UPDATE! Help Micro Adjusting the new Canon 24-70 II

philthejuggler

philthejuggler

Messages
7,167
Name
Phil
Edit My Images
Yes
I have recently purchased the 24-70II which performs wonderfully on my 5DII

On my 1DsIII however, it needs adjusting.

Trouble is the adjustment required for 24mm is about 4 and for 70mm it is 10. As you can only enter a single value, I chose 9 (closer to the long end as recommended) but focus is clearly off at the wide end when I do this.

Any suggestions?

Thanks

Phil
 
Last edited:
Phil,

I think you've gone the wrong way with your compromise unless the majority of your usage is likely to be at the long end.

The MA units are 1/8 of the single sided depth of focus (ie, depth of field in the image plane). For the same aperture value then depth of focus will increase coincident with focal length increasing...the opposite to depth of field when distance is constant.

So, given your values or 4 and 10 then I think 6 is going to give you a better compromise.

Bob
 
Canon Bob said:
Phil,

I think you've gone the wrong way with your compromise unless the majority of your usage is likely to be at the long end.

The MA units are 1/8 of the single sided depth of focus (ie, depth of field in the image plane). For the same aperture value then depth of focus will increase coincident with focal length increasing...the opposite to depth of field when distance is constant.

So, given your values or 4 and 10 then I think 6 is going to give you a better compromise.

Bob
Click to expand...

Ok will try it - thanks. It has needed more MA since going to Canon for a recall - I might send it to them to check out again.

Phil
 
That's a difference for sure Phil, but not a massive one IMHO - it is micro adjustment after all. Definitely bias towards the long end.

What distance are you testing at? Is it unrealistically close and what target are you using? Reason I ask is because of the discrepancy and also you sound very sure about the exact adjustment. In my experience with wide-angles, at realistic shooting distances generous depth of field makes it very hard to be that confident, even at f/2.8. And for the same reason, MA is less critical.

Edit: crossed post with above. While AF works by assessing depth of focus (because it can't do anything else, being behind the lens) it's depth of field that matters and any errors will be far more noticeable at the long end.
 
Last edited:
HoppyUK said:
That's a difference for sure Phil, but not a massive one IMHO - it is micro adjustment after all. Definitely bias towards the long end.

What distance are you testing at? Is it unrealistically close and what target are you using? Reason I ask is because of the discrepancy and also you sound very sure about the exact adjustment. In my experience with wide-angles, at realistic shooting distances generous depth of field makes it very hard to be that confident, even at f/2.8. And for the same reason, MA is less critical.

Edit: crossed post with above. While AF works by assessing depth of focus (because it can't do anything else, being behind the lens) it's depth of field that matters and any errors will be far more noticeable at the long end.
Click to expand...

Thanks for the responses.

I just did some tests on a gravestone - camera on tripod, AF and then move to live view and try to improve sharpness with MF. Gravestone was about 5m from camera.

70mm seemed perfect - no adjustment made significant difference. At 24mm end significant improvement by adjusting MF at 10x in live view.

Original values from Reikan Focal.

Phil
 
philthejuggler said:
Thanks for the responses.

I just did some tests on a gravestone - camera on tripod, AF and then move to live view and try to improve sharpness with MF. Gravestone was about 5m from camera.

70mm seemed perfect - no adjustment made significant difference. At 24mm end significant improvement by adjusting MF at 10x in live view.

Original values from Reikan Focal.

Phil
Click to expand...

At 5m, 24mm f/2.8, your DoF is a whopping 16m (2m in front, 14m behind). MA is capable of way way finer adjustments than that, but there's room for a big error within that range. If you're seeing 'significant improvement' in live view, I'm inclined to think there's summat wrong. With a subject like a gravestone, you should be able to see where the actual focus point is on the grass. Any clues there?

Then you say everything's fine on your 5DII, but not on the 1DsIII...:thinking: Camera must be suspect too. You'll need a few more very careful tests before drawing conclusions.
 
Yes I agree - I was surprised by the result - perhaps at that setting the focus point was catching something nearer to the camera.

Phil
 
HoppyUK said:
At 5m, 24mm f/2.8, your DoF is a whopping 16m (2m in front, 14m behind).
Click to expand...

Remember that DoF calculations are for a normal-sized print viewed at a normal distance. When you start zomming in to the individual pixels you drastically reduce the size of the circle of confusion, which reduces the DoF. So zooming in 10x on a 1Ds will give you a DoF of around 70cm.
 
hollis_f said:
Remember that DoF calculations are for a normal-sized print viewed at a normal distance. When you start zomming in to the individual pixels you drastically reduce the size of the circle of confusion, which reduces the DoF. So zooming in 10x on a 1Ds will give you a DoF of around 70cm.
Click to expand...

Yes, quite true - that was the point I was trying to make previously with the 'way way' comment. Hard to put exact numbers on it though your estimate instinctively seems somewhere right.

But under those conditions, there will still be quite a broad zone where sharpness appears to be effectively spot on, and let's say it is around 70cm. That camera/lens should be able to nail focus comfortably within that zone with ease, yet the OP says he can see significant error.

That's what's making me think there is summat amiss, and it may not just be with the lens, especially when apparently it's fine with the 5D2. Needs more careful testing and process of elimination, but unless that comes up with some clear answers, it may be better to do as Daugirdas suggests and send both the lens and 1Ds3 for calibration. You can chase these things around and go half mad/blind.
 
Hmm - reran the test today on a larger target without issue - perhaps yesterday something in the foreground was just clipping the AF area.
 
philthejuggler said:
Hmm - reran the test today on a larger target without issue - perhaps yesterday something in the foreground was just clipping the AF area.
Click to expand...

Haha, you sounded so certain and precise... :D
 
HoppyUK said:
Haha, you sounded so certain and precise... :D
Click to expand...

Well I was very careful to put the centre point dead on the centre of the gravestone, but I can't think of any other explanation.
 
philthejuggler said:
Well I was very careful to put the centre point dead on the centre of the gravestone, but I can't think of any other explanation.
Click to expand...

Fair comment. Just goes to show how tricky MFA can be, prone to different test methods, and (though I'm not saying this applies here) very commonly subject to user error.

Could also be just an inconsistency thing. You do get the odd miss-focus, that usually passes unnoticed, except in tests like this. But if there is an error, it will always be there and will show up every time in a sequence.
 
This is the shot @ 24mm - you can clearly see where the focus point sits, but the gravestone to the left and closer to the camera is much sharper:

p1735730237.jpg
 
I took quite a few shots with the lens over the weekend - mostly at the longer end, but had reasonably consistent focus accuracy. There was a series of three similar shots where only one was in focus, but I think that was because the focus point was on a low-contrast area.
 
I used the lens for a model shoot last night - I think the combination of the 1DsIII and this lens is just susceptible to some random focussing errors. A percentage of shots (not worked out how many yet) were simply out of focus and I don't think it was user error.

Phil
 
I had extremely similar issues with a 5DII and 24-70 mkI. It was so infuriating that I sent both back for a refund within a week of buying them. With my combo, live view was always perfect, but AF through the viewfinder was variable, to the point where AF adjustments were pretty much useless. The body worked fine with other lenses and the lens worked fine with other bodies, but the pair took a disliking to each other.

Have you tried both bits of kit on other equipment?

Also check you have the latest body firmware. Canon sometimes updates the firmware to work better with newer lenses.
 
Last edited:
Orville said:
I had extremely similar issues with a 5DII and 24-70 mkI. It was so infuriating that I sent both back for a refund within a week of buying them. With my combo, live view was always perfect, but AF through the viewfinder was variable, to the point where AF adjustments were pretty much useless. The body worked fine with other lenses and the lens worked fine with other bodies, but the pair took a disliking to each other.

Have you tried both bits of kit on other equipment?

Also check you have the latest body firmware.
Click to expand...

Interesting you had this issue. The lens seems spot-on with my 5DII although not taken as many shots with that combo yet. The error rate is not huge - maybe one in ten or fifteen. I also can't be 100% that there is no user error looking at the mis-focussed shots, but I am confident that some are not user error.
 
philthejuggler said:
Interesting you had this issue. The lens seems spot-on with my 5DII although not taken as many shots with that combo yet. The error rate is not huge - maybe one in ten or fifteen. I also can't be 100% that there is no user error looking at the mis-focussed shots, but I am confident that some are not user error.
Click to expand...

Just a thought... I'm not aware of any compatibility issues with that combo, but maybe worth an email to Canon?
 
HoppyUK said:
Just a thought... I'm not aware of any compatibility issues with that combo, but maybe worth an email to Canon?
Click to expand...


Good thinking - thanks for that. Might call CPS support tomorrow!
 
philthejuggler said:
Good thinking - thanks for that. Might call CPS support tomorrow!
Click to expand...

Latest firmware? Might be summat like that.

Inconsistent errors are a nightmare and almost impossible to pin down.
 
HoppyUK said:
Latest firmware? Might be summat like that.

Inconsistent errors are a nightmare and almost impossible to pin down.
Click to expand...


Yeah - the body went back to Canon a couple of weeks ago for a free mirror fix that I didn't know anything about until it started locking up and I googled it!
 
philthejuggler said:
Yeah - the body went back to Canon a couple of weeks ago for a free mirror fix that I didn't know anything about until it started locking up and I googled it!
Click to expand...

If the mirror, and especially the piggy-back AF mirror are not mechanically 100%, that would throw up inconsistencies for sure. Did it do that before the mirror fix?
 
Ok - I've now had chance to analyse the shots from the shoot in more detail and I have correctly focussed shots at all focal lengths. I've also checked with Canon and checked the firmware

The ones where there were variations in focussing were in lower light (shaded) areas and it would appear that it's locked on the subject's shoulder - not her face. In better light the hit rate was 100%. I also checked the same lens under studio lighting conditions and there was no mis-focussing.

CPS said they did not adjust the mirror at all and no known issues with the 24-70II and the 1DsIII.

Conclusion - User error - :(
 
LOL Are you using auto AF-point select? Centre point only for checking accuracy.

Even when the AF point is displayed, all it shows is which point was used, not necessarily what part of the subject it was over when focus was locked. There's often a bit of focus-recompose going on, even when it's not deliberate.
 
You must log in or register to reply here.
Back
Top