Soft image from Sony GM 16-35 f/2.8 at f/16 - diffraction?

A

Andy9

Well-known member
I shot some 4K XAVC-S footage at f/16 with the Sony a7Siii and GM 16-35 2.8 (towards the wide end) and it is noticeably soft. I've done a bit of testing and can reproduce this (f/2.8 much sharper than f/16) with video. I am shooting 4K for a 1080 timeline and it is particularly noticeable (of course) when I scale the image up so I initially assumed it was a PPro scaling issue. It isn't though, I am never scaling beyond 100% so even though I'm not using all of the FOV of the lens, it's still an oversampled image. And I can reproduce this in any software.

Strangely though, I don't see the same effect with stills. Is the a7Siii applying diffraction reduction processing to stills but not to video? There is some mention in Sony marketing material of "diffraction reduction" being a feature of various older Sony cameras. Perhaps this is so routine now that it is no longer considered a feature worth listing and is always turned on.

Or have I got the wrong end of the stick entirely? To be honest, I was surprised to see such a dramatic effect. I'd heard of diffraction, but I assumed that at f/16 you'd need to really pixel peep to detect it. I'm normally shooting with larger apertures so it's not something I'd ever considered much before.
 
IDK what the camera/lens is doing in its photography mode, but in general, and according to Christopher Frost (lens testing legend), diffraction starts around f/11 in this lens (although he specifically said that at the 35mm testing part so it's not entirely clear if that applies to 16mm as well).

If so, f/16 would likely be terrible, which actually surprises me just like you (since its max is f/22).

And, IMO, it would be better to test on a higher MP camera for stills to see how it performs.

 
NorBro said:
...f/16 would likely be terrible...
Click to expand...

Yeah, no, not so much. It's likely you'd just begin to see it. In a still. If you magnified it enough. If you were really hunting for it. You'd probably be hard pressed to identify diffraction softening in a video, because the point of a motion picture is that something is in motion.

As an example, Ansel Adams, well known for the sharpness and detail in his prints, and a co-founder of Group f/64, shot a huge amount of 8x10 film. To keep everything he wanted in focus (which for Adams was usually everything visible in the print) he basically lived in the diffraction zone (for him, anything between f/22 and f/128 or so). But no one complains that his prints are anything less than razor sharp. Because... they are razor sharp. His prints hang in many art museums today, so there's no need to take my word for that -- go find a print and have a look.

Diffraction softening is a thing, but it's not that important a thing. Just sayin'...
 
I disagree. diffraction is just like waves going through a harbour gate .. a physics phenomenon not attached to any specific lens but only to the size of the hole.

Getting a cooke or zeise wnt cure it.

Its gonna be bad a f11./16 on ff35 yes a problem.

andsel had a chip 100 odd times bigger so f128 was a lot bigger hole. Wheil the problem might have been as bad.. as a proportion of chip size/final image it wasnt.

Differaction is proportional to hole size whereas f is fuunction of hole size and chip size? did I just make that up|?

To get sharpness and a lot of DOF you need a small chip camera?
 
Bruce Watson said:
Yeah, no, not so much. It's likely you'd just begin to see it. In a still. If you magnified it enough. If you were really hunting for it. You'd probably be hard pressed to identify diffraction softening in a video, because the point of a motion picture is that something is in motion.

As an example, Ansel Adams, well known for the sharpness and detail in his prints, and a co-founder of Group f/64, shot a huge amount of 8x10 film. To keep everything he wanted in focus (which for Adams was usually everything visible in the print) he basically lived in the diffraction zone (for him, anything between f/22 and f/128 or so). But no one complains that his prints are anything less than razor sharp. Because... they are razor sharp. His prints hang in many art museums today, so there's no need to take my word for that -- go find a print and have a look.

Diffraction softening is a thing, but it's not that important a thing. Just sayin'...
Click to expand...

Yeah, I mean, I can't argue with what you may or may not see with your own eyes (and our brains process results differently from one another).

I can only go by my own tests from my own experience, and diffraction is pretty obvious in video and it is a problem.

I can see the difference between using a lens in its optimal aperture zone and the end of the aperture zone. There is no doubt about it.

And sometimes when people don't see comparisons between one result and a better result, ignorance is bliss, so to speak. So they may think it's sharp because they have never seen anything sharper and don't know any better.
 
In my experience, diffraction is a problem with every lens and every camera I have ever owned. Plus, smaller apertures will reveal all the dust and crap on your sensor. Therefore, 99.9% of everything I shoot is wtihin 1-2 stops of wide open.
 
..which is all very well until your client needs some jewels or something filmed.

Ive hit this a couple of times and shooting loose cropping in post usually gets me as far as I can go. (f8 at 1m has more DOF than F8 at .5m)

focus stacking for stills or 3d scan for video will take you further

scroll down to the CU shots.. coins.. my buddy nick .. focus stack with the nikon z6 last week

you can be working for a train company and still end up shooting small things.

https://www.telegraph.co.uk/news/202...rthed-hs2-dig/
 
Andy9 said:
I shot some 4K XAVC-S footage at f/16 with the Sony a7Siii and GM 16-35 2.8 (towards the wide end) and it is noticeably soft. I've done a bit of testing and can reproduce this (f/2.8 much sharper than f/16) with video. I am shooting 4K for a 1080 timeline and it is particularly noticeable (of course) when I scale the image up so I initially assumed it was a PPro scaling issue. It isn't though, I am never scaling beyond 100% so even though I'm not using all of the FOV of the lens, it's still an oversampled image. And I can reproduce this in any software.

Strangely though, I don't see the same effect with stills. Is the a7Siii applying diffraction reduction processing to stills but not to video? There is some mention in Sony marketing material of "diffraction reduction" being a feature of various older Sony cameras. Perhaps this is so routine now that it is no longer considered a feature worth listing and is always turned on.

Or have I got the wrong end of the stick entirely? To be honest, I was surprised to see such a dramatic effect. I'd heard of diffraction, but I assumed that at f/16 you'd need to really pixel peep to detect it. I'm normally shooting with larger apertures so it's not something I'd ever considered much before.
Click to expand...
How images are captured by camera sensors have grown in complexity. Back in the film days we could narrow such a discussion to only the lens. You might want to read this article https://www.bhphotovideo.com/explora...RoC54wQAvD_BwE
While it might not give you a solution you'll get a better understand of why its not as simple as you might think.

What has always been true is at either extreme of aperture openings will result in poorer image quality. Of course lens can be designed to be sharper at either end but generally 5.6 is ideal.

I think the root of your issue is simply closing down the lens to give a full frame camera a large dof has its limitations. This is why there is still a place for smaller sensors. In fact that's what Hollywood commonly uses. There is also the reverse issue of small sensors trying to get shallow dof by apertures at the opposite extreme. Neither larger or small sensors are better they have limitations that can't always be overcome by the aperture of the lens.
 
Last edited:
You must log in or register to reply here.
Back
Top