50 1.2L back-focuses with lateral focus points

[faustino]

I am experiencing a strange issue with my 50mm f1.2L lens:
when I use any of the lateral focus points (from a 5dmk3), also the cross type, the lens back-focuses. It happens also when the aperture is set to 1.2. When I choose a central focus point, the focus is perfect.

It is not an issue of the camera as I don't see the same problem with the 85mm f1.2L.

I am wandering if this issue affects only my copy or is it normal; maybe it is due to the same spherical aberration that causes the well known focus shift issue (backfocus when stopped down).

Do you have the same issue with your 50 1.2L lenses?

Thanks for your replies (and sorry for my imperfect English)!
Fausto

 

[Loswr]

The 50L has some field curvature, and that's what you're seeing. Unlike focus shift, which is not seen wide open, field curvature is most evident at max aperture.

 

[Random Orbits]

Mine focuses about as sharply using the outer points as the center point wide open on a 5DIII. How much back-focusing are you talking about?

 

[faustino]

Thanks for your replies.

The 50L has some field curvature, and that's what you're seeing. Unlike focus shift, which is not seen wide open, field curvature is most evident at max aperture.

I thought the focus had to be "precise" at the point where one is focusing, and that the field curvature would affect the remaining area of the frame.
I mean:
1. if I focus in the center, the field curvature would make closer the "focus sphere" (should be "focus plane") at the left and right parts of the frame;
2. If I focus on one side (e.g. left), the side where I focus should be spot on; while, the field curvature should make closer the focus sphere on the other side of the frame (center and even more on the right).
Isn't the above correct?

Anyhow, thank you for the explanation!
I will try to understand a bit more about the behavior of the autofocus with field curvature.

Mine focuses about as sharply using the outer points as the center point wide open on a 5DIII. How much back-focusing are you talking about?

To compensate I have to set a +4 in the microadjustment settings; in that case I end up with the center point front focusing.

P.S.
I brought the lens to the Canon Service Center in Milan (where I live), they said that the softness on the corner is normal. Anyhow, it seems they misunderstood the issue.

 

[nightbreath]

I thought the focus had to be "precise" at the point where one is focusing, and that the field curvature would affect the remaining area of the frame.
I mean:
1. if I focus in the center, the field curvature would make closer the "focus sphere" (should be "focus plane") at the left and right parts of the frame;
2. If I focus on one side (e.g. left), the side where I focus should be spot on; while, the field curvature should make closer the focus sphere on the other side of the frame (center and even more on the right).
Isn't the above correct?

My experience shows that it's not correct. For a weird reason (caused by the lens's spherical abberrations, as far as I know) you should have differenet compensation adjustments for each focus point. Or shoot using one focus point. Otherwise you'll have those focus issues.

To compensate I have to set a +4 in the microadjustment settings; in that case I end up with the center point front focusing.

P.S.
I brought the lens to the Canon Service Center in Milan (where I live), they said that the softness on the corner is normal. Anyhow, it seems they misunderstood the issue.

The dispersion is within the specification. I have AFMA range of ~7 steps accross focus points (from -3 to +3).
That's the price you pay for the bokeh you get with this lens.