How does Mk. III handle 2 micro AF adjustments?

[Z]

So... enough time has passed for a lot of users to have calibrated their 5D Mark IIIs. I'm left wondering, out of curiosity's sake, how the new AFMA works. Let me expand on that...

With the new AF system in the 5D3/1DX, users can now provide 2 AFMA values; one for the long and one for the wide end of zoom lenses. Let's say for argument's sake my 70-200 worked best with +4 at 70mm and +12 at 200mm - does the camera scale the AFMA across the zoom range? (i.e. automatically guesses +8 at 135mm, halfway between the two extremes?).

I'm not sure how anyone would have this information, but I'd love to be enlightened.

 

[Loswr]

Thanks for posting this, I am also curious (in fact, I previously wondered about this).

I'm not sure how anyone would have this information

The AFMA value is displayed on the LCD during image review, and written in the EXIF (can be viewed in DPP). So, it would be as simple as applying to different values to the wide vs. tele ends of the zoom, taking shots through the zoom range, and looking at the AFMA that is applied.

 

[Z]

... it would be as simple as applying to different values to the wide vs. tele ends of the zoom, taking shots through the zoom range, and looking at the AFMA that is applied ...

Excellent - in that case, somebody hook us up with some values!

 

[bkorcel]

So far I have not had much luck using AFMA on the wide end as the DOF on the wide end typically does not allow for a good sense of how far focus may be off if any. Any guess would be just that. So far I have left the wide end set in the middle.

According to the documentation the camera will extrapolate AFMA between the two ranges for a zoom.

I do like the fact that it will register lenses based on serial numbers as well as extender combinations which is desireable if you have multiples of same model but dont have time to remember which lens was calibrated with which camera.

I've also found that trying to evaluate AFMA on the camera screen is not a good idea. Transfer the raw image to DPP or photoshop and evaluate there. Takes more time but it is much easier to evaluate without any guessing.

So I discovered that my 300 2.8LIS by itself and with 1.4x was front focusing by about 4. With the 2X it was dead on.

70-200 2,8L IS II on the 200 side was front focusing about 2.

That's probobly within technical tolerances I would think. Someone might be able to shed more light on that.

 

[Mt Spokane Photography]

Looking at my AFMA test shots, the AFMA value is interpolated between the two extremes.

Thus, for my 28-300mmL, at

28mm, AFMA = -13
135mm, AFMA = -8
300mm, AFMA = 0

 

[Loswr]

According to the documentation the camera will extrapolate AFMA between the two ranges for a zoom.

What it states is: "When shooting with the intermediate range (focal length) of a zoom lens, the AF's point of focus is corrected relative to the adjustments made for the wide-angle telephoto ends."

'Corrected relative to the adjustments' is a bit ambiguous. For example, if you enter -5 at the wide end and +7 at the long end, will it run a gradient of adjustments across the zoom range (progressing from -5 to +7 in small steps as you zoom)? Or will it abruptly switch from -5 to +7 adjustment at the midpoint of the zoom range (or some other arbitrary point)?

That's the question on the table. To answer it doesn't even require that the adjustments be the right ones to give correct focus - just picking random (different) adjustments for each end and seeing what the middle looks like would do the trick.

 

[Loswr]

Looking at my AFMA test shots, the AFMA value is interpolated between the two extremes.

Thus, for my 28-300mmL, at

28mm, AFMA = -13
135mm, AFMA = -8
300mm, AFMA = 0

Thanks! How finely? For example, what do 85mm and 200mm look like on the 28-300mm?

 

[Loswr]

Also worth noting, below are the FoCal results for my 28-300mm on my 5DII. So, if I adjusted only at the two ends, the interpolation would be useless...

 

[Mt Spokane Photography]

Looking at my AFMA test shots, the AFMA value is interpolated between the two extremes.

Thus, for my 28-300mmL, at

28mm, AFMA = -13
135mm, AFMA = -8
300mm, AFMA = 0

Thanks! How finely? For example, what do 85mm and 200mm look like on the 28-300mm?

I sold the lens last week, so I only have the images taken at those three points after calibrating it and my 5D MK III with Focal. They indeed do not look like a linear interpolation from two points is going to be very good on a 10:1 zoom.

I bought the lens locally from Craigslist, but decided it wasn't my style of lens. I use two zooms for outdoor walkaround, 24-105mm f/4 L, 70-200mm f/4L IS, and 100-400mm L. Indoors its all primes. My 70-200mm f/4 gets seldom used, so it may be sold too.

 

[Loswr]

I bought the lens locally from Craigslist, but decided it wasn't my style of lens. I use two zooms for outdoor walkaround, 24-105mm f/4 L, 70-200mm f/4L IS, and 100-400mm L.

As did I, and for me the jury is still out. It worked well as a travel lens, combined with the 35L for indoors/night. Not sure how much use it'll get outside of that - am hoping it will be useful for local outings with my daughters, where previously I'd take the 24-105mm and 70-200mm f/2.8L IS II. But I'm thinking I may ultimately prefer the combo of the 24-105 and the 70-300L.

As with all my CL purchases, I bought cheaply enough that I can re-sell and most likely make a profit.

 

[markko]

Thanks! How finely? For example, what do 85mm and 200mm look like on the 28-300mm?

I don't have a 28-300m, but I did some tests with my 70-200mm II and 100-400mm:

 

[Loswr]

I don't have a 28-300m, but I did some tests with my 70-200mm II and 100-400mm:

Wonderful, thanks!

 

[Z]

Thank you very much markko, I think that answers our question!

 

[markko]

Thank you very much markko, I think that answers our question!

Welcome

One question remains unanswered, though: how do you pick the right value for adjustment!? :-\

I have the Lens Align tool, but still: I just used it by night in the living room and the difference in the values is so little... I think I have to use it at daylight, so that I'm really really really sure that I see the red dot through the center hole (and the lens is aligned properly with the tool).

The odd thing happened that the tool suggested an adjustment that I would expect the opposite when I was looking at real world pictures.

 

[Loswr]

One question remains unanswered, though: how do you pick the right value for adjustment!?

No simple answer there. In selecting a single AFMA (for my 5DII and 7D), I consider both the results at the different distances in terms of DoF for a given focal length, and also the distances which I commonly shoot for those lenses. I test at two distances - 25x and 50x the focal length, and for zoom lenses I test at both ends, and multiple intermediate points. For example, the value I chose for the 35L is biased toward the closer distance, since that's where DoF will be thinner; I picked the value for the 85L based on the fact that I usually shoot that lens at about 7-8 feet for portraits, which is ~25x the focal length. For lenses like the 70-200 II, which I shoot at varying focal lengths and distances, I selected values biased toward the long end of the zoom range, since that's where DoF will generally be thinnest and focus most critical. Selecting adjustments for the 1D X will be similar - weighting the choices for both ends based on what the values in the middle are, where I shoot the lens most, and where DoF will be thinnest.

 

[TrumpetPower!]

One question remains unanswered, though: how do you pick the right value for adjustment!? :-\

There's only one reasonable answer: Reikan FoCal. It (automatically) makes multiple exposures at multiple AFMA settings and does a statistical analysis to determine the setting most likely to result in the highest percentage of in-focus shots.

As you've noticed, the system's margin of error is bigger than the smallest AFMA adjustment. Thus, if you take a single shot at 0, click the step adjustment in tethered live view three times to get it in focus, and conclude the "right" value is therefore +3...well, it could very easily be that the single shot was a fluke and you'll instead get better results with a setting of -1.

Unless you perform the type of statistical analysis that FoCal does, you'll never actually know for sure.

Cheers,

b&

 

[Loswr]

One question remains unanswered, though: how do you pick the right value for adjustment!? :-\

There's only one reasonable answer: Reikan FoCal. It (automatically) makes multiple exposures at multiple AFMA settings and does a statistical analysis to determine the setting most likely to result in the highest percentage of in-focus shots.

Yes...for the subject distance you tested at, and for a zoom lens, the focal length you tested. I can do the same thing FoCal does, manually with the LensAlign - it just takes longer. The fact that my manual values closely match those from FoCal confirms that. But if you change the subject distance and/or focal length, the determined value usually changes. Unless you always shoot at a particular distance and, for a zoom lens, a specific focal length (or two, if you have a 5DIII or 1D X), selecting the AMFA (or two of them) to assign to a lens (one that's applicable to the range of distances and focal lengths) requires an informed compromise. There's more to it than 'f/ocal and forget it'.

 

[TrumpetPower!]

There's more to it than 'f/ocal and forget it'.

Well, of course the system isn't perfect. And, again of course, you should put some thought into how you use your tools. If you always shoot one particular way, optimize your system for that; otherwise, the 20x focal length is a good all-around compromise. And you need to know and understand the limits and how to work with and around them.

FoCal is the best tool I'm aware of for determining the optimum AFMA setting, but it's up to the photographer to determine the optimum environment in which to use that tool.

b&