Canon adds focus breathing correction support to f/2.8 and f/4 RF zoom lenses via firmware

[mxwphoto]

It's interesting they only have it for zoom lenses. I would have thought primes would be easier to calculate and correct for focus breathing as you only have one focal length to test and program for.

 

[Skyscraperfan]

Does it really come without a loss of sharpness, if the video is dynamically oversampled from different resolutions because of the focus breathing?

 

[Del Paso]

Focus shift is a 'feature' (meaning a distinctive attribute) of the RF 100mm Macro lens design. Not a desirable one, but a feature nonetheless. Apparently Canon felt the design tradeoff of going to 1.4x magnification was worth the consequence of focus shift. While it would be possible to address it with firmware, it's a more difficult proposition than correcting focus breathing.

I had understood "feature" as an intentional characteristic, and not as an "accepted" consequence of the S.A. Control implementation.
My bad.

 

[neuroanatomist]

I had understood "feature" as an intentional characteristic, and not as an "accepted" consequence of the S.A. Control implementation.
My bad.

It’s a consequence of the high magnification, not the SA control.

 

[AlanF]

Kai shows it briefly in his preview, from 6:23 till 6:33:

I thought that focus breathing was the change in focal length with distance. Does it have another meaning as well as Kai is talking about something completely different?

 

[neuroanatomist]

I thought that focus breathing was the change in focal length with distance. Does it have another meaning as well as Kai is talking about something completely different?

Yes, focus breathing is the change in effective focal length as focus distance is changed.

In the video, Kai literally just mentions the setting and says, “It’s great if you don’t like breathing, well, not actual breathing.” Then he moves on to talk about…ramble on about may be more apt…a different menu setting controlling how the camera behaves when a subject leaves the frame.

 

[paul]

One simple way (pointed out above) that would work on MILCs is to have the lens focus at the selected f/stop instead of focusing wide open then stopping down (since it's the latter that causes the shift). That would be quite easy to implement in firmware. It could be implemented as an option, perhaps only being applied at focus distances >0.5x magnification.

Alternatively, the focus distance of the lens could be shifted slightly based on the selected aperture. The problem with the latter approach is that the magnitude of the shift varies with both the extent to which the aperture changes relative to wide open and the distance to the subject, and the latter is very difficult for the camera to estimate at macro working distances.

Isn't that an option on the R3, choice to focus with selected F-stop instead of wide-open? Which I hate, focussing stopped down, and the main reason to move back to Canon from Sony( after 3 years)

 

[neuroanatomist]

Isn't that an option on the R3, choice to focus with selected F-stop instead of wide-open? Which I hate, focussing stopped down, and the main reason to move back to Canon from Sony( after 3 years)

Nope. There’s the Display Simulation mode with DoF where the lens is stopped down to the selected aperture while you look through the VF. The R3, R7, R10 and R6II have that capability. But when you actually take the shot, the aperture opens up, the camera focuses, then it stops back down.

That’s why not all lenses support the feature. All RF lenses do, but only the more recent EF lenses have a fast enough iris diaphragm motor.

Canon : Supplemental Information : Lenses Compatible with [Exposure+DOF] in [Display simulation]

 

[EOS 4 Life]

No, SA adjustment is a feature, focus shift is an issue . And if Canon can correct focus breathing, they sure could correct via firmware focus shift.

Focus breathing correction is just doing a video edit in camera that can be done in post.
It is a big deal for live video though.
Focus shift compensation would be making something not in focus in focus.
I say all that to say that I still think it can be fixed.
It is just that fixing one is no indication that the other can be fixed.
They are very different issues with very different fixes.

 

[EOS 4 Life]

Is this something we can select or no? I don't care for it. It's interesting in an age where people are concerned that modern lenses look too sterile in their rendering, we still have an obsession with focus breathing. I feel like a lot of cinema icons prefer it even.

You are the first person I have ever heard of to prefer focus breathing.
I guess it can be used artistically like everything else.

 

[EOS 4 Life]

Does it really come without a loss of sharpness, if the video is dynamically oversampled from different resolutions because of the focus breathing?

Yes, but it might not be much.
The way Sony does it is kind of trash.
I assume that it does not work in 6K RAW and there is plenty of room between 6K and 4K to still remain sharp.

 

[paul]

Nope. There’s the Display Simulation mode with DoF where the lens is stopped down to the selected aperture while you look through the VF. The R3, R7, R10 and R6II have that capability. But when you actually take the shot, the aperture opens up, the camera focuses, then it stops back down.

That’s why not all lenses support the feature. All RF lenses do, but only the more recent EF lenses have a fast enough iris diaphragm motor.

Canon : Supplemental Information : Lenses Compatible with [Exposure+DOF] in [Display simulation]

Ah, is that how it works( having R5's myself) . Anyway, I prefer to focus wide open, and push that button with my finger, the old DSLR way maybe, if I want to check aperture and depth of field.

 

[neuroanatomist]

Ah, is that how it works( having R5's myself) . Anyway, I prefer to focus wide open, and push that button with my finger, the old DSLR way maybe, if I want to check aperture and depth of field.

I really like seeing the DoF, personally. On the R3, there is an OVF simulation mode as well (basically HDR for the EVF), which is mutually exclusive with Display Simulation so I use the latter. I'm typically in Auto ISO, so the VF remains consistently bright and I can see the DoF change dynamically as I change aperture.

 

[SHAMwow]

You are the first person I have ever heard of to prefer focus breathing.
I guess it can be used artistically like everything else.

I guess my clarification would be that I treat it like vignette and profile correction. I like my photos to have rendering from their respective lenses. For instance, some people widely correct distortion on their wide angle lens, but for me that is the point of the wide angle lens. To get that distortion falloff. It's iconic. I want my image to have a character that matches the design of the lens. I'm not some old lens purist, but I do think that mechanically a lens is what it is. Focus breathing is part of that to me. I've been shooting for twenty years, and I've never once said "man that focus breathing killed my work or the shot I wanted".

 

[koenkooi]

I guess my clarification would be that I treat it like vignette and profile correction. I like my photos to have rendering from their respective lenses. For instance, some people widely correct distortion on their wide angle lens, but for me that is the point of the wide angle lens. To get that distortion falloff. It's iconic. I want my image to have a character that matches the design of the lens. I'm not some old lens purist, but I do think that mechanically a lens is what it is. Focus breathing is part of that to me. I've been shooting for twenty years, and I've never once said "man that focus breathing killed my work or the shot I wanted".

Do you mean focus breathing specifically or spherical aberration?

 

[tq0cr5i]

This old coach may never take videos. Focus breathing is harmful for both film makers and audience. Cinema lens and broadcasting lenses usually have little focus breathing and much expensive than photo lenses.

 

[t.linn]

I'm hoping for that as well, but reports on other fora imply that not every copy has the same amount of shift. So far I've been lucky enough to be the weakest link when using the RF100, the shift hasn't been an issue yet.

My impression is that focus shift on the RF100 requires a pretty specific set of shooting conditions for it to become an issue: high magnification, close focus distance, and a particular aperture range. I am not suggesting that there is no copy-to-copy variation but it wouldn't surprise me if people who go looking for the problem but don't really understand it have trouble finding it. I think this is the reason Canon isn't too worried about it. That said, it would be nice if the issue could be addressed. I would like to see Canon address it by allowing AF at smaller-than-maximum apertures on all lenses.

 

[danfaz]

My impression is that focus shift on the RF100 requires a pretty specific set of shooting conditions for it to become an issue: high magnification, close focus distance, and a particular aperture range.

Yes, it comes into play on apertures between 2.8 and 11. I've replicated it on mine, but honestly I tend to shoot at either 2.8 (and focus stack) or 16, so it doesn't bother me.

 

[SHAMwow]

This old coach may never take videos. Focus breathing is harmful for both film makers and audience. Cinema lens and broadcasting lenses usually have little focus breathing and much expensive than photo lenses.

Broadcast no, and new cinema no. But I see focus breathing in cinema so often I couldn't tell you how many films it's in. And the only people I've seen worried about it are pixel peeping on charts. To each their own, just sharing my thoughts on it.

 

[SwissFrank]

I'm hoping for that as well, but reports on other fora imply that not every copy has the same amount of shift. So far I've been lucky enough to be the weakest link when using the RF100, the shift hasn't been an issue yet.

I'd love to see the info you refer to as it would be educational for me. From what I know, it'd be impossible for these lenses to have sample-to-sample variation of a human-detectable amount under normal conditions. Maybe if you're shooting test targets from a fixed tripod and looking at individual pixels? Yet even then, the entire point of focus breathing is that you're comparing the size of something when it's in focus to when it's NOT in focus... and how can a human eye make that comparison accurately when the out-of-focus shot is blurred to the order of 50 or even 500 pixels?? Tolerances of the parts in the lens are far, far below a millimeter and even what little variation there is has a lot of adjustments.

Again, happy to believe I'm wrong and you're right, but I'd like to see the specifics of what you're referring to. Is it side-by-side apples-and-apples comparisons? Or is it just some guy who had two copies of a lens and is reporting the old one had a different amount of focus breathing, without giving apples-and-apples examples?