+15 AFMA on a New 24mm f/1.4L II - Keep or Return

[conropl]

I bought a new 24mm f/1.4L II lens, and did the AFMA last weekend using FoCal. I ended up with a +15. In the past I have lived with what I got with a new lens... e.g., my macro has a -16 AFMA. My macro works great, but I have read others return lenses that are that far off. I like the lens, but I was wondering what others would do with a new lens that requires that much adjustment.

What's your thoughts? My first inclination is to keep it... but then I wonder if I am missing something.:confused:

Pat

[neuroanatomist]

Personally, I'd return it. It's fine for this body, of course. Hypothetically (with completely made up numbers), what if your current body is –8, and your next body is 0. Current lens plus new body would then need a +23 adjustment, and that's outside the range. I think I'd return any lens that needed more then |10| units of adjustment.

[HDNitehawk]

I would agree with Neuro to return it, but first a short discussion about this lens from a person that owns it.

My experience has been the wider the lens the tougher it is to get accurate results.

The 24mm is hard to nail when doing an AFMA. For one thing you have a wide DOF with this lens so a AFMA is far less critical than it is with a longer lens.
Because of the wide DOF it will be prone to random results. With the wide lenses you have to be extra critical to get good lighting, right distance and a level set up.
I would repeat the test a few times on this lens before you ship it back. Make sure it gets +15 or close to each time.

I assume you are using it with your 1D X and all of your lenses are not +15, just these two?

If you really get a +15 and you think it is right it is easy to test. Set up your 1D X with this lens at +15, set your camera on center focus and a wide aperture (at wide open or close to). Take a picture of the wife or kids eyes and head just a few feet from them and see if it hits a sharp crisp focus on their eyes. Close up that is when it is going to matter with this lens. At 10 or 15 feet your DOF has grown so large you may never notice. It is my opinion that close up is what matters with these lenses. The proof that it is really that far out is in the real life situation. If it isn't really +15 you are going to see it.

This lens is a tough one. I would check it a few times first then ship it.

[Jonathan Huyer]

Hmmm... I'd also be a bit concerned, for sure. I've never had a lens need anywhere near that much AFMA. My identical setup needed only a +3 adjustment. The whole point of having a f/1.4 lens is to use it wide open, where focus accuracy is key.

[Kayaker72]

Hi Pat,

I'd be in favor of exchanging it as well. But, in case you haven't already, I thought I'd recommend checking your set up and retesting a couple of times. I've found FoCal to be pretty senstive to a couple of items within the set up. Specifically light on target (more the better, I've seen benefit pushing the EV to > 10), making sure the target is perfectly "square"/perpendicular to your lens/parallel to image sensor, and distance to target (https://dl.dropboxusercontent.com/u/11469157/FoCal/Docs/FoCal%20Test%20Distance_1.0.pdf). I am actually trying to make improvements to my setup, specifically with the lighting.

All that said, I doubt it would make +/-15 AFMA units difference. So, if the rest of your lenses check out (so it isn't the 1DX body), I'd likely be sending the 24 f/1.4 II back.

Good luck,
Brant

[Jonathan Huyer]

Okay I just bought FoCal, and it is telling me that I need +12 for this same setup (1DX and 24 f/1.4 LII), versus +3 using LensAlign. I tried it twice at different distances and it came up with the same result. By comparison, all the other lenses I tested came in very close to the LensAlign result... this was the only one that differed by more than a few points. Unfortunately there is not enough data on the FoCal website to compare this result with others just yet. But checking the results with the 5D3, most of the tests came in between +5 to +10. The furthest outside was a +15, indicating a very rare result indeed. On my 5D3, the lens came in at +7. So there might be an issue with this lens and the 1DX body, for some reason.

[conropl]

Thanks all for the responses.


Personally, I'd return it. It's fine for this body, of course. Hypothetically (with completely made up numbers), what if your current body is –8, and your next body is 0. Current lens plus new body would then need a +23 adjustment, and that's outside the range. I think I'd return any lens that needed more then |10| units of adjustment.

I see what you mean. My macro was +10 on my 7D and -16 for the 1DX. Although I did not adjust the macro using FoCal because I wanted to adjust it for the distance that I us it at... which is almost always at macro distance (close to the MFD). So I have a home grown way that gets me very accurate at the MFD.


I would agree with Neuro to return it, but first a short discussion about this lens from a person that owns it.

My experience has been the wider the lens the tougher it is to get accurate results.

The 24mm is hard to nail when doing an AFMA. For one thing you have a wide DOF with this lens so a AFMA is far less critical than it is with a longer lens.
Because of the wide DOF it will be prone to random results. With the wide lenses you have to be extra critical to get good lighting, right distance and a level set up.
I would repeat the test a few times on this lens before you ship it back. Make sure it gets +15 or close to each time.

I assume you are using it with your 1D X and all of your lenses are not +15, just these two?

If you really get a +15 and you think it is right it is easy to test. Set up your 1D X with this lens at +15, set your camera on center focus and a wide aperture (at wide open or close to). Take a picture of the wife or kids eyes and head just a few feet from them and see if it hits a sharp crisp focus on their eyes. Close up that is when it is going to matter with this lens. At 10 or 15 feet your DOF has grown so large you may never notice. It is my opinion that close up is what matters with these lenses. The proof that it is really that far out is in the real life situation. If it isn't really +15 you are going to see it.

This lens is a tough one. I would check it a few times first then ship it.

Thanks for the insight on the lens. What you say makes sense.

I did test it several times, and I must say there were varrying results. I also did the semi auto method and adjusted it up and down to confirm it was higher than +10; but as I said, the results did vary a bit. I am pertty sure I had the set up good (4 shop halegen lights on it; mid day overcast; I was steady on a good tripod; I performed the target set up; I got it lined up good with a positive target test; and the distance was good).

My other lenses (100-400L and 24-105L) were +3T/+4W and +2T/+3W respectively. So they look good, and those settings were consistent with the settings with my 7D.

After reading your response, I did a little testing close up (couple of feet), and walked away a little less sure. I found enough apparent variation to make me want to set it back up on a tripod and check it out some more at closer distances. I may even try FoCal again at closer distances. In the past FoCal had some specific requirements for testing distance, but they now seem to be waffling on that. The FoCal set up manual talks about the Canon recomendation for testing distance; but then Focal goes on to say they thought there was some effect for distance... however, the effect was usually negligable. So perhaps if I try it at closer distances, there will be enough DOF change to get more consistent results.


Hmmm... I'd also be a bit concerned, for sure. I've never had a lens need anywhere near that much AFMA. My identical setup needed only a +3 adjustment. The whole point of having a f/1.4 lens is to use it wide open, where focus accuracy is key.


Okay I just bought FoCal, and it is telling me that I need +12 for this same setup (1DX and 24 f/1.4 LII), versus +3 using LensAlign. I tried it twice at different distances and it came up with the same result. By comparison, all the other lenses I tested came in very close to the LensAlign result... this was the only one that differed by more than a few points. Unfortunately there is not enough data on the FoCal website to compare this result with others just yet. But checking the results with the 5D3, most of the tests came in between +5 to +10. The furthest outside was a +15, indicating a very rare result indeed. On my 5D3, the lens came in at +7. So there might be an issue with this lens and the 1DX body, for some reason.

... or maybe not. I would guess the lens align set up was closer just becuase the lens is so wide on a FF. I am going to put a little more time into this lens testing to see if I can get consistent results, and may end up bypassing FoCal on this one if I can not get consistent results. Thanks for the feedback.


Hi Pat,

I'd be in favor of exchanging it as well. But, in case you haven't already, I thought I'd recommend checking your set up and retesting a couple of times. I've found FoCal to be pretty senstive to a couple of items within the set up. Specifically light on target (more the better, I've seen benefit pushing the EV to > 10), making sure the target is perfectly "square"/perpendicular to your lens/parallel to image sensor, and distance to target (https://dl.dropboxusercontent.com/u/11469157/FoCal/Docs/FoCal%20Test%20Distance_1.0.pdf). I am actually trying to make improvements to my setup, specifically with the lighting.

All that said, I doubt it would make +/-15 AFMA units difference. So, if the rest of your lenses check out (so it isn't the 1DX body), I'd likely be sending the 24 f/1.4 II back.

Good luck,
Brant

The rest of my lenses look good (including the 24mm end of my 24-105 which was shot at the same distance). I think my set up was good. I think HDNiteHawk has it right, and I have to start looking at this lens a little different.

Thank you all for the great insights.

Pat

[HDNitehawk]

Focal recommends 25x to 50x for target distance.

For some quick conversions that is 23.6" and 47.2"
The DOF for a 24mm on a 1D X at f/2.8 is listed on the DOF calculator at 47" as 16.7" Deep
Here is my theory, at 47.2" the 24mm will have a DOF is 35% of the distance.
It is a big target to pinpoint an exact point that is the optimal setting for accuracy.
Longer lenses would have a similar DOF but a smaller target by percentage.

I would get as close as you can.
Change the settings and at stops wider than F/2.8 as well.

I had trouble getting consistent results. A few of the things I did and maybe these will help if you are not doing them already. Set the camera and the target at the exact same height as the target. Set your camera as close as you can to square with the target both horizontal and vertical. Lighting on the 24mm is tough because you are so close. You get shadows if your lighting is only from one direction. It helps to get lights from both sides and no shadows, diffusion helps. Use the lens hood, you get stray light coming from the sides. Cover the view finder, this is very important, I have seen tests several points off because I forgot to cover the view finder before the test started.

I did quite a bit of testing on my 24mm in manual mode. I would get results from +4 to -2 that were very consistent and would have good IQ.
I have 32 files saved from the individual tests I ran with focal. I probably deleted that many when I was testing. I finally arrived at an average of +1 to +2 as the optimum setting to use.

As a comparison as primes go, I have 1 test I saved and only ran a few with my 500mm to confirm it is +0.
I ran just a few test and have 3 I saved for my 300mm f/2.8 to confirm it at +5
I have 24 files saved from my 35mm.
The wider it gets the harder the test to confirm.

Good luck with the test.
I still wouldn't hesitate sending it back if you are not sure.

[conropl]

Focal recommends 25x to 50x for target distance.

For some quick conversions that is 23.6" and 47.2"
The DOF for a 24mm on a 1D X at f/2.8 is listed on the DOF calculator at 47" as 16.7" Deep
Here is my theory, at 47.2" the 24mm will have a DOF is 35% of the distance.
It is a big target to pinpoint an exact point that is the optimal setting for accuracy.
Longer lenses would have a similar DOF but a smaller target by percentage.

I would get as close as you can.


I am pretty confident in my set up other than the distance. I was a distance quit a bit further than the distance you calculated. FoCal does publish Canon's recomendation of 50X focal distance, but suggested a distance closer to the intend use. So I moved back as far as I could and still get a lock on target and pass the "Target Test". In reading your logic, I think I agree with your approach more, and will retest at a closer distance to see how it goes. Yesterday, at high speed bursts and focussed at short distances, I was seeing a lot of variation (but that was a quick test and not as ridgedly mounted as I would like.... back to the tripod and test set up).

Below is the latest from FoCal on target distance (the last paragragh is what I focused on):

"The distance to the target will usually affect the AF microadjustment value by a small amount. As you approach minimum focus distance (MFD), the values can sometimes become quite different to the rest of the range.

What does this mean? Well, in reality, it means that you can either calibrate your lens to work as well as possible very close and not very good far away, or go for the opposite way round.

In one of their MA calibration documents, Canon recommends calibrating at 50x the focal length of the lens. This generally keeps well away from MFD and is not such a bad "standard" for all lenses.

FoCal won’t stop you calibrating at any distance, but if you are using the FoCal target and Target Validation, it will inform you if you are testing away from Canon’s recommended distance.

In practice, the distance you calibrate at is the distance that the selected AF microadjustment value will be correct for. So you should try to pick a distance that is around the "most used" distance for your particular shooting style."


Change the settings and at stops wider than F/2.8 as well.



Can you chose the apeture limits, or is this set by the software?

[HDNitehawk]

Expert mode with the Pro version lets you change the limits.

[neuroanatomist]

I test all my lenses at 25x and 50x the focal length (and I also tested the 100L Macro near the MFD with the small FoCal target).

Rick, you say test wider than f/2.8 - do you mean narrower? Lenses should be tested wide open, f/1.4 in this case (which is wider than f/2.8). Wide open gives the thinnest possible DoF for the test, which should mean more accuracy. Normal AF is always done with the lens wide open.

If a narrower aperture shows backfocus, that indicates the lens has focus shift (e.g., the 50/1.2L). In that case, you can actually use AFMA to compensate for the focus shift. Not really 'on the fly', but if you're shooting a series of shots at the same aperture (say, f/1.6 on the 50L) and subject distance, and you have determined the AFMA for those conditions (which would be different than the f/1.2 AFMA with a close subject).

[HDNitehawk]

Widest equals F/1.4

I thought focal had chosen f/2.8 as a default when I tested my copy.
Is this not the case John?
Maybe these were the settings I had and I started there.

From 1.4 through 2.0 I had near identical results. From 2.0 and on only slight variations.
No indication of focus shift.

[neuroanatomist]

FoCal always tests at the widest aperture, unless you've altered the settings from default.

[Jonathan Huyer]

Update: I retested my 1DX - 24 f/1.4LII combo using LensAlign, and it seems my original adjustment of +3 worked better than the FoCal recommendation of +12. So I changed it back. With the 5D3, the FoCal result was +7, with LensAlign giving me a +5. I'm guessing the differences are due to the extreme subtleties involved here, and my testing procedures which I'm sure are not up to laboratory standards. Most of the time I'll be using this lens in manual focus mode with live view anyway.

[conropl]

I set up and tested again this evening. I tested at 50X and 25X the focal length (47" and 23.5"). The results ended up better, but boarder line. The results were:

At 50X focal length test distance I tested it 4 times with the following results:
+11
+12
+9
+10
Avg = 10.5

At 25X focal length test distance I tested it 3 times with the following results:
+10
+14
+11
Avg = 11.7

The 25X was much harder to get set up, and the set up has more effect on the results at that distance (obviously). I got a positive target set up, but kept getting errors. So I went back to the target set up program, and got it set up better. That caused the readings to drop from the 17-20 range (with many errors) to what you see above (avg around +12) and it ran very quickly with no errors. With the two averages at 10.5 and 11.7, it looks like a good place to set it is at +11.

:confused::confused::confused:

Edit: Now I need to do a little shooting to verify that is where it should be.

[Jonathan Huyer]

....Now I need to do a little shooting to verify that is where it should be.

After spending the better part of a day testing my lenses with both camera bodies, I am thinking the same thing! The real world is the best test there is.