Autofocus and critical sharpness

[Frank Kolwicz]

Preliminary note: I still haven't gotten satisfaction from Canon on the repair of my EF600/4L IS(I).

As part of that long on-going quest I recently performed a series of target and field test images for submission to Canon's repair department, including star photos of Polaris and vicinity. All three sets of test images show focus defects due to optical misalignment. The star photos were interpreted by a scientist who specializes in optical imaging for the space program as showing decentering and possibly "wedging" of one or more elements. The terrestrial targets show the inability to produce critical sharpness under the best of conditions. The field shots show wide variation in plane of best focus in AF, both in phase-detect and contrast-detect modes, as well as poor sharpness in the best region.

I hadn't realized this until yesterday, but it seems likely that the erratic and out-of-spec, AF performance I have been getting is most likely due to the optical defect(s) of the lens (I have good performance with a 100-400L, obviously better than the 600).

For my 70d bodies and the f/4 600mm lens,  "Canon specifies the precision of the AF system to be within one depth of focus for a lens’ maximum aperture for standard-precision AF points" (From the Digital Picture article on Microfocus adjustment http://www.the-digital-picture.com/Photography-Tips/AF-Microadjustment-Tips.aspx). I wasn't able to find a direct quote from Canon with a Google search.

When a lens's focus degrades, it seems to me that there will be a wider range of almost-in-focus area that probably prevents the AF system from getting a tight fix on the best focus zone and images will show a much greater variation of plane of best focus, further in front and behind the subject - which is what I've been seeing. I didn't understand what was happening before and thought that it was me screwing-up in my technique, despite my years of experience with 600mm lenses and previous ability to get critical sharpness at shutter speeds down to below 1/100th of a second with my current setup when I had a lens that worked properly. There are times when AF is so far off that I get scores, even up to hundreds, of totally out of focus frames in a session, but then I'll get a few that are barely acceptable for small prints and so on, over and over. It's not me, it's the lens, dammit!

[mjperini]

Frank,
Sorry to hear about the lens problem, and while I'm not qualified to speak to those lens issues ,  I did recently purchase a 70D/18-135 combo for my wife to use as a light weight backup camera to her 7D /70-200 f/2.8 L IS combo  and have noticed a high percentage of slightly soft pictures. I'm currently trying to get to the bottom of that.
In my research I came across the fact that Canon has published a list of lenses that have limited compatibility with the 70D focus system . A 600mm f/4 is on the list. (I'm not sure if that's your lens or not) I also came across a rather technical explanation of 70D AF by Chuck Westfall
Limited Compatibility Lens list:http://www.the-digital-picture.com/News/News-Post.aspx?News=7260
Chuck Westfall:  http://www.imaging-resource.com/news/2014/02/27/canon-dual-pixel-cmos-af-autofocus-secrets-of-the-canon-70d-explained

I'm not sure if any of this could be part of your problem or not, but I thought I'd pass it along.

[Frank Kolwicz]

Thanks, Bob,

The 600 on the list is the non-IS model, the one previous to mine, which is the first IS version, so presumably it is compatible.

I'm interested in seeing what an incompatible lens will do on my 70d and I happen to have one, a 1990's era 100-300 f/5.6L. If it is anything notable, I'll send a note.

[mjperini]

There seems to be quite a hubbub about 70D AF performance on various web locations.  (non- video AF) in my very limited use of video and an STM lens I was very impressed.
As I reported earlier , my Wife used here regular wedding combo of 7D & 17-55 2.8 and 70-200 f/2.8 L IS for 7/8 of the last wedding and switched to the 70D 18-135 stm for the tail end.
There were a lot of unsharp pictures, and AWB seemed all over the lot.
I have not yet had a chance to do any controlled testing, nor have I gone through and set up the most optimal settings for still AF in dim light, with and without flash.
I need to do that, but the still AF in the 70D is very similar to the 7D which has always been a great focusing camera for her.
But I do think SOMETHING is up with the 70D AF, even if it is just that it needs more user attention to setup. I also don't know if the STM lenses need different AF optimization than standard USM.
When I get a chance to do some controlled testing I'll report back if I find anything interesting. I don't want to be critical without doing my homework, but I was taken aback by how different the two cameras performed.  I expected the 70D to be better than the 7D.  It may come down to user / setup  error  (or an 18-135 issue)
Michael

[Frank Kolwicz]

I got a call from Canon Pro Services yesterday (I recently signed-up hoping for a better response than previously) and the service tech said that he did see that there was a problem and that the case would be kicked up the chain for a response and that I would hear back tomorrow morning.

In the mean time, I've rented a 500/4LIS(II) and 1.4 extender (III), which I will start using this evening. I got the 500 instead of a 600 because I wanted to see if the loss in magnification would be slight enough to make the easier handling worthwhile with my arthritis and shoulder damage, in case I need to make a purchase decision in the future. It will also give me an easy way to find the images for comparison, if I ever need them, because Canon's exif data does not include the lens serial number, so I can't search for files from the several rental lenses among the archives of my images.

I've also been doing some file editing and gone back to 2012, when I had a good lens, and was well gratified to see how good the images should be!

Frank