[Klaus]

Taken from another thread:

http://www.bokehtests.com/page2/index.html

 

Quite interesting topic. I have never tried to recreate this so far.

 

Any experience ?

[popo]

I had a go... this test was not easy. Basic procedure. Using 7D + 100-400L at 400mm, set to manual, f/5.6, auto ISO. First take some shots at 1/640s with IS off. Then turn on IS and set shutter to 1/50s and repeat. When taking shots with IS, I half pressed for a moment to let the IS settle before taking the shot.

 

Test 1: Use a close subject, relatively close background. Of a ball park distance of subject at 2m, background at 5m say. This test was inconclusive as the background was so blurred out, it was impossible to make out differences. There were bigger differences from me not framing exactly the same between shots.

 

Test 2, Let's try a not so close subject and relatively closer background. I don't know the exact distances, but let's say subject 5m, background 6 or 7m. This was more interesting. Most of the shots look similar, but one part of one of the IS on shots looked worse than the others.

 

Test 3. I went to a bedroom window and focused on the rooftop opposite, with distant houses in background. As I flicked through the IS on shots, I was seeing variation in the bokeh characteristic. Have I found it? No, that was when I noticed the focus plane was moving between the shots. I'm more likely seeing the effects of varying defocus there. That's an easy fix, I did one AF operation and switched to manual focus mode. Now I can eliminate focus as a contributing factor. This mechanism likely invalidates the 2nd test.

 

Selected results of test 3 are attached. I would add at this point images were taken in raw and were processed in DxO 10 with defaults. I didn't change anything. The first shot is the IS off, fast shutter speed one. I'm getting too old I think, but there was quite a bit of movement in the viewfinder so even this may not be movement free. Anyway, shots 2 and 3 were taken with the longer shutter time and IS on. Shot 2 is noticeably blurred in the foreground, so I must have had more shake on that.

 

So, can you see a difference in the defocus areas? I don't think I can. And if we have to look this hard, I'd argue if there is any effect, it is insignificant in real world situations.

 

Maybe there is a worst case subject and background distance that will show up the effect more clearly. Or some lens negative bokeh characteristic that could be amplified by this mechanism. If I have to try this hard to try and find it, I don't think it is something I need to worry about.

 

Edit: I note the forum image uploader has recompressed the images, but I think it will still suffice to look in general at out of focus areas.

[Klaus]

I think it's probably very hard to see the effect in the general blur but more so in OOF highlights.

[stoppingdown]

BTW, Thom Hogan commenting on the A7m2 just wrote:

 

One of the side effects of IS that I dislike the most is how it sometimes distorts out of focus areas in unusual ways. 

[Klaus]

Yeah, this is my point. An OOF highlight is no longer circular. The question is when this is getting relevant (at which focal length and shutter speed). No idea really.