Canon RF-S 11-22mm f/4.5-5.6 IS STM coming first half of 2023

OskarB said:

Is this the same for the 105?

Click to expand...

With 105mm being a shorter focal length, I expect it to behave much better than the 150mm. But I don’t have direct experience with the 105, only the 150.

entoman said:

The RF100-400mm isn't weather-sealed, and the build quality is lower, so it's less durable. Optically it's more than good enough and the bokeh is really nice IMO. It focuses closer than the RF100-500mm, but it isn't supplied with a tripod collar, lens hood or case. But it's a fraction of the cost of the RF100-500mm.

For safaris or bird photography from hides, weight and portability aren't significant factors, so I use the RF100-500mm on my R5, which gives me about an extra stop of light at most focal lengths - and that's valuable for wildlife in winter light or at dusk and dawn when most animals are active.

For insects and other small wildlife, the RF100-400mm is for me a much better choice, as it's so easy and light to carry, focuses closer, and is much easier to manoeuvre.

Luckily I have both lenses, but if I had to choose between one or the other, I'd probably get the RF100-400mm and poke it on a R7, which will give more reach than the RF100-500mm will on FF.

Click to expand...

We agree 100% on this, apart from the last point. In theory, a 32 Mpx APS-C gives 36% more resolution than a 45 Mpx FF sensor, and a 500mm lens gives only 25% more resolution, so on paper you are right as it looks like 400mm on an R7 should give 8.8% more reach than a 500mm on an R5. In practice, however, my measurements show it's the other way round that the 100-500mm on the R5 gives about 10% more reach at base iso and it increases with increasing iso. The following reasons come into it. The R5 sensor has a better AA-filter and its larger pixels are slightly less affected by diffraction so the resolution advantage of the R7 is only about 20-25% more at base iso, and the 100-500 is slightly sharper and has a small advantage for diffraction. At higher isos, the smaller pixels are affected more by noise and the R7 progressively loses its reach advantage. It's for the this reason, I'll be taking the 100-500mm with me on our next birding safari, but for insects I would go for the 100-400 too.

Dragon said:

The linear resolution is related to the square root of the pixel count, so the R7 has 1.16 x more linear resolution than the A6400 or 1.16x400 = 465mm equiv on the A6400 for the same pixel density on the subject. I didn't see any sign of motion issues in your shots. If you have a tracking mount and an R7, just buy the 800 f/11 rather than spending all that rental money. It takes awesome pictures and supports a 1.4 TC well even on an R7. I bought the 800L used before there was a mention of anything "R" and it has served me well, but it still cost more than an R5 plus an R7 plus an 800 f/11 . I probably won't sell it but given how well the f/11 lens works, I wouldn't buy the L today. It has landed me some very cool shots and it supports a 2x TC well, albeit best at f/16. Here is a shot I took with the 800L and the 2x converter at f/16, ISO 1600, and 1/125th. This was during the recent near flyby of Jupiter, which was very bright. This was on a tripod with delayed shutter and the IS turned off.

The second shot is with the R7, the RF800 f/11 with 1.4 TC at f/16, ISO 6400, and 1/400th handheld, but note that the stabilizer is good down to 1/30th if you have a reasonably steady hand and the subject is not moving. That hummingbird was about 35 feet away and the attached image is downscaled almost 2:1 to give you an idea of the magnification.


View attachment 207036 View attachment 207037

Click to expand...

That Jupiter shot is incredible! Not just the bands and moons, but all of the background stars too. Nice job!

You're right about the pixel density, it's linear pixels/mm, not area in total MP.

But the multiplier is even larger because Canon's APS-C is smaller than everybody else's. The R7 sensor is 22.3 x 14.8 mm, whereas the Sony a6400 is 23.5 x 15.6 mm. So when you look at the actual pixels per mm, the density multiplier for Canon is more like 1.22.

Being a math nerd, I put together a spreadsheet to compare the two. I used as an example two of my lunar eclipse lenses, the Sony 200-600 at 600 mm on the a6400, and the Sigma 100-400 at 400 on the R7.

In order to get the same pixel density with the R7 as with the Sony at 600 mm, I only need 491 mm. Similarly, if I used the Sigma at 400 on the Canon, it would be equivalent to 489 mm on the Sony.

mdcmdcmdc said:

That Jupiter shot is incredible! Not just the bands and moons, but all of the background stars too. Nice job!

You're right about the pixel density, it's linear pixels/mm, not area in total MP.

But the multiplier is even larger because Canon's APS-C is smaller than everybody else's. The R7 sensor is 22.3 x 14.8 mm, whereas the Sony a6400 is 23.5 x 15.6 mm. So when you look at the actual pixels per mm, the density multiplier for Canon is more like 1.22.

Being a math nerd, I put together a spreadsheet to compare the two. I used as an example two of my lunar eclipse lenses, the Sony 200-600 at 600 mm on the a6400, and the Sigma 100-400 at 400 on the R7.

In order to get the same pixel density with the R7 as with the Sony at 600 mm, I only need 491 mm. Similarly, if I used the Sigma at 400 on the Canon, it would be equivalent to 489 mm on the Sony.

View attachment 207048

Click to expand...

Correct. I spaced the 1.5/1.6 crop factor difference between Sony and Canon in my response, and I know better. My only excuse is that it was late . In any case, 600mm still wins the reach game. and 800 is even better. Here is a 100% crop of a shot on a very dark day with the R7 and the 800 f/11 bare (no TC). ISO 6400 and 1/80th second handheld. As you see, it doesn't want for detail. Processed through DXO Photolab 6.1.1 and final tweaked in LR. It is amazing how the hummer feathers light up when they are looking at you even in crappy lighting. Actually, dim gray lighting works best because if the sun is behind you, the contrast goes up to about 20 stops and you can't recover the detail in the feathers. And rember, this bird is only about 2-1/2 inches long from tip to tip.

Dragon said:

Correct. I spaced the 1.5/1.6 crop factor difference between Sony and Canon in my response, and I know better. My only excuse is that it was late . In any case, 600mm still wins the reach game. and 800 is even better. Here is a 100% crop of a shot on a very dark day with the R7 and the 800 f/11 bare (no TC). ISO 6400 and 1/80th second handheld. As you see, it doesn't want for detail. Processed through DXO Photolab 6.1.1 and final tweaked in LR. It is amazing how the hummer feathers light up when they are looking at you even in crappy lighting. Actually, dim gray lighting works best because if the sun is behind you, the contrast goes up to about 20 stops and you can't recover the detail in the feathers. And rember, this bird is only about 2-1/2 inches long from tip to tip.

View attachment 207049

Click to expand...

I know I focused (no pun intended) on the Jupiter shot earlier, but your bird shots are terrific too.

My only concern with buying the 800mm f/11 is that I wouldn't use it that often. I'm not a birder as such; the "birds" I like to shoot usually have metal wings and propellers, and even 400mm on the R7 is more than enough most of the time (there's always humidity and haze in the air where I live, so I've learned that if I have to zoom all the way to 400, it's probably not going to be a great shot anyway).

But there's always the moon and planets, and there are two solar eclipses coming up in the next 15 months, so you've got me thinking about it!

Edit: Somebody on my local Craigslist has an FD 800mm f/5.6L listed for $1900. Since I’ll probably only use it in MF anyway, I’m tempted, but the price is too much. If I were serious I would try to talk him down to maybe $1200, but I’m not sure I’m that serious yet.

11-22 is so random. Then again I just do photos. I bet that is super helpful for those who do video. 28 has been more than enough for my weddings and etc.
-Cody M
Charlotte Wedding Photographer

koenkooi said:

The OS is junk on the 150 and the IBIS on the R5 actively fights it. So the non-OS 150 would likely give you better stabe on an IBIS body. Turning off OS on my 150 also turns off IBIS

Click to expand...

For what it is worth, I just did a comparison test with my Sigma 150 OS. On the 5DSr, the OS has about a 90% hit rate at 2 stops and about 60% at 3 stops. That is in the ballpark for the original specification of the lens, which, as I remember, only claimed about 2 stops. I then put the lens on the R5 and the results were slightly worse with several random weird results, so the theory that the lens doesn't play well the IBIS may be correct or it may simply not be fully compatible with the R5. I then put the lens on the R7 and the results were slightly better than with the 5DSr (actually quite a bit better when you take the crop factor into account), so the lens does appear to be compatible with the R7. These kinds of issues are common with third party lenses and later released Canon bodies. For lenses that are still in production, Both Sigma and Tamron do try pretty diligently to provide firmware updates, but with discontinued lenses, you are on your own and some bodies never get supported. As to getting better performance with the non-OS lens and IBIS, I doubt that. IBIS works wonders on short lenses but falls off very fast over about 70mm. The sensor simply cannot move far enough to make the necessary corrections. In any case, all three Sigma OS macros (105, 150, and 180) are much older lenses and the OS is pretty primitive compared to modern lenses. There is also, no claim from anyone, least of all Canon, that the IBIS will be cooperative with 3rd party EF lenses. The good news is that, at least in my limited testing, the 150 seems to be happy on the R7. With the R5, the best solution may well be to simply turn off the OS (and thus, also the IBIS). I did run a test on the 150 with the R5 a while back to see if the lens worked with focus stacking and in limited testing, it seemed to work correctly. For focus stacking, you should have the camera on tripod with the IS/OS turned off, no matter what lens you are using, so depending on your application, the OS compatibility may or may not be a major issue. I have both the 150 and the 180 and optically they are both stellar with the 180 being slightly sharper in the center wide open.

mdcmdcmdc said:

I know I focused (no pun intended) on the Jupiter shot earlier, but your bird shots are terrific too.

My only concern with buying the 800mm f/11 is that I wouldn't use it that often. I'm not a birder as such; the "birds" I like to shoot usually have metal wings and propellers, and even 400mm on the R7 is more than enough most of the time (there's always humidity and haze in the air where I live, so I've learned that if I have to zoom all the way to 400, it's probably not going to be a great shot anyway).

But there's always the moon and planets, and there are two solar eclipses coming up in the next 15 months, so you've got me thinking about it!

Edit: Somebody on my local Craigslist has an FD 800mm f/5.6L listed for $1900. Since I’ll probably only use it in MF anyway, I’m tempted, but the price is too much. If I were serious I would try to talk him down to maybe $1200, but I’m not sure I’m that serious yet.

Click to expand...

The FD 800L is even heavier than my EF 800L and I don't think it will be any sharper than the RF800 f/11 and maybe just about match it after you stop the FD down to f/11 . Trust me, using an 800mm without stabilization is very challenging. You need a really good tripod and NO wind to get decent results. One of the big features on the EF 800L is that the IS is tripod sensing and continues to work correctly on a tripod. I haven't tried the RF 800 f/11 on a tripod to see if it has that feature, but I have seen no claim that it does. OTOH, it works so well handheld that I have never seen the need for a tripod. For astro, you would turn off the IS in most cases, but for terrestrial shots, it is awesome. I did look at an FD 800L before I bought the EF 800, but that was before Canon mirrorless and FD to EF has the same -2mm flange distance problem that M to R has so I passed. Also good to remember that there is no sensible way to get one of the big FD lenses repaired as all repair parts are long gone at Canon. That makes the investment less attractive.

codym90 said:

11-22 is so random. Then again I just do photos. I bet that is super helpful for those who do video. 28 has been more than enough for my weddings and etc.
-Cody M
Charlotte Wedding Photographer

Click to expand...

The M series 11-22 has been around for quite a few years and it is a fine little lens. 11-22 crop is 17.6-35.2 FF equivalent, so not really random, but rather a very decent ultra-wide range. Beyond video, it is very useful for landscape, architectural, and real estate photography.

mdcmdcmdc said:

That Jupiter shot is incredible! Not just the bands and moons, but all of the background stars too. Nice job!

You're right about the pixel density, it's linear pixels/mm, not area in total MP.

But the multiplier is even larger because Canon's APS-C is smaller than everybody else's. The R7 sensor is 22.3 x 14.8 mm, whereas the Sony a6400 is 23.5 x 15.6 mm. So when you look at the actual pixels per mm, the density multiplier for Canon is more like 1.22.

Being a math nerd, I put together a spreadsheet to compare the two. I used as an example two of my lunar eclipse lenses, the Sony 200-600 at 600 mm on the a6400, and the Sigma 100-400 at 400 on the R7.

In order to get the same pixel density with the R7 as with the Sony at 600 mm, I only need 491 mm. Similarly, if I used the Sigma at 400 on the Canon, it would be equivalent to 489 mm on the Sony.

View attachment 207048

Click to expand...

It's easier for me to work the relative resolutions of sensors using the pixel pitch directly, which is basically doing the same as you. For example, the length of the A6400 pixel is 3.89µ, R7 3.2µ, R5 4.38µ, R10 3.7µ, R3 5.99µ etc so the A6400 needs a lens 3.89/3.2 times longer etc to get the same linear resolution. But, this ignores the effects of diffraction, noise, AA-filter etc. In bright light with a sharp f/2.8 lens, the relative pixel sizes will approximate to the real relative resolutions. But, at high iso and a narrow lens the resolution advantages of the smaller pixels become negated.

mdcmdcmdc said:

That Jupiter shot is incredible! Not just the bands and moons, but all of the background stars too. Nice job!

Click to expand...

It's a nice shot but if you look closely it's not a single exposure, the planet has a halo that shows it was pasted into the background, that's the only way to do this, as you can't expose for both the surface details and the much dimmer starfield.

AlanF said:

My second comment that naysayers are not much fun to read was in response to your comment:

Click to expand...

Sorry for being so tardy, I was away on a 3-week trip...

Fair enough, although my quip wasn't about a specific lens being good or bad, it was about other people assuming that their likes (or dislikes) should be shared by everyone else.

While it may be "fun" to read high praises about a lens, I personally prefer to hear both from people who liked it and people who didn't, since we all have our specific use cases and different tolerances to the compromises each lens has.

neuroanatomist said:

Not sure if you mean overlap in a full kit, or overlap in what gets packed for a specific outing/trip.

Click to expand...

Both, actually Apologies for the long time, I was on a 3-weeks trip with family in Italy. I brought 3 lenses (RF 100-500, RF 85 1.2 and TS-E 24 II) which have all seen some use on my R5, plus my GoPro 10 (not so happy with it) and my DJI Mini 3 Pro (super happy with it)

neuroanatomist said:

For the former, I see no problem with significant overlap. My preference is to have the right tool available for the job. You can pound a fence post in with a little hammer or drive a small picture-hanging nail into your wall with a sledgehammer, but neither is optimal. I currently have 6 RF lenses (counting the one being delivered today by FedEx), 8 EF lenses, 8 EF-M lenses, and four TCs. For FF that focal range spans 11mm to 1200mm (11-24/4, 600/4 + 2xTC), with ample overlap. For ‘fast’ lenses, I have the RF 28-70/2L and EF 85/1.4L IS, and I suppose you could count the RF 70-200/2.8 as well.

That means when I’m going to shoot something, I can pick the lenses best suited to my subject(s). I feel the same about how those lenses are carried. I have an array of cases/backpacks that can hold a camera with pretty much any lens or set of lenses without empty space (unless by design when I want to also carry non-photo items).

The only thing I don’t want in my lens kit are lenses that I don’t use.

In terms of selecting lenses for a particular outing/trip, the same rules apply except with significant trips it’s preferable to have a backup body at least. Previously, I would bring an M body and a few lenses (typically the M11-22, M18-150 and M22/2). I’d use the M set as a convenient daytime walk around kit with family, then take the FF kit out later for solo blue hour / night photography. The EF-M adapter allowed the M body to serve as a backup, although I never actually needed it. Still, that’s a function the M6II cannot serve with my mostly RF kit. It’s why, if the rumored RF-S 11-22 comes out, I’ll consider an R8 or R10 as a second camera for the trip.

Regarding UWA overlap, as I typed the list of lenses for a planned urban European trip this summer (EF 11-24/4, RF 14-35/4, RF 24-105/4, RF 100-400, and TS-E 17), I actually removed the RF 14-35 then added it back. The reason is the weight difference – the 11-24 is a beast, and for the days I won’t need wider than 14mm I’d prefer to carry the lighter lens and leave the other in the hotel. I would bring both while traveling, but not carry both on any given outing.

Click to expand...

It's all good, as mentioned my post was not meant as a critique to your approach. I have a slightly different philosophy I guess. I have the RF 85 1.2 and if I had the RF 85 2 I know for a fact that the f/2 lens would stay in the drawer unused, even if it is significantly smaller and lighter than the 1.2 one. Again, this is just me.

neuroanatomist said:

As for lenses being f/4 or slower, I’ve found that not to be a problem for my travels. I use fast lenses (mostly the RF 28-70/2 and RF 70-200/2.8 now) for portraits and for indoor/night-lit events with moving subjects. Most of my travel shooting is outdoors at narrower apertures, and indoor shots are typically more of interiors (cathedrals, etc.), where I don’t need a fast shutter and thus can benefit from IS. The (up-to) 8-stops of stabilization with lenses like the RF 14-35 and 24-105 is a huge benefit for interior shots in locations where a tripod is not practical or not permitted.

Click to expand...

As I mostly shoot fashion with the Hassy & IQ 180 and active kids with the R5, when I use the R5 I value speed to isolate subjects and to keep fast shutter speeds even in bad light conditions. Different use cases. Of course if I'm traveling I shoot other things as well: I've shot landscapes and cityscapes in Italy with my lenses stopping down as needed... I am aware I could have used lighter lenses for those images, but that won't change my mind.
Moreover, I am probably being an old curmudgeon , but I am not a fan of going high ISO. With the IQ 180 I am really comfortable only at ISO 35 or 50... but that's an old CCD. With the R5 I can go 1600 in a pinch but really I feel 800 is the ceiling for me.

neuroanatomist said:

Given others’ reports, testing and —more importantly— results with the RF 100-400, I have a reasonable expectation the lens will perform well. Did your friend properly test the lens when new, or did you? I ask because QC is not perfect and bad lenses are shipped. I had to return the first copy of my Rokinon 14/2.8, the second performed as expected. All my Canon lenses have been fine, but for example (I know I’m repeating myself from other threads), when I was reviewing the M18-150 for TDP, my results were much better than what Bryan showed in his ISO 12233-type charts. I told him that, and he ordered and tested a second copy of the lens and it was much better, those are the results now on his site. While such issues are more common with consumer-grade lenses, they happen with L-series lenses, too. Bryan tested four copies of the EF 24-70/2.8L, the first two were sub-par.

Click to expand...

In short ,no. I just took some shots in the same settings I was using my 100-500 (a wolf reservation) and looked at them side by side at home. It's not that it is a bad unusable lens (no clue if that particular one was sub par or not), but I prefer the output of the 100-500 which, since I do indeed have a 100-500 already, made the decision not to buy a 100-400 a very easy one for me...

neuroanatomist said:

My RF 100-400 arrives today, I’ll properly test it next week, if it’s not up to snuff I’ll exchange it, and if the second copy also fails to perform well I’ll probably just return it and decide whether or not to take the 100-500L on the trip.

Click to expand...

Happy shooting!

scyrene said:

It's a nice shot but if you look closely it's not a single exposure, the planet has a halo that shows it was pasted into the background, that's the only way to do this, as you can't expose for both the surface details and the much dimmer starfield.

Click to expand...

Sorry, but that was a single shot with a lot of shadow pull and noise reduction. The halo is due to dimming the planet with a mask in LR so the starfield could be pulled up without burning out the planet, but it was a single shot albeit heavily tweaked in post.

Dragon said:

Sorry, but that was a single shot with a lot of shadow pull and noise reduction. The halo is due to dimming the planet with a mask in LR so the starfield could be pulled up without burning out the planet, but it was a single shot albeit heavily tweaked in post.

Click to expand...

Ok fair dos, unexpected. I would suggest some (many?) of the 'stars' likely to be noise/hot pixels, in that case. A good illustration of what's possible, anyhow

scyrene said:

Ok fair dos, unexpected. I would suggest some (many?) of the 'stars' likely to be noise/hot pixels, in that case. A good illustration of what's possible, anyhow

Click to expand...

It was exposed to the right to start with, so you are seeing the max dynamic range of an R5. The noise was just below the stars that are showing so I doubt that more than a few are noise. That shot was at ISO 600 and it would have been a bit better if I had had a tracking mount to be able to drop the ISO to 100. It does go to show that a lot of dynamic range is good, and more is better . If Canon can get that photon counting sensor tech up to normal camera resolution, it will be really interesting because more bits in an A/D always translates to longer readouts.

Dragon said:

It was exposed to the right to start with, so you are seeing the max dynamic range of an R5. The noise was just below the stars that are showing so I doubt that more than a few are noise. That shot was at ISO 600 and it would have been a bit better if I had had a tracking mount to be able to drop the ISO to 100. It does go to show that a lot of dynamic range is good, and more is better . If Canon can get that photon counting sensor tech up to normal camera resolution, it will be really interesting because more bits in an A/D always translates to longer readouts.

Click to expand...

Cool! Nothing taxes DR like astro. Do you find ETTR helpful even now? My understanding was since Canon's newer sensors (~5D4 onwards?) it was no longer really necessary, but I haven't tested that assumption tbh.

scyrene said:

Cool! Nothing taxes DR like astro. Do you find ETTR helpful even now? My understanding was since Canon's newer sensors (~5D4 onwards?) it was no longer really necessary, but I haven't tested that assumption tbh.

Click to expand...

I think ETTR is always helpful when you have very high contrast and want to preserve the highlights. With a very small object like a planet, often the easiest approach is to take a series of exposures and then pick the optimum one in Post as I did in this case. The issue is not sensor related, but rather a metering issue. In-camera metering is useless for astro (other than maybe an 800mm moon shot) because the mass of black will cause the metering to overexpose all the shiny things . For terrestrial stuff with objects big enough to meter on, I find spot metering to be quite useful, particularly for birds against both bright and dark backgrounds, but it is very hard to use for BIF unless you have a camera that tracks the AF point with the Spot metering (think 1 series).

AlanF said:

It's easier for me to work the relative resolutions of sensors using the pixel pitch directly, which is basically doing the same as you. For example, the length of the A6400 pixel is 3.89µ, R7 3.2µ, R5 4.38µ, R10 3.7µ, R3 5.99µ etc so the A6400 needs a lens 3.89/3.2 times longer etc to get the same linear resolution. But, this ignores the effects of diffraction, noise, AA-filter etc. In bright light with a sharp f/2.8 lens, the relative pixel sizes will approximate to the real relative resolutions. But, at high iso and a narrow lens the resolution advantages of the smaller pixels become negated.

Click to expand...

True on all counts. In fact, resolution or pixel density aren't true figures of merit because you'll rarely get features in your images whose boundaries align exactly with the pixel boundaries on the sensor.

I wish there was a straightforward way to quantify pixel density because I feel "crop factor" is an anachronism. Twenty years ago, when APS-C DSLRs became popular and most people viewed photos as 4x6" minilab prints, crop factor made some sense. But given that most high-end APS-C shooters do a lot of cropping in post anyway, it's less relevant. But mathematically, crop factor is easy. Just multiply by 1.6 (for Canon). Pixel density is a little more complicated.

In my experience, diffraction is often worried about more than it really needs to be. Yes, it's real, it can be calculated (I've done it), it can be measured, and sharp-eyed observers might be able to spot in an image viewed at 100% on a monitor. But in my experience, using the R7 at the f/8 to f/11 where I normally shoot, I've never felt limited by it. I'm not talking about f/22, but people on various forums who say the R7 can't be used at slower than f/5 in my opinion haven't tried.

I look forward to a day when diffraction limits me with the R7 at f/11, and not my own handling of the camera.

scyrene said:

It's a nice shot but if you look closely it's not a single exposure, the planet has a halo that shows it was pasted into the background, that's the only way to do this, as you can't expose for both the surface details and the much dimmer starfield.

Click to expand...

Dragon said:

Sorry, but that was a single shot with a lot of shadow pull and noise reduction. The halo is due to dimming the planet with a mask in LR so the starfield could be pulled up without burning out the planet, but it was a single shot albeit heavily tweaked in post.

Click to expand...

Even if it wasn't a single exposure, so what? That's what HDR is.

Similar to what @Dragon said, the lunar eclipse image I posted a couple of pages back was a single exposure, but I used a mask in LR to tone down the brightness of the non-totally eclipsed limb. I was really surprised by how much detail became visible on that part. Kudos to the R7 sensor!