bought another Tamron 24-70 f2.8. If its bad I will return it (local shop). If good I will send the original lens back to Tamron.

Well seems like the Tamron that I just got is all good. MA does change a bit with distance to target but only by 1 or 2.

I kept putting the old one on and then the new one. I used my Reikan Focal System and then the lens align and followed that all up with pictures of real stuff.

Here is what I found beyond my blog post here:

Bobkeenan Photography Autofocus AF, Microadjustment MA Epiphany - Bobkeenan Photography

In fact I will have to update it.

The first tamron was calibrated with Reikan FoCal system. And at least for the beta version 1.5 it was giving me inflated MA adjustment. I verified this with the Lens Align system and real shots.

When I went to the lens align system I got good MA adjustment but the required MA changed too much as you moved from 8' to 2' to the target on the Canon 5DMkII and the movement was about double on the Canon 5DMkIII.

All these variables are making me dizzy.

Bottom line with the original Tamron lens adjusted for sharp images at 8' it would not take sharp images with AF at 2,3,4,5 ft. And vice versa.

The new lens changes but over the distances that I was checking from 2-8' It only varied by 1 MA and for real shots I might not notices that.

Not sure what's going on with the Focal system. That was working good on the other lenses. I may go back and try their 1.4 version.

Moral of the story. If the images are not sharp ..... return the lens. The new lens is behaving well and of course I still love the VC.

I will send the old one back to Tamron for repair and then selling it on EBAY..... look for a good deal in a few weeks and I will only sell it if its fixed. Although it does work really well on the Canon 5DMkII now. Isn't that weird??

[Duce]

Like you I recheck focus on my lens camera combination and it seems to be more of a problem with zoom lens than prime lens. My macro lens is a Tamron 60mm and seems to be very stable even with travel
Duce

[stef@apl]

Bob, I notice you often say things like "Nice fit" regarding the curves. I think the curve fit is only important near one area: the maximum. If you get a shotgun effect at a bad adjustment, that's not a problem... in fact, it's probably a good thing. If you get a shotgun effect at the peak (which is presumably the best micro AF adjustment), that would be a very bad thing. I realize the AF adjustment program is probably trying to use statistics to find the most likely AF area, but I wonder if there's some wiggle room, which focuses very well even when a few mm off. If so, it would allow you to find a better spot.

+13 on the 5d3 is kind of crazy. But if you look at the actual points and ignore the fitted curve, +16 might be a better number to use which will focus short/med/far.

Also, did you do the tests on a tripod with the AF turned off? I don't know if tamron's AF senses tripods...I would guess it does, but I'd be interested to know if that changes anything.

[scoundrel1728]

I would understand the mechanical feedback affecting the vibration compensation when the lens is mounted on a tripod, but how would tripod mounting affect the AF (autofocus)?

[stef@apl]

Originally Posted by scoundrel1728

I would understand the mechanical feedback affecting the vibration compensation when the lens is mounted on a tripod, but how would tripod mounting affect the AF (autofocus)?

If it's not tripod sensing, it could make shots go out of focus, causing them to look as if they had motion blur.

[swampler]

Originally Posted by stef@apl

If it's not tripod sensing, it could make shots go out of focus, causing them to look as if they had motion blur.

That would be VC doing that, not AF. You would turn off VC.

[stef@apl]

Originally Posted by swampler

That would be VC doing that, not AF. You would turn off VC.

Is there a point I'm missing? If it's not in focus due to not sensing the tripod, it's still out of focus and will show up in the graph on the micro-AF test.

Maybe you were asking if the tripod affects the AF circuitry? I don't think so. But oof is oof as far as most software if concerned, and something testing for it specifically won't care how the blur was caused.

[swampler]

Originally Posted by stef@apl

Is there a point I'm missing? If it's not in focus due to not sensing the tripod, it's still out of focus and will show up in the graph on the micro-AF test.

Maybe you were asking if the tripod affects the AF circuitry? I don't think so. But oof is oof as far as most software if concerned, and something testing for it specifically won't care how the blur was caused.

The blur is caused by the VC trying to still a non moving image and creating blur. Has nothing to do with AF at all. Canon even tells you on older lenses to turn off IS when using a tripod.

[scoundrel1728]

Originally Posted by swampler

The blur is caused by the VC trying to still a non moving image and creating blur. Has nothing to do with AF at all. Canon even tells you on older lenses to turn off IS when using a tripod.

Exactly. Besides, I don't think anyone would suggest running a reasonably careful focusing test using a handheld camera.