Canon announces the Canon RF 10-20mm f/4L IS STM

[GMCPhotographics]

This new lens is a great opportunity for RF customers. I think the light weight, small size and sharp image Quality looks like compelling reasons to purchase this lens. However, for my needs, the current EF 11-24mm perfectly fullfills my current shooting needs. I aslo get to drop in rear mounted filters which is a serious advantage with top performing EF glass. Looking at some reviews and examples, the lack of front mounting filters will sevearly limit it's applicable use in genres like low light landscape work. If canon were to introduce internal filters into their bodies (CP and ND options) this would change things substantially.
the other issue that I've seen is the Chronic vignetting in the corners. It's beyond thee stops when shooting wide open. Sure, this can be corrected via plugins / software but 3 stops will eat into your corner iso noice lattitude. There's a fair argument that a photographer will stop down substantially for depth of field so it's a mute argument. But as it stands, wide open, this lens has heavy vignetting in the corners. I guess that's the trade off for it's relative size vs the EF version.
It's also very expensive, but then so is all of the RF glass. We know they are great, but eye wateringly expensive. I also suspect that Canon have been quite dellberate with their long end focal length choice for this particular lens. Cutting it to only 20mm really does limit it's versatility in my opinion. I can't imagine anyone choosing this lens as their one and only ultrawide lens of choice. My EF 11-24mm F4 L is my UWA lens of choice, (I keep an old EF 16-35IIL for wet work, near salt water coast lines). I can't imagine anyone doing the same for this lens. So it's quite clear to me that Canon is exepcting us to buy this lens in addition to a 15-35 type of lens. I can't imagine anyone choosing this lens to be mated to a 24-105 and a 100-400 as an F4 trinity of lenses. I think this niche versatility will relegate this lens to a few photographers who need the heavy rectilinear nature and it's probable architectural applications. Estate Agents will love this lens, but I don't think I'll be seeing it around on holiday in Portugal.

 

[EOS 4 Life]

If canon were to introduce internal filters into their bodies (CP and ND options) this would change things substantially.

I agree but this lens would be a lot bigger.
Canon seems to really emphasize the size and weight advantage of this lens.
Gel filters are a pain though.

 

[AlanF]

This new lens is a great opportunity for RF customers. I think the light weight, small size and sharp image Quality looks like compelling reasons to purchase this lens. However, for my needs, the current EF 11-24mm perfectly fullfills my current shooting needs. ..... I think this niche versatility will relegate this lens to a few photographers who need the heavy rectilinear nature and it's probable architectural applications. Estate Agents will love this lens, but I don't think I'll be seeing it around on holiday in Portugal.

Can't say I've seen many EF 11-24mm on holidays in Portugal or anywhere. Both the EF and the RF are specialist lenses. 10% wider at the short end is more versatile than 20% shorter at the long as you can't crop up. But, maybe you use 24mm a lot

 

[neuroanatomist]

from my practical experience, you really don't see that difference between .6 and .7 in the corners and it also has better astigmatism control as well, so should look much smoother. I did say show a noticable difference but i'll make that a bit more clearer. also in your superimposed chart, there's two 30lines/mm the 11-24 is better than the 10-20 on the sagittal not on meridional. think you'd see that much difference? also the 10-20 has higher center resolution at 20mm versus the 11-24mm as well, as it doesn't fall to under .7 until 15mm (think asp-c crop), while the 11-24 drops to .7 at 8mm. so overall i'd call them a draw. overall the 10-20 has more consistent center performance than the 11-24mm.

Yep - totally my mistake. My fault for initially flipping back and forth between tabs on my phone, then not actually looking at the plots as superimposed. Even this morning, I thought I had mislabeled them on the graph, when they were only mixed up in my head. Big oops.

They are similar at the wide end, the RF lens has a definite improvement over the EF at the (not quite as) long end.

My apologies!

 

[neuroanatomist]

I recall that the wide end of the 14-35 is actually at around 12-13mm, so that a 14mm image is achieved by both cropping and stretching the corners.

That is incorrect. There was speculation some time back (including by me) that there was cropping going on. IIRC, that was based on misinterpretation of RAW images from the 24-240 and 16/2.8, where people were comparing framing after applying a distortion correction of 100 in LR, not realizing that didn’t mean 100% and that only ~40 was needed to fully correct the distortion. Cropping would mean the camera was secretly upscaling the images, it’s not.

In any case, there’s no cropping. Distortion correction is sufficient to fill the corners. Comparing RAW images to optically corrected lenses and looking at the relevant patents confirm that these force-corrected lenses aren’t wider that their designated focal length.

After correction in DPP (or with an in-camera jpg), the FoV of the 14-35 at the wide end matches the framing of the EF 11-24 set to 14mm. With 3rd party RAW converters, there is more correction in their profiles. ACR/LR give a bit wider than 14mm, DxO gives ~13.5mm. In addition, with DxO you can opt to not force a 3:2 aspect ratio – because the image is a rectangle inscribed in the image circle, the image needs more stretching on the horizontal axis, thus you get an even wider (but not taller) FoV by not cropping off the sides.

If the image circle for the 10-20/4 at the 10mm end does not cover the FF corners, there's not much in-camera corrections can do about the dark corners short of cloning. I wonder if the wide-open end of the 10-20/4 is actually wider than 10mm if it needs fairly heavy corrections.

The patent says it’s 10.3mm at the wide end.

 

[Hector1970]

I'm in a similar position, except I already have the RF 14-35/4. It's an excellent lens, and by comparison to the 11-24 the 14-35 actually delivers ~13.5mm FoV after distortion correction and is as sharp in the corners. Very happy with the 14-35/4 as a replacement for the EF 16-35/4.

Yes that would be an attractive lens to me. I’m still on mainly RF lens. Adapter is fine but sat some point I need to fully embrace RF lens

 

[neuroanatomist]

Adapter is fine but sat some point I need to fully embrace RF lens

Why? Honestly curious.

Consider how difficult it will be to filter the RF 10-20mm. Gels are annoying at best, and there's no CPL option for them. 186mm filters are a PITA to carry around. For lenses like the 11-24 and the TS-E 17, the drop-in filters are a much more convenient option.

I guess 'at some point' may be far in the future. No MP-E 65 and no TS-E lenses for RF, yet.

But if one uses only 'mainstream' lenses (trinity-type zooms, etc.), then completely switching to RF makes sense even today, affordability notwithstanding.

 

[Berowne]

It is a new stabilization algoritme that should help against the wobbling effect, seen in the corners in video.

Canon RF 10-20mm F4L IS STM

Discover Canon RF 10-20mm F4L IS STM Lens, the ultimate choice for capturing breathtaking landscapes and architectural details with maximum precision.

www.canon-europe.com

Thanks, But why is it only working with the R5? Is the digital IS at work in this case?

 

[HMC11]

That is incorrect. There was speculation some time back (including by me) that there was cropping going on. IIRC, that was based on misinterpretation of RAW images from the 24-240 and 16/2.8, where people were comparing framing after applying a distortion correction of 100 in LR, not realizing that didn’t mean 100% and that only ~40 was needed to fully correct the distortion. Cropping would mean the camera was secretly upscaling the images, it’s not.

In any case, there’s no cropping. Distortion correction is sufficient to fill the corners. Comparing RAW images to optically corrected lenses and looking at the relevant patents confirm that these force-corrected lenses aren’t wider that their designated focal length.

After correction in DPP (or with an in-camera jpg), the FoV of the 14-35 at the wide end matches the framing of the EF 11-24 set to 14mm. With 3rd party RAW converters, there is more correction in their profiles. ACR/LR give a bit wider than 14mm, DxO gives ~13.5mm. In addition, with DxO you can opt to not force a 3:2 aspect ratio – because the image is a rectangle inscribed in the image circle, the image needs more stretching on the horizontal axis, thus you get an even wider (but not taller) FoV by not cropping off the sides.


The patent says it’s 10.3mm at the wide end.

Thanks for the explanation - appreciated please. However, I am still somewhat confused. If you look at the photos in this link (https://admiringlight.com/blog/review-canon-rf-14-35mm-f-4l-is-usm/2/#Image-Quality) of the Ohio Supreme Court, it does seem as if there is some cropping along both the horizontal and vertical directions. If the second picture is the corrected version at 14mm, would the uncorrected one then have a FOV that is wider than 14mm? The dark corners seem to have been removed via cropping. Can you help to interpret the photos? Thanks.

 

[YuengLinger]

Why? Honestly curious.

Consider how difficult it will be to filter the RF 10-20mm. Gels are annoying at best, and there's no CPL option for them. 186mm filters are a PITA to carry around. For lenses like the 11-24 and the TS-E 17, the drop-in filters are a much more convenient option.

I guess 'at some point' may be far in the future. No MP-E 65 and no TS-E lenses for RF, yet.

But if one uses only 'mainstream' lenses (trinity-type zooms, etc.), then completely switching to RF makes sense even today, affordability notwithstanding.

I agree with the case by case approach. For example, my EF 70-200mm f/2.8L IS was ergonomically awkward for me when using vertically on the EOS R. The extender was just long enough to make the lens tip slightly front heavy for me. I just couldn't get comfortable with it, so I sold it and bought the RF version.

My EF 35mm f/1.4L II is so good that I would not upgrade it for any new features in an RF version. I intend to use it as long as it works and remains in my possession. I also feel the same way about the EF 100 f/2.8L IS Macro. Plus I'm keeping some non-L lenses, such as the ef 85mm f/1.8 for my kids, as the oldest will be ready soon to start using an 80D. (Eventually she will end up with the R6, with the 80D, should it survive, being handed down to her brother).

All in all, Canon sure did a great job of transitioning!

 

[P-visie]

Thanks for the explanation - appreciated please. However, I am still somewhat confused. If you look at the photos in this link (https://admiringlight.com/blog/review-canon-rf-14-35mm-f-4l-is-usm/2/#Image-Quality) of the Ohio Supreme Court, it does seem as if there is some cropping along both the horizontal and vertical directions. If the second picture is the corrected version at 14mm, would the uncorrected one then have a FOV that is wider than 14mm? The dark corners seem to have been removed via cropping. Can you help to interpret the photos? Thanks.

Quoted from the review text: “The uncorrected image is also a fair bit wider than 14mm, with the field of view reduced a little when correcting the distortion, however, natively it’s not quite as dramatic as shown in the second image, as this also was perspective corrected for keystoning, further reducing the angle of view. “.
The second image has been corrected in post processing (e.g. in Lightroom or Photoshop) for perspective to prevent the building from falling over. This perspective correction makes that the resulting image is cropped.

 

[neuroanatomist]

Thanks for the explanation - appreciated please. However, I am still somewhat confused. If you look at the photos in this link (https://admiringlight.com/blog/review-canon-rf-14-35mm-f-4l-is-usm/2/#Image-Quality) of the Ohio Supreme Court, it does seem as if there is some cropping along both the horizontal and vertical directions. If the second picture is the corrected version at 14mm, would the uncorrected one then have a FOV that is wider than 14mm? The dark corners seem to have been removed via cropping. Can you help to interpret the photos? Thanks.

For the RF 14-35mm, the uncorrected FoV is slightly less than 13mm. The distortion correction reduces the FoV, but because the image is 'stretched out' (after being 'squished' by the distortion), not through cropping. Having said that, there is some cropping at the sides to keep the 3:2 aspect ratio, because there's more stretching on the longer axis of the image, resulting in an EOS R output that's 305 pixels (4.5%) wider than the native 6720 pixels, and those extra pixels cropped away by Canon/Adobe (but optionally not by DxO) to keep the 3:2 ratio and native output dimensions.

All of those details (and more) are included in my tests of the 14-35 at the wide end:

RF 14-35mm f/4L IS – Distortion Correction Testing

Mirrorless cameras lack an optical viewfinder, so looking through the lens is always an electronic process. This fact enables lens manufacturers to include consideration of digital optical corrections in their lens designs and make trade-offs accordingly – more optical aberrations can be...

www.canonrumors.com

Simple explanation for the cropping in the article you link. "Supreme Court of Ohio – Distortion Corrected and perspective corrected," then, "The uncorrected image is also a fair bit wider than 14mm, with the field of view reduced a little when correcting the distortion, however, natively it’s not quite as dramatic as shown in the second image, as this also was perspective corrected for keystoning, further reducing the angle of view."

Unlike the site you link, in my tests I ensured that the camera was level and orthogonal to the building to avoid vertical and horizontal distortions. Perspective correction can result in substantial cropping, depending on the amount applied. Here's a rather extreme example of that, taken with the RF 14-35/4 @ 14mm as I walked past the Parthenon in Rome.

Perspective corrected shot:


Original shot:


Perspective correction step (force-rectangle in DxO), showing how much cropping is needed after the keystoning is corrected:


One more point regarding focal lengths. Even though the 14-35/4 gives an uncorrected FoV equivalent to ~13mm, the actual focal length is still 14mm (slightly longer, since Canon always rounds in their favor). Consider that the diagonal AoV of the RF 15-35/2.8 at the 15mm is 110°, while the diagonal AoV of the EF 15/2.8 fisheye is 180°. Both are 15mm, but the first is nearly rectilinear (distortion mostly corrected optically), while the fisheye isn’t corrected (circular projection, massive ‘black corners’). Think of the RF lenses requiring distortion correction as one step down the path from rectilinear to fisheye.

 

[Bonich]

Great tip! I downloaded the app and will test it this evening (it's 7.00 am now in Germany).

You can do these statistics in Lightroom simply via applying Metadata Filters.

 

[Amingo]

This new lens is a great opportunity for RF customers. I think the light weight, small size and sharp image Quality looks like compelling reasons to purchase this lens. However, for my needs, the current EF 11-24mm perfectly fullfills my current shooting needs. I aslo get to drop in rear mounted filters which is a serious advantage with top performing EF glass. Looking at some reviews and examples, the lack of front mounting filters will sevearly limit it's applicable use in genres like low light landscape work. If canon were to introduce internal filters into their bodies (CP and ND options) this would change things substantially.
the other issue that I've seen is the Chronic vignetting in the corners. It's beyond thee stops when shooting wide open. Sure, this can be corrected via plugins / software but 3 stops will eat into your corner iso noice lattitude. There's a fair argument that a photographer will stop down substantially for depth of field so it's a mute argument. But as it stands, wide open, this lens has heavy vignetting in the corners. I guess that's the trade off for it's relative size vs the EF version.
It's also very expensive, but then so is all of the RF glass. We know they are great, but eye wateringly expensive. I also suspect that Canon have been quite dellberate with their long end focal length choice for this particular lens. Cutting it to only 20mm really does limit it's versatility in my opinion. I can't imagine anyone choosing this lens as their one and only ultrawide lens of choice. My EF 11-24mm F4 L is my UWA lens of choice, (I keep an old EF 16-35IIL for wet work, near salt water coast lines). I can't imagine anyone doing the same for this lens. So it's quite clear to me that Canon is exepcting us to buy this lens in addition to a 15-35 type of lens. I can't imagine anyone choosing this lens to be mated to a 24-105 and a 100-400 as an F4 trinity of lenses. I think this niche versatility will relegate this lens to a few photographers who need the heavy rectilinear nature and it's probable architectural applications. Estate Agents will love this lens, but I don't think I'll be seeing it around on holiday in Portugal.

I agree but this lens would be a lot bigger.
Canon seems to really emphasize the size and weight advantage of this lens.
Gel filters are a pain though.

Why? Honestly curious.

Consider how difficult it will be to filter the RF 10-20mm. Gels are annoying at best, and there's no CPL option for them. 186mm filters are a PITA to carry around. For lenses like the 11-24 and the TS-E 17, the drop-in filters are a much more convenient option.

I guess 'at some point' may be far in the future. No MP-E 65 and no TS-E lenses for RF, yet.

But if one uses only 'mainstream' lenses (trinity-type zooms, etc.), then completely switching to RF makes sense even today, affordability notwithstanding.

Guys I think you missed that this lens does in fact have a Rear Filter Holder and you won't have to use front mounted filters at all. Canon Rumors didn't mention it in their post but plenty of other sources do.

 

[EOS 4 Life]

Guys I think you missed that this lens does in fact have a Rear Filter Holder and you won't have to use front mounted filters at all. Canon Rumors didn't mention it in their post but plenty of other sources do.

It has a gel filter holder.
I believe GMCPhotographic was referring to drop-in filters like the big whites tend to have.
Drop-in filters can be changed without removing the lens.

 

[neuroanatomist]

Guys I think you missed that this lens does in fact have a Rear Filter Holder and you won't have to use front mounted filters at all. Canon Rumors didn't mention it in their post but plenty of other sources do.

Speaking for myself, I didn’t miss it. It’s already been discussed in this thread. I presume you’re aware that the ‘rear filter holder’ is a gel filter slot on the mount? Useful for ND filters (well, sort of useful, they’re fiddly though perhaps you’ve never used them – I have, with my EF 11-24/4L). There are IR gels if you want that.

But if you want a CPL, a grad ND, or some combo of those with ND, that ‘rear filter holder’ is useless. Did you, perhaps, miss that point?

Oh, wait…I said that already. You even quoted it.

Gels are annoying at best, and there's no CPL option for them.

 

[EOS 4 Life]

Thanks, But why is it only working with the R5? Is the digital IS at work in this case?

I read somewhere that firmware updates will be coming to other cameras with IBIS eventually.

 

[Amingo]

Speaking for myself, I didn’t miss it. It’s already been discussed in this thread. I presume you’re aware that the ‘rear filter holder’ is a gel filter slot on the mount? Useful for ND filters (well, sort of useful, they’re fiddly though perhaps you’ve never used them – I have, with my EF 11-24/4L). There are IR gels if you want that.

But if you want a CPL, a grad ND, or some combo of those with ND, that ‘rear filter holder’ is useless. Did you, perhaps, miss that point?

Oh, wait…I said that already. You even quoted it.

I've seen a couple of posts from you before and expected that your response would be extremely passive aggressive, just like that. I wasn't even trying to be offensive or anything.

 

[neuroanatomist]

I've seen a couple of posts from you before and expected that your response would be extremely passive aggressive, just like that. I wasn't even trying to be offensive or anything.

I am always happy to respond reasonably to reasonable posts, and to help people with reasonable questions. I simply don’t suffer fools lightly, and someone who posts that several people are missing the point when that point was already discussed extensively in the same thread (meaning that someone couldn’t be bothered to read the thread before weighing in) falls into that category.

 

[Exploreshootshare]

You can do these statistics in Lightroom simply via applying Metadata Filters.

Thx for the hint. I noticed that a while back but I don't my pics in the LR library. After I edit the pics, I take most of them out of Lightroom.