50mm f1.2L

Just wanted to get a quick survey. Does it still have issues with focus shift? I've heard some mixed things and wanted to gain some confirmation on the issue or non-issue.

I know it's a little soft at f1.2 but its the sacrifice you make to shoot at f1.2 - but how's the sharpness at f1.4-f2?

No issues with my 6 month old 50 f/1.2 - and sharpness increases rapidly from f/1.4-2 but is usable (i.e. good but not amazing sharpness) at f/1.2. Color and contrast are awesome from f/1.2 on, unlike the 50 f/1.4 that needs to be stopped down to f/2 before the contrast is acceptable.

No issues with mine. It's a later model. Sharp and pretty much trouble free.

pxchoi wrote in post #14504148
Does it still have issues with focus shift? I've heard some mixed things and wanted to gain some confirmation on the issue or non-issue.

Yes, focus shift is by design. The design has not changed, so it's still there. Whether or not it's an issue for you is up to you.

For me, I have 2 50L's and any focus shift they may or may not exhibit isn't an issue.

Yes, I bought mine new last year and it demonstrates focus shift at close focusing differences. However, whilst the plane of focus moves back, usually the front portion of the area in focus is just about where I want to focus, thus whilst it's sometimes is an issue, never at longer distances (over say 6 ft) and usually taking a couple of shots means I got focus. Conclusion, don't worry about it.

I do not have any issues with mine and it is a couple of months old.

Thanks for the feedback!

It has all the issues you've ever heard about, but it's an awesome lens.

timnosenzo wrote in post #14504788
Yes, focus shift is by design. The design has not changed, so it's still there. Whether or not it's an issue for you is up to you.

For me, I have 2 50L's and any focus shift they may or may not exhibit isn't an issue.

+1
Focus shift is indeed by design. The more one stops down, the more the focus shift.
Shoot wide open = no focus shift.
In any case, this issue could be compensated in firmware. Not sure if this is done on newer bodies though...

CheshireCat wrote in post #14507975
In any case, this issue could be compensated in firmware.

Really!? First time I hear about this. Care to elaborate or share the source where you got this information?

It seems that 5D3 changes everyone's attitude on 50/85 1.2, they became a focus-whatever-you-point lenses..
5D2's center focus point is accurate as hell, just use that one.

Judsonzhao wrote in post #14508356
It seems that 5D3 changes everyone's attitude on 50/85 1.2, they became a focus-whatever-you-point lenses..
5D2's center focus point is accurate as hell, just use that one.

Agreed, the center point is the most accurate for me. Depending on light and aperture, some of the far outer points are a hit and miss. Not a big issue for me at least, i work my way around it. I haven't seen a focus shift with my lens.

pixel_junkie wrote in post #14508096
Really!? First time I hear about this. Care to elaborate or share the source where you got this information?

Sorry, the source is my brain

The focus shift is a function of aperture and focus distance: f(a, d).
No matter how complex this function is, just embed in firmware a 2D lookup-table with discrete samples of f(a, d).
Now, the lens focuses wide open and reports a and d to the CPU. CPU interpolates the 2D LUT and moves the focusing group to compensate accordingly. This may slow down autofocus a bit due to the extra step, and only work in single-shot AF, but I believe it is feasible.

Another solution may be to AF with the lens already stopped down at the actual aperture. AF system performance will be decreased, but this will work with AF servo, down to f/8 on 1D bodies.

But then again, why should you guys shoot at other apertures than f/1.2 ?

mackguyver wrote in post #14504223
Color and contrast are awesome from f/1.2 on, unlike the 50 f/1.4 that needs to be stopped down to f/2 before the contrast is acceptable.

... but color is not going to improve stopping down the 50/1.4. Glass quality matters.

CheshireCat wrote in post #14508458
Sorry, the source is my brain

The focus shift is a function of aperture and focus distance: f(a, d).
No matter how complex this function is, just embed in firmware a 2D lookup-table with discrete samples of f(a, d).
Now, the lens focuses wide open and reports a and d to the CPU. CPU interpolates the 2D LUT and moves the focusing group to compensate accordingly. This may slow down autofocus a bit due to the extra step, and only work in single-shot AF, but I believe it is feasible.

Another solution may be to AF with the lens already stopped down at the actual aperture. AF system performance will be decreased, but this will work with AF servo, down to f/8 on 1D bodies.

But then again, why should you guys shoot at other apertures than f/1.2 ?

Yah, can't happen though. Focus shift varies from camera to camera and it shows itself at different apertures from lens to lens. Some have it much worse than others. So that general algorithm you speak of that may compensate for it may work for one lens on one body and be totally off on another combo and you're back to square one. It may work if you write the software on specific individual lens/body combo basis.