Microadjustment (fluke?)

[msdarkroom]

I just spent a lot of time doing AF MA on the 5dm3.
When I was finished I wrote down the adjustments for future reference.

Here they are:
24-70 wide: -5 tele: +5
70-200 wide: +5 tele: 0
50: -3

When I wrote down the numbers I noticed that the adjustments seem to correspond to the focal length, regardless of lens.
Wide is negative 5 and increases to positive 5 by the time it hits 70 for all lenses. By the time I reach 200, I am at zero MA.

Is this just a fluke or are most people seeing this?

 

[Loswr]

I'd say fluke. Lenses vary, bodies vary, and the interaction varies. I've done 10 lenses on each of two bodies (7D and 5DII), with zooms at wide, tele, and 1-3 intermediate focal lengths. No real pattern like that.

 

[LanceF]

How are you testing these lenses?

 

[msdarkroom]

How are you testing these lenses?

I'm shooting 5 shots wide open and then 5 at various apertures depending on the lens until I get to f/8.
I'm using a tripod, SpyderCal, remote shutter, and mirror lockup.
I'm checking the images at 100% in Lr and then going back to the camera and adjusting as needed until I get the focus point plane to be the sharpest point of the image.

When I am satisfied with the tripod tests I then handhold to verify that I am getting the same results.

 

[te4o]

May I squeeze a question in here:
Can I MA manual focus lenses? Dones this make sense - I thought the MA corrects the lens USM where to stop...

 

[Mt Spokane Photography]

May I squeeze a question in here:
Can I MA manual focus lenses? Dones this make sense - I thought the MA corrects the lens USM where to stop...

If you have a chipped manual focus lens that causes the AF light to flash, doing AFMA will adjust the body so the light flashes when in exact focus.

 

[Loswr]

Yes, you can AFMA manual lenses. It's true that AFMA alters the movement of the USM motor, but it also alters the behavior of focus confirmation for MF lenses. That includes Canon's MF lenses, like the TS-E series and MP-E 65mm (although I really prefer LiveView for those applications). It also applies to the focus confirmation of AF lenses when they are set to MF via the switch on the lens.

 

[skitron]

May I squeeze a question in here:
Can I MA manual focus lenses? Dones this make sense - I thought the MA corrects the lens USM where to stop...

As the others have said, YES if it has an auto confirm chip on it.

I have the Rokinon 35mm f 1.4 manual lens and added an autoconfirm chip and it did indeed need AFMA on both 5D2 and 50D to get the focus light/beep to trigger on the deisred focus plane. And even though the Rokinon 35 doesn't have this problem, a manual lens with focus shift will be as much a problem using a chip as a full auto would be.

Once I did this, the lens becomes in effect "semi-manual" or "semi-auto" (take your pick) since it can be shot in AP - you just manually adjust f-stop on the lens and then set the body to the same f value, and auto exposure now works. Then just press the shutter half way and twist focus until you get the usual red light and beep and after AFMA, very sharp focus is achieved.

Full manual would be a real handful without the "semi-auto" features...a definite must-have especially when you can get the programmable version af confirm chips for under $20 on eBay. Basically your hands become the motors for both focus and aperture and the camera body does the same things as usual with the chip installed on lens.

 

[drjlo]

How are you testing these lenses?

I'm shooting 5 shots wide open and then 5 at various apertures depending on the lens until I get to f/8.
I'm using a tripod, SpyderCal, remote shutter, and mirror lockup.
I'm checking the images at 100% in Lr and then going back to the camera and adjusting as needed until I get the focus point plane to be the sharpest point of the image.

Try this moire interference microadjust method. I really like how fast and reproducibly it works. Even +/- 1 adjustment clearly shows the difference instantly on fast lens like 85L.

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

 

[msdarkroom]

How are you testing these lenses?

I'm shooting 5 shots wide open and then 5 at various apertures depending on the lens until I get to f/8.
I'm using a tripod, SpyderCal, remote shutter, and mirror lockup.
I'm checking the images at 100% in Lr and then going back to the camera and adjusting as needed until I get the focus point plane to be the sharpest point of the image.

Try this moire interference microadjust method. I really like how fast and reproducibly it works. Even +/- 1 adjustment clearly shows the difference instantly on fast lens like 85L.

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

Thanks. I stumbled on that page a few days ago and thought it was a cluttered mess. I didn't give that method a try. I'll revisit it.
Thanks for the link.

-MS

 

[Loswr]

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

Thanks. I stumbled on that page a few days ago and thought it was a cluttered mess. I didn't give that method a try. I'll revisit it.

FWIW, I found that method to be inaccurate. I can move the camera forward and backward for a reasonable distance without seeing any change in the moiré patterns, even with fast lenses. OTOH, on my LensAlign Pro, I can pick up much finer differences. The key thing when doing manual AFMA testing is to take many shots at each setting (personally, I take 8 shots per adjustment value). That's why I really like Reikan FoCal - it automates the process. Unfortunately, it doesn't support the 5DIII yet...but it should fairly soon. I'll be in that boat when my 1D X arrives, so likely I'll manually calibrate my most commonly used lenses, and just wait for FoCal support to do the rest of them.

 

[msdarkroom]

Thanks Neuro, I'll skip that method. My SpyderCal seemed to have worked fine with 5 shots, but it is definitely hard to tell the difference between a single digit MA.

I recently emailed the FoCal guys and I'll have to wait until later on (they hope for Q2) to use their stuff as I'm a Mac guy.

 

[Loswr]

I recently emailed the FoCal guys and I'll have to wait until later on (they hope for Q2) to use their stuff as I'm a Mac guy.

Yeah, so am I. I run FoCal on a Parallels virtual machine running on my 17" MBP, and it works just fine.