p.3 #1 · 5D3 outside AF point issue with wide lenses
Related to my other point, if it is field curvature, the 5D3 points are farther from the center and therefore more effected by the issue. I assume if Canon really wanted to, they could apply a 'default' field curvature adjustment for wide angle primes, but I'm guessing it would be difficult and could cause more problems than it solves.
p.3 #2 · 5D3 outside AF point issue with wide lenses
AGeoJO wrote:
Nice try, Phil, mine is a stellar copy, too. I just have to be careful in selecting the AF point and minimize recomposing. My copy is also used, I meant, previously owned...
Damn. I was feeling all special for a while
Funny... if I want spot on focus it's ALWAYS the lens I grab..
p.3 #3 · 5D3 outside AF point issue with wide lenses
Yeah... Basically we summed up the issue here, as it seems it happens with most everybody. Quite a few here and on POTN, it seems to be a curvature issue. Would be nice if Canon could compensate for the curvature, since each lens is basically the same as far as distortion goes... So if they add some type of algorithm to adjust while using outer AF points it would make them more reliable. But if both Nikon and Canon have the same issue, maybe it is not so easy or correctable... Especially when both flagship AF systems experience the exact same thing with fast, wide lenses.
I am just happy it is not my specific 5D3 or lens issue. We can pretty much confirm it is lens curvature and how the AF is responding to it.
p.3 #4 · 5D3 outside AF point issue with wide lenses
The soft AF can be a result from a combination of the contrast reading reported by AF sensor(s) and how smart the AF firmware to drive the lens. Use speedlite to assist the outter AF point(s) to trouble shot before we can blame the issue on any lens.
p.3 #7 · 5D3 outside AF point issue with wide lenses
Yohan Pamudji wrote:
Contrast detect AF, anyone? PDAF quirks like this are why I look forward to the day when CDAF is good enough in all applications to kill off PDAF.
So what's the verdict? Don't use outer AF points on wide angle lenses at large aperture because of field curvature?
here is your ans it's all about how smart the AF firmware to decom the reading report from all AF sensor(s) and active sensor
p.3 #8 · 5D3 outside AF point issue with wide lenses
Invertalon wrote:
Correct ^
And mttran, you do not need a speedlite to assist outer AF points when it is bright daylight shooting f/1.4 at 1/8000 second at ISO 100
Basically this is just field curvature, nothing more.
Blargh! So much for the 5DIII + 35L as the ultimate environmental portrait combo. Back to focus-recompose like the bad old days, a.k.a. my current M.O.
p.3 #9 · 5D3 outside AF point issue with wide lenses
Invertalon wrote:
And mttran, you do not need a speedlite to assist outer AF points when it is bright daylight shooting f/1.4 at 1/8000 second at ISO 100
Please note, not just one active sensor but all AF sensors report to firmware and sometime it is not easy for AF firmware to weight the reports. Extra help from speedlite is just to trouble shot the missing pieces from firmware. Speedlite assist works with my 5d2 so it should work with 5d3.
p.3 #10 · 5D3 outside AF point issue with wide lenses
Well, this is highly disappointing. One of the major advantages of having outer focus points (that work!) is the ability to avoid focus-recompose. Especially with wide angle lenses, where focus-recompose will require the greatest angular rotation.
Does field curvature of the lens mean that a different MA will be required for outer points relative to the center, and thus most all fast wide angle lenses will be "off" somehow?
Did people have this same problem with the 1DsIII? It's got good AF sensors on the outer points. Why, on the Nikon side, are these problems only now being reported with the D800? Didn't they show up with the D3, D700, D3x?
Is the only solution to MA the lens/camera combo to an outer point, or to set the MA midway between the readings for the center point and the outer points? Nothing will be tack sharp at wide apertures that way.
My 5D3/35L appears to focus properly wide open both with the center point and the outer points. I haven't tried my 24-70 yet. That lens looks to have much more noticeable field curvature, one reason I never liked doing the focus-recompose dance with it. Looking up at someone on stage, using the center point and then recomposing, always gave me an out of focus face. Reliable outer AF points on the 5D3 were supposed to solve this problem.
p.3 #11 · 5D3 outside AF point issue with wide lenses
In an ideal world, the portion of the image that corresponds to the auto-focus point used should be in focus independent of whether the lens has field curvature or not. For a lens with significant field curvature however, the focus setting for the lens would not be the same for a center point and and edge point, hence the possible lens ring movement.
Using an edge sensor on a lens with field curvature could throw the center part of the image out of focus but the portion of the image covered by the edge sensor should be in focus. If it is not, that would seem to me to be a defect or weakness of the camera.
There may or may not be a problem with the 5D3 regarding outer focus points on wide-angle lenses. The only point I an trying to make is that a lens ring shift when changing from center- to edge-point is not enough to arrive at that conclusion IMO. I think you need to compare 100% crops of the portion of the image covered by the auto-focus point used to come to a more solid conclusion.
p.3 #12 · 5D3 outside AF point issue with wide lenses
Jim Levitt wrote:
My 5D3/35L appears to focus properly wide open both with the center point and the outer points.
If it works for one lens it should work for other lens. Sound like AF firmware has no issue since your 5D3/35L combo is working fine. Other 5D3 might has QA issue required AF sensor point(s) reading to be re-calibrated, assumed MA = 0.
p.3 #13 · 5D3 outside AF point issue with wide lenses
+1 for Eyeball.
I would expect the image immediately under the auto-focus point to have a sharp focus, but as you move in the image away from the auto-focus point, field curvature may cause out-of-focus faster than expected.
Instead of lens-cap shots, I think some brick wall tests will be needed to check this.
p.3 #14 · 5D3 outside AF point issue with wide lenses
hahaha. dont worry. MA kind of fixes it. For my 35L,the center point sharpness doesn't seem to suffer too much even after calibrating MA with outer focus point. Its still considered sharp.
But for my copy of the 50L, you'll see the difference in sharpness - but i don't worry about it too much because I prefer to use the outer points anyway,
I do wish we could use a focus screen for the 5D3. It basically guarantees sharpness once you learn to trust it. No more viewing on LCD screen to check sharpness.
Yohan Pamudji wrote:
Blargh! So much for the 5DIII + 35L as the ultimate environmental portrait combo. Back to focus-recompose like the bad old days, a.k.a. my current M.O.
p.3 #16 · 5D3 outside AF point issue with wide lenses
j5daniel wrote:
hahaha. dont worry. MA kind of fixes it. For my 35L,the center point sharpness doesn't seem to suffer too much even after calibrating MA with outer focus point. Its still considered sharp.
But for my copy of the 50L, you'll see the difference in sharpness - but i don't worry about it too much because I prefer to use the outer points anyway,
I would rather have the best IQ possible with center point and edge points, instead of compromising. If I adjust for outer points, my center point (although still looks OK) does not have the same detail and clarity if I leave it at my adjusted MA for center point only. Kind of sucks... But I guess we have to deal unless Canon comes up with a fix for it.
p.3 #17 · 5D3 outside AF point issue with wide lenses
I just wanted to update this thread... (it's been a while, I know!)
I just purchased a brand new 35L recently... It does not show ANY problem at all using outside AF points. It does not shift at all and nails focus at center or farthest out point with the 5D3. I was expected to have to just use center AF wide open after my last experience and the reason I made this topic in the first place, but I guess I was wrong!
Not sure why the previous 35L game me such a shift when using the outside AF points, but this new copy of the 35L is perfect. I can use any AF point with NO problem what-so-ever. So happy.
So what I thought was the camera, may have been the lens itself... Which I find weird, but only Canon would know why I suppose.
Either way, it is nice to have a 35L with the 5D3 again... And able to use any AF point I wish. Makes me very happy!
PDAF quirks like this are why I look forward to the day when CDAF is good enough in all applications to kill off PDAF.