The MA results are sometimes front focused, sometimes back focused from one shot to the next. This is with a newish 50 1.4 and 7D. The lens is all over the place on 7D - mostly front focused in real shooting conditions, including good outdoor sunlight. 7D has been good with my other lenses. And the 50 1.4 is perfect on 5D2, no MA needed.

I am MA-ing in my basement using a big flourescent modeling light - I think 150W equiv. My halogen modeling lightbulb popped the other day, and I haven't replaced it yet. Maybe halogen is better? My computer is in the basement, I'm shooting tethered from a tripod. I am focusing to infinity between shots, then focusing with the computer, then popping into liveview to check results.

I am hoping the light source is the issue? Can anything else cause this type of focus hocus-pocus? It's going to drive me to drinking (mmmm, microbrew).

I can post photos later if I take another set. I got fed up yesterday and deleted them all.

TIA,
Michelle

Hey Michelle,

I did autofocus microadjust (AFM) last month, on quite a few lenses and two bodies; 1DsIII and 1DIV. I also used my computer with LiveView, with the process described at the link below, which seems to be somewhat different from your process. In some cases, I used a USAF 1951 target at the 50x focal length advocated by Canon and at the site below. In other cases, I used the closer distance advocated by the instructions for the LensAlign II. Some of my adjustments were made and/or verified in good, indirect sunlight (i.e. big windows with light curtains, on a sunny day), and others were done with incandescent (halogen) lighting.

My results seemed to be very stable. I didn't do my 50/1.4, as I wanted to concentrate on my fast primes and zooms, with enough Extender combinations that I was limited by the number of cases that can be encoded in the camera.

Maybe you can get more consistent results with good, indirect sunligbht, or halogens. I don't like fluorescents.

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

Cheers, Jim

What kind of exposure times are you getting from your fluorescents? Hopefully at least 1/200 at f1.4

I always have trouble with MA but I believe bright indirect sunlight is best for that procedure. Bright incandescent lighting is probably second best.

This might help:

http://learn.usa.canon.com/app/pdfs/quickguides/CDLC_Accurate_EOS_AF_QuickGuide.pdf

mco_970 wrote:
...focus hocus-pocus...
Michelle

Michelle, you definitely get the wordsmithing prize of the month for that!

I'd try outside, weather permitting. But it sounds like you just might have stumbled on a problematic combination of camera and lens, since your issue doesn't arise with that lens on other cameras or other lenses on that camera. Some have reported similar mismatches with the 35L on a 5DII, so it's not an unheard of scenario.

mco_970 wrote: ...I am focusing to infinity between shots, then focusing with the computer, then popping into liveview to check results. ...I am hoping the light source is the issue? Can anything else cause this type of focus hocus-pocus?

Yes, the wrong choice of -- or misalignment of -- a target on which to focus can cause hocus-pocus due to changing focus locus.

I use outdoor light or a continuous light source for setting MFA.

EBH

I had similar problem with my 7D and 24-70 f2. 8.
Using florescent lights for MA caused back focus issues in non florescent Lighting. Whereas doing Marketing in incandescent caused front focus when shooting in other lighting.

Other people on the site have also commented that they carry a little MA matrix with them for certain lens/camera combos for focal length -distance-and lighting.

Guy

gschlact wrote: ...Whereas doing Marketing in incandescent caused front focus when shooting in other lighting.

"Marketing"?

Three soup cans placed diagonally to each other outside works just fine for me.

You might need to do the magic 7D reset before MA procedure. I don't own one but i have read couple threads about 7D magical reset fixes in the past.

My favourite way to MA is the Real Life Method™
For longer lenses i use birds and shorter distance a human eye.
100% in LR and you see a pattern in which way to adjust.
Sure there is a risk to get an OD of fresh air but what the heck, i can take it!

Ok, I am blowing a BAC 0.0 again. I tried soup cans this morning on the dining table with natural light and a reflector. Got it down to either +2 or +3 for MFA. Great. I decided to spin through and test different apertures and see where my lens looked the sharpest and most contrasty, and got a perfect example of the lens weirdness. My lens was good at f-stops 1.4 to 1.8. At f/2 it started getting worse. This is on a tripod with a remote shutter release, focusing to infinity between shots and refocusing. I would have thought it would only get better as I stopped down.

Is this a focus issue I'm seeing or a shutter speed issue? Is it the lens? Is the 7D's unforgiving pixel pitch part of the issue? The sample image is at 1:2 viewing, so not even a 100% pixel peeping crop.

On that last one with the pair of soup can shots, it's impossible (for me, anyway) to tell whether it's a focus issue or some other lens issue. If you shot something with a ruler at a 45-degree angle, you'd be able to detect whether the lens was focused in front or behind the plane where the AF point was. But with the can, you can't know with certainty if it is or isn't focusing in front.

I'd hhave your lens looked at. There are plastic parts inside that can crack and caue focusing issues that vary from shot to shot. I had Canon fix my 50mm f/1.4 after the issue became very bad, and its perfect now.

Here's a set I just took with a ruler in the sunlight coming in through a greenhouse window & a sample of the random FF phenomenon. Tripod, cable release, focus to infinity between images, etc. The 12mm mark on the rule should be in the exact plane of focus. I am using small focus point and it's on the R49 mark (insulation measuring thingy from Home Depot ).

scalesusa wrote:
I'd hhave your lens looked at. There are plastic parts inside that can crack and caue focusing issues that vary from shot to shot. I had Canon fix my 50mm f/1.4 after the issue became very bad, and its perfect now.

I hate to think about it, but I suppose it is possible. I will throw the lens on the 1D2n or 5D2 and take a set of test photos and see if I can repro the issue with some more bleepity-bleeping test shots.

I bought the 50 new exactly 2 months ago, so hopefully its not broken already.

I did finally put the 28mm on the 7D, and it looks good enough not to bother with MFA. Hooray for that.

Michelle -- the shot of the cans looks like you might have focus shift as you stop down the lens. You need to try thing with more apertures and more detailed subjects off into the distance to confirm. I say this because the first shot has the emblem very sharp on the front curve. The second has the blue lettering on the lower right and the very edge impressions on the left side of the can in better focus than the first shot. Looks like the focus went back nearly 0.5". Since yo focus wide open you don't see the focus shift until the exposure is recorded.

Another check is to use Live View and see if the focus migrates backwards as you close the lens from wide open to about f/8 (after that aperture, DOF starts to cover up any focus shift).

Sorry for your hassles, but I'm sure you'll learn a lot from this.

What I see is that, if that ruler is coming toward the camera at the bottom as it appears to be doing, you've got front-focusing going on in the left-hand shot and rear-focusing in the right-hand shot. They're both labeled as f1.4, so it looks like your intention there is to demonstrate that you're seeing inconsistent results even at the same aperture, right? I think a trip to Canon is in order for that lens.

Hey, thanks guys. It's the lens, definitely. I put it on 1D2n and mounted it where I could see the top of the lens as it focuses. The 50 has the window on top with a distance scale, and where it is stopping and confirming focus varies wildly. Ok, it varies just a bit each direction, but it is enough to cause what looks like random large shifts in front focus.

These are the top distance window in successive shots, using cable release, going to infinity between shots, not touching camera, natural light, etc. on 1D2n. Definitely have learned to watch that window next time I'm having doubts about a lens!