Canon RF 35mm f/1.4L VCM coming tomorrow at Cinema EOS event

A 35mm 1.4 prime the size of this lens shouldn't need that much distortion correction. The RF 35 1.8 Macro doesn't have much, so I'm not expecting it here on this L lens either.
Worth noting the RF 35L is smaller and lighter than its EF predecessor, and takes a smaller filter size. I suspect that means we will see some degree of digital correction required, but time will tell.
 
  • Like
Reactions: 1 users
Upvote 0
What's ridiculous about this?

To my mind, lens design is always a big tradeoff game: improve one spec and another gets worse. Distortion is one aberration that can be totally corrected in software and I believe undetectably so, so any time you can improve any other aspect of the lens--be it size, weight, cost, autofocus speed, out of focus highlight shapes and disk contours, lateral chromatic aberration, or any other, and make distortion worse... and then correct the distortion perfectly in software--you're far ahead, no? Would you prefer less distortion and more of other aberrations that cannot be easily and perfectly corrected?
That's pretty self explanatory, it's an excessive amount of lens corrections that was previously not seen by any L-series lens before it. My full quote in context is "This was the first lens I encountered from Canon that leaned on ridiculous distortion corrections to generate a usable image. After I got over that, there was no denying that the photos looked spectacular and somehow remain very sharp." Meaning I was surprised at how significant the corrections were, but there was no denying the final results looked excellent.
 
  • Like
Reactions: 1 user
Upvote 0
it's an excessive amount of lens corrections that was previously not seen by any L-series lens before it. My full quote in context is "This was the first lens I encountered from Canon that leaned on ridiculous distortion corrections to generate a usable image.
I read your whole post several times. I'm still at a loss what's ridiculous about it? It seems like the smartest possible move to me.

Sure, the final results will look excellent. Basically the worst case scenario would be if a post-correction pixel were composed of the average of four pre-correction pixels. But how terrible would that be, exactly? The R5 has 45MP, so 228 pix/mm: sqrt(45M*2/3)/24mm=228.22, right? That means 114lp/mm features would result in zero contrast, even if the lens produced lines of pixel values of 255 and 0. But in fact even if the lens can resolve 114lp/mm, it'd require the lines in the image to align perfectly with the sensor pixels, or 114lp/mm contrast would already be compromised to a worst case of the same zero contrast. And Canon's own MTF charts only go to 30lp/mm.

And that's the case with any amount of distortion correction, whether it's minor or major. Its effect doesn't vary by amount, because as long as you're moving data at least a half-pixel, this is your worst case: obliterating 114lp/mm details that probably aren't even rendered by the lens and even with a perfect lens may be completely obliterated anyway due to lines in the scene falling half-way between two pixels. Whether you're correctly 0.05% or 20%, the worst case is basically the same, and basically beyond the system's ability to resolve anyway.
 
Upvote 0
I read your whole post several times. I'm still at a loss what's ridiculous about it?
He means a lot. That's it.

When you think about it, having a lens with sufficient barrel distortion that the image height is less than 1/2 the sensor diagonal, i.e. the image circle cannot cover the sensor, is pretty crazy. It works, obviously. But it would not work with film, and wouldn't work with a DSLR. Well, maybe you would not have freaked out if you looked through the OVF and saw black corners, but I'd immediately check that the lens was mounted correctly, that the hood was mounted correctly, and if they were then I'd be packing the lens off to Canon for service.
 
Upvote 0
would they migrate to Canon based on the new lens?
They left early because there was no mirrorless 35 L option at all.
They did not leave now.
This lens would have made most of them happy.
They did not like either the EF 35 L or the RF 35 f/1.8.
I love the RF 35 f/1.8 but it has issues in the corners.
The way I use it, the corners do not matter much.
IS does, so I will not be getting this lens.
I am disappointed but I am not angry.
 
  • Like
Reactions: 1 user
Upvote 0
They left early because there was no mirrorless 35 L option at all.
They did not leave now.
This lens would have made most of them happy.
They did not like either the EF 35 L or the RF 35 f/1.8.
I love the RF 35 f/1.8 but it has issues in the corners.
The way I use it, the corners do not matter much.
IS does, so I will not be getting this lens.
I am disappointed but I am not angry.
Yes, IS would have been nice. I can't imagine anyone not happy with the EF 35mm f1.4L II.
 
  • Like
Reactions: 1 users
Upvote 0
What's ridiculous about this?

To my mind, lens design is always a big tradeoff game: improve one spec and another gets worse. Distortion is one aberration that can be totally corrected in software and I believe undetectably so, so any time you can improve any other aspect of the lens--be it size, weight, cost, autofocus speed, out of focus highlight shapes and disk contours, lateral chromatic aberration, or any other, and make distortion worse... and then correct the distortion perfectly in software--you're far ahead, no? Would you prefer less distortion and more of other aberrations that cannot be easily and perfectly corrected?
I agree.. but an interesting article on non-optical lens solution ie
"Metalenses are flat surfaces that are designed to replace full-size optics with a wholly different method. Instead of curved optics, they reduce the aforementioned multiple glass optical systems by compressing them into a much smaller system that bends light using nanoparticles and metasurfaces."
https://petapixel.com/2024/06/03/scientists-have-made-the-thinnest-lens-ever-at-just-3-atoms-thick/
 
  • Love
Reactions: 1 user
Upvote 0
I'd like to understand why the new RF lens has more space between the lens elements and the outer shell of the lens barrel, yet Canon chose not to include IS. I've never missed IS with my current ef 35mm f/1.4L II with the R5's IBIS, but I'd just like to understand this new lens's design a little better. (Sorry I couldn't find a direct, side by side comparison of both lenses straight on.)

Thanks.
 

Attachments

  • 1717598848_IMG_2258857.jpg
    1717598848_IMG_2258857.jpg
    80.2 KB · Views: 2
  • 1468349463_IMG_653993.jpg
    1468349463_IMG_653993.jpg
    78.1 KB · Views: 2
  • Like
Reactions: 1 user
Upvote 0
I'd like to understand why the new RF lens has more space between the lens elements and the outer shell of the lens barrel, yet Canon chose not to include IS. I've never missed IS with my current ef 35mm f/1.4L II with the R5's IBIS, but I'd just like to understand this new lens's design a little better. (Sorry I couldn't find a direct, side by side comparison of both lenses straight on.)

Thanks.
My guess would be weight.
 
  • Like
Reactions: 1 user
Upvote 0
Does the IS mechanism really weigh so much?
I suspect that adding IS to one of the lensgroups would have changed the barrel designs significantly, increasing the diameter significantly (1cm/half an inch) and needing more material to strengthen the construction.
In one of the interviews a Canon rep alluded to the VCM taking up a lot of space, but I always take launch day interviews with a grain of salt, I wouldn't be able to keep my facts straight after 3 straight hours of being asked the same questions :)
 
  • Like
Reactions: 1 user
Upvote 0
My guess would be weight.
I frankly don't know if IS focus groups can simply be added to any lens design, or whether it requires compromises to the overall design in order to make it possible to stabilize images with such a group. I'm leaning towards the latter though. I'd guess the same lens with IS would be bigger, heavier, and more expensive, but ALSO not as sharp, have worse aberrations, distortion, OOF highlight shapes, slower AF, focus breathing, etc. etc.

@YeungLinger I'd suggest watching some of the lens teardowns on Lens Rentals. It seems like there's not much dead space inside these lenses. The high-impact plastic is sturdier than metal (resisting bending, etc.) but also takes more space for equivalent strength. I recall Lens Rentals saying that Canon lenses have the most spots for micro-adjustments during manufacture to get every part better-aligned. The auto-focus motors are wonders of compactness on their own yet really contribute a lot to girth.

As a separate matter, the move to mirrorless, and putting the lens back closer to the sensor, has generally caused a trend I've seen noted in a Canon USA white paper that the best lens designs for RF have much larger rear elements, as a rule, than the similar EF lens, but don't need as big a front element. So while the front element size may seem smaller, it's not that there's nothing behind it that is big. Here's a comparison of the EF35/2 and RF35/1.8. If you try to estimate the amount of "non-optics space" in the lens I don't think it's night and day different.

1718207854476.png
 
  • Like
Reactions: 2 users
Upvote 0
No doubt Canon could have designed it with IS, but evidently it wasn’t a priority.
Actually I meant to add a note to that effect.

Canon sites a bit higher hand-holdability for IS lenses on IBIS bodies than non-IS lenses on IBIS bodies, but it's not a night and day difference. And even if that difference really sounds useful for stilling camera shake, it obviously doesn't do anything for subject movement. And just due to focal length being 35mm, the "reciprocal rule" suggests you get an extra stop or more on the 35mm compared to say the 85mm anyway.

I'm sure those without IBIS would disagree, but since moving from the R to R5 I'm actually keen to see as FEW lenses with IS as possible. The IBIS already allows insanely long shots (Even my worst shots at 1/4 with the 50/1.8 are acceptable, and most 1/2 are OK). I'd prefer the lens aberrations and other features be improved, instead of IS. I'm certain others are doing real-world shooting who would totally disagree, but as for my shooting that's where I stand.
 
  • Like
Reactions: 1 user
Upvote 0