From Good to Great!

[Badger]

Ever see those posts about people complaining that they "have a bad copy" of a lens? Or have you bought a lens that everyone raves about but you are not quite as excited about your "copy"?

AFMA! I'm a believer! Adjust your body to your lens. You might be pleasantly surprised I knew that adjustment was available on my camera but adjusting it seemed more complicated than my mind could wrap around. A few days ago, I took the plunge and purchased Reikan's Focal and was shocked to find out my nifty fifty, which I am learning to love, was off by +16 on my 6D! All my lenses needed some adjustment (the 50mm needed the most).

You Pros already know this, but if you are anything like me, go through the AFMA exercise. I think Focal does an amazing job, it is fully automated, easy to use, but certainly not free. The good news is that there are several free options out there that are probably just as effective but maybe not as easy. Pick one, and do it. You might find a lens in your collection that you were happy with but are now happier with

 

[Deleted member 91053]

From great to greater?
I checked out all my lenses (except a manual focus Contax-Ziess) on my 1D4 and focus accuracy was spot on - perhaps I am just lucky?
However (if it EVER stops raining here in South Wales) I am going to re-check them now I have moved to the 1DX. So far I have only had a chance (in over 2 months!) to use my 300 & 800 Canons on it and it appears to be even better than the Mk4 - I hope it's the same for the others!
If not I shall be checking out your recommendation..........

 

[Ivan Muller]

Badger, thanks for the info...I didn't even know they existed, but I shall keep them in mind!

 

[neuroanatomist]

A 'quick and dirty' test to see if you'd benefit from AFMA is to set up on a tripod, timer or remote release, lens with open, tape a newspaper page or sheet if printed text to a wall with plenty of light, and take ~5 shots with Live View AF (not Quick) or manual focusing using live view at 10x, then take ~5 shots with phase detect (viewfinder) AF with mirror lockup. Pick the sharpest shot of each set, and if the Live View shot is noticeably sharper than the viewfinder AF shot, AFMA will likely help.

 

[sagittariansrock]

A 'quick and dirty' test to see if you'd benefit from AFMA is to set up on a tripod, timer or remote release, lens with open, tape a newspaper page or sheet if printed text to a wall with plenty of light, and take ~5 shots with Live View AF (not Quick) or manual focusing using live view at 10x, then take ~5 shots with phase detect (viewfinder) AF with mirror lockup. Pick the sharpest shot of each set, and if the Live View shot is noticeably sharper than the viewfinder AF shot, AFMA will likely help.

Nice one! Also a good test to confirm your AFMA has actually worked.

 

[cayenne]

From great to greater?
I checked out all my lenses (except a manual focus Contax-Ziess) on my 1D4 and focus accuracy was spot on - perhaps I am just lucky?
However (if it EVER stops raining here in South Wales) I am going to re-check them now I have moved to the 1DX. So far I have only had a chance (in over 2 months!) to use my 300 & 800 Canons on it and it appears to be even better than the Mk4 - I hope it's the same for the others!
If not I shall be checking out your recommendation..........

I was looking at the Focal stuff awhile back, but it seems that I read on the site, that it didn't fully work with the 5D3.

Does anyone know if they've fixed it yet?

Cayenne

 

[rpiotr01]

I was looking at the Focal stuff awhile back, but it seems that I read on the site, that it didn't fully work with the 5D3.

Does anyone know if they've fixed it yet?

Cayenne

It works just fine, but one small piece isn't fully automated. When it's running the AFMA test, it can't automatically change the AFMA settings in camera for you. So it'll do it's thing, take a bunch of shots, then a prompt will come up on screen that says "Set AFMA to -20/-10/0 etc. and click OK" - you do that, and it continues to run the test automatically. Really not a big deal, takes a couple seconds each time it asks.

 

[cayenne]

I was looking at the Focal stuff awhile back, but it seems that I read on the site, that it didn't fully work with the 5D3.

Does anyone know if they've fixed it yet?

Cayenne

It works just fine, but one small piece isn't fully automated. When it's running the AFMA test, it can't automatically change the AFMA settings in camera for you. So it'll do it's thing, take a bunch of shots, then a prompt will come up on screen that says "Set AFMA to -20/-10/0 etc. and click OK" - you do that, and it continues to run the test automatically. Really not a big deal, takes a couple seconds each time it asks.

Ok, thanks for the info. I wonder why it can't do the full auto on the 5D3 like it can other cameras?

Hmm..trying to get to the Reikan site...seems they're having some webserver problems.

I'll try to look closer at it again soon.

I'm planning this year, to buy a copy of the Canon 50L f/1.2, and I want to make sure I get it properly tuned for my camera.

C

 

[neuroanatomist]

Ok, thanks for the info. I wonder why it can't do the full auto on the 5D3 like it can other cameras?

It's not really Reikan, it's a limitation in the Canon SDK for the 5DIII (and 1D X) - the AFMA value cannot be changed by the tethered computer. Those were the first two cameras with that store 2 values per zoom lens (W and T). The SDKs for newer cameras with that capability (6D and 70D) do support changing the AFMA from the computer, so FoCal is able to work in fully automatic mode on those cameras.

Personally, I don't even use the Manual Settings Change mode, rahter I just shoot all my shots without the computer connected, then load them into FoCal's Manual Analysis mode. I take more shots than I need - 83 shots per test at each focal length+distance combo, two distances per focal length, so that's two tests for a prime and 4-10 tests for a zoom (depending on the range). But I get great curve fits, and it only takes me about 10 mintues for each 83-shot test.

 

[ishdakuteb]

A 'quick and dirty' test to see if you'd benefit from AFMA is to set up on a tripod, timer or remote release, lens with open, tape a newspaper page or sheet if printed text to a wall with plenty of light, and take ~5 shots with Live View AF (not Quick) or manual focusing using live view at 10x, then take ~5 shots with phase detect (viewfinder) AF with mirror lockup. Pick the sharpest shot of each set, and if the Live View shot is noticeably sharper than the viewfinder AF shot, AFMA will likely help.

might i have a question about your suggestion on using live af focus instead of quick focus. per my understanding, i do think that we have to use live af for liveview shot, quick af does not work in liveview mode. but live af type is somewhat not a phrase detection type like you have mention above "...take ~5 shots with phase detect (viewfinder) AF with mirror lockup..." along with your previously suggestion to modify it to live af type. i am kind of confused.

in short, which one should i use as if I am using viewfinder only for most of my shoot? live af (contrast dection) instead of quick af (phrase dection)? i am currently using quick af type...

thanks...

 

[sagittariansrock]

Ok, thanks for the info. I wonder why it can't do the full auto on the 5D3 like it can other cameras?

It's not really Reikan, it's a limitation in the Canon SDK for the 5DIII (and 1D X) - the AFMA value cannot be changed by the tethered computer. Those were the first two cameras with that store 2 values per zoom lens (W and T). The SDKs for newer cameras with that capability (6D and 70D) do support changing the AFMA from the computer, so FoCal is able to work in fully automatic mode on those cameras.

Personally, I don't even use the Manual Settings Change mode, rahter I just shoot all my shots without the computer connected, then load them into FoCal's Manual Analysis mode. I take more shots than I need - 83 shots per test at each focal length+distance combo, two distances per focal length, so that's two tests for a prime and 4-10 tests for a zoom (depending on the range). But I get great curve fits, and it only takes me about 10 mintues for each 83-shot test.

Can you explain which 83 shots?
This is probably a better way than the semi-automatic mode. I have to try this.

 

[Skulker]

A 'quick and dirty' test to see if you'd benefit from AFMA is to set up on a tripod, timer or remote release, lens with open, tape a newspaper page or sheet if printed text to a wall with plenty of light, and take ~5 shots with Live View AF (not Quick) or manual focusing using live view at 10x, then take ~5 shots with phase detect (viewfinder) AF with mirror lockup. Pick the sharpest shot of each set, and if the Live View shot is noticeably sharper than the viewfinder AF shot, AFMA will likely help.

might i have a question about your suggestion on using live af focus instead of quick focus. per my understanding, i do think that we have to use live af for liveview shot, quick af does not work in liveview mode. but live af type is somewhat not a phrase detection type like you have mention above "...take ~5 shots with phase detect (viewfinder) AF with mirror lockup..." along with your previously suggestion to modify it to live af type. i am kind of confused.

in short, which one should i use as if I am using viewfinder only for most of my shoot? live af (contrast dection) instead of quick af (phrase dection)? i am currently using quick af type...

thanks...

I suspect that it is intentional.

Take some shots with normal AF. Take some with live view. Then compare the two.

As the AFMA only affects normal AF, and using live view should be accurate as it works of the image on the sensor.

It seems to make sense to me.

 

[neuroanatomist]

might i have a question about your suggestion on using live af focus instead of quick focus. per my understanding, i do think that we have to use live af for liveview shot, quick af does not work in liveview mode. but live af type is somewhat not a phrase detection type like you have mention above "...take ~5 shots with phase detect (viewfinder) AF with mirror lockup..." along with your previously suggestion to modify it to live af type. i am kind of confused.

in short, which one should i use as if I am using viewfinder only for most of my shoot? live af (contrast dection) instead of quick af (phrase dection)? i am currently using quick af type...

First off, this isn't for everyday use, but is a quick test to see if AFMA could help - there are other reasons your lens might not be delivering the sharpness you expect (decentered or misaligned element, etc.) for which AFMA wouldn't help. AFMA only affects phase detect AF with the mirror down (i.e., AF when looking through the viewfinder; note that I'm referring to the dedicated AF sensor only, not the on-CMOS phase detect in more recent cameras, which operates with the mirror up anyway). The idea is you use a focusing method other than phase detect, to see how sharp the lens could be when optimally focused, and compare that to phase detect AF (as good as possible, thus mirror lockup, in Live View it's already up), and if there's a difference, AFMA can correct for it. In each case, you take several shots (because nothing is perfect) and pick the sharpest for the comparison.

For that 'other' focusing method as a benchmark for sharpness, there are several options. The best is not really focusing at all - focus once (any method) then move the whole camera forward/backward to focus bracket, using a camera rail/slider - that's a pain (so I didn't mention it before), but it's the way lens sharpness testing should really be done. The other options use Live View, either manual focus with 10x magnification, or autofocus in Live View. For the latter, there are three modes (you pick one in settings): Quick AF, Live AF, and Face AF. Quick AF flips the mirror down and uses the dedicated phase detect AF sensor - obviously, that's not going to work as a benchmark for that same sensor. Live AF (aka contrast detect AF) uses the CMOS image sensor itself, and that's the one you want. Contrast detect AF can be imprecise in older cameras, thus taking several shots and picking the best is quite important.

You compare the best shot with phase detect (viewfinder) AF to the best shot with the other method, and if the phase AF is softer, try AFMA. If they're the same, and you still feel the lens isn't as sharp as it should be, it might need to go in for service.

Hope that clarifies...

 

[neuroanatomist]

Can you explain which 83 shots?
This is probably a better way than the semi-automatic mode. I have to try this.

Two shots at each even numbered AFMA value between |20| and |12| (inclusive), one setting the focus to infinity before AF, the other starting from the MFD, then three shots at every AFMA value between –10 and +10 (inclusive), one setting the focus to infinity before AF, another starting from the MFD, and the third without adjusting the focus ring before AF.

I had to use Manual mode before FoCal supported the 1D X, and I've just kept doing it that way. I still tether for the aperture sharpness test, multipoint test, etc.

 

[Jack Douglas]

This will sound like heresy but since the AF setting is so easy to change at least with my 6D I just change it at will if there are repetetive shots such as at my feeder and see what gives me the best result. I strongly suspect that many bird "eye" focussed shots are influenced by a forward shoulder and so I'm leaning towards a slightly + AFMA value. I'd say there is a lot of overkill in all this! Of course, to each his own.

Jack

 

[Dukinald]

Been holding off on doing afma calibration. Maybe its time to explore this now. Will try to do the manual stuff first as suggested. I see that people here mostly use reikan focal, any other alternative that might be cheaper thats also reliable? Doesn't ML have a calibration module built into it? How reliable is it?

 

[neuroanatomist]

I see that people here mostly use reikan focal, any other alternative that might be cheaper thats also reliable? Doesn't ML have a calibration module built into it? How reliable is it?

There's the LensAlign product. I've used their tool (the old Pro version that's not available anymore), and it works well (results match FoCal); they now have software, too. No idea how well it works, but IIRC it's cheaper than FoCal.

Haven't tried ML, but it's based on DotTune which I have tried and found to be unreliable (some results matched FoCal/LensAlign, others were off enough to matter).

 

[ishdakuteb]

Hope that clarifies...

thanks very much for helping to clarify... excellent explanation as always!...

 

[100]

Ever see those posts about people complaining that they "have a bad copy" of a lens? Or have you bought a lens that everyone raves about but you are not quite as excited about your "copy"?

AFMA! I'm a believer! Adjust your body to your lens. You might be pleasantly surprised

AFMA is great because it lets you get the best out of your combination of camera body and lens but it doesn’t solve the “problem” of copy variation between lenses.

It would be great if Canon came up with something like an automated in camera version of Sigma’s USB-Dock system which lets you adjust the lens at multiple (4) focusing distances and with zoom lenses at multiple (also 4) focal lengths too, so 16 values all together instead of the 2 we can do with the current AFMA http://www.lensrentals.com/blog/2013/06/sigma-optimization-pro-and-usb-dock

 

[Badger]

It would be great if Canon came up with something like an automated in camera version of Sigma’s USB-Dock system which lets you adjust the lens at multiple (4) focusing distances and with zoom lenses at multiple (also 4) focal lengths too, so 16 values all together instead of the 2 we can do with the current AFMA http://www.lensrentals.com/blog/2013/06/sigma-optimization-pro-and-usb-dock

That sounds even more amazing 100, but understand that I just discovered or started using AFMA so I'm still thrilled by the 2 settings I get with the 6D. Give me a week or two, and I'll start complaining and wishing for 4 ;D