Canon RF 28-70 2.8 and optical limits review.

alphaZ

Veteran Member
Messages
5,111
Solutions
1
Reaction score
3,967
Location
UK, UK
I never really tested the RF 28-70 2.8 properly but recently decided to try the Fuji system again for travel and use the R5ii for just wildlife/action and when I needed the extra performance. The X-T5 is a cracking compact camera but trying to find a decent zoom lens for it has proven very challenging. I bought the X-T5 with the 16-50 a tiny compact 24-75 4-7.2 equivalent and its pretty good but it can be slow so set about trying some other options, first of all the Tamron 17-70 2.8, a fairly compact 25-105 f4 equivalent, but it had a lot of issues, mis-focus, ca was high in a lot of situations and I didn't bond with the lens and sold it. Next up was the new Fuji 16-55II 2.8 as expensive as the RF 28-70 2.8 and the first copy was lousy, I returned it for a refund and put it down to bad luck. Buying Fuji gear, especially some of the later lenses you have to join a que and wait, so had to wait weeks and weeks for another copy of the lens. It arrived too late for me to take it away on holiday and on my return I have been testing this copy.

Suffice to say, I am left totally confused by optical limits review of the 28-70 2.8 and what Fuji have released as their greatest 2.8 constant zoom lens, which after all is actually F4.

I set this little test up, which evolved after I was doing some ad-hoc tests around the house with the first lens and these 3 gin bottles became a good place to start and so I set them up again when this second copy of the Fuji 16-55ii arrived and embellished it a little. Again I tested it against the RF 28-70 2.8, which is leaving me a little puzzled as this lens does not seem anywhere near what optical limits is saying it delivers at 70mm and I'm actually thinking of cancelling my patreon to them as I'm not sure I trust their reviews if this is anything to go by. What do others think?

Tripod, es, ibis/ois off, af single on the central body of the cental bottle, smallest af point. 10 sec delay
Tripod, es, ibis/ois off, af single on the central body of the cental bottle, smallest af point. 10 sec delay

Same set-up as above.
Same set-up as above.

Note the coins at the extreme far sides are forward of the everything else and certain things behind the main plain of focus too. This was the second copy of this Fuji lens and the RF 28-70 2.8 has out-performed both of them imo, which has surprised me as I was expecting the extra stop of dof on aps-c and a lens without ois too to be particularly sharp across the field. It also raises the question about optical limits review imo as this lens at 70mm does seems very good across the frame wide open, which is quite impressive for a non L zoom.
 
I've read and watched conflicting information about this lens's performance.

Sometimes it is shown/described as being weaker at 28mm in the corners of the image and being great at 70mm.

The Digital Picture sample images support this claim (link to the comparison between 28 and 70). The corners look a lot better at 70mm than they do at 28mm.

Other times it's the exact opposite: great corners at 28mm and weak in the corners at 70mm.

The video review by Christopher Frost shows this behaviour (link to the video review), with the corners looking much better at 28mm than at 70mm.

So which is it? It's a weird situation.
 
Last edited:
TL;DR: True, but It's by design and it's the right design choice/s optically for the nature of the lens.

Their review is accurate. Lens corners are soft wide open at 70mm, but that doesn't matter as most shots @70mm and @2.8 are portraiture anyways where soft corners are fine. Can't say it's ever bothered me / affected how I shoot or created any problems. Typically, you have a lot of bokeh already at that focal length and aperture, so corner performance becomes not applicable in real life, and really just a synthetic benchmark only (no offense). When I shoot 70mm landscapes? On either of my 28-70s (I have it's big brother the 28-70 f/2L)... You'd be wise to shoot f/4 or tighter, where both clean up and have a larger DoF. Not to pick on my L for a minute, it may be sharper at 70mm and f/2.8, but it's weak spot (if it has one), is also long and wide open... Keep in mind, the largest part of the glass, is going to be for the longest focal, and largest aperture on both lenses in this case as it's not an ultra wide where the largest element would be the shortest focal length with the largest aperture. Naturally, the most compromise on a lens trying to be smaller, lighter, is going to reflect that. Now for me? I shoot at 28mm more than I do 70mm, by a lot I might add. I am very happy with the performance of both lenses btw and I think Canon has designed both (RF 28-70s) well for what they are.

It does rely heavily on correctional data, sure, but that doesn't actually impact the metrics aside from the above... In other words, it's a top performer with that single caveat, which is a small feat for the size, price and capability of the lens.

I agree with OL on the price though. It's still commanding top dollar, even though it's a fantastic offering. I am going to disagree with OL... No lens hood is included at the price point, but, I never shoot this lens with a hood, anyways, as the point of the lens is small and compact but powerful. Lens hoods encroach on that thinking / workflow. It is true though that Canon doesn't include lens hoods on non-L, and this is a quasi-L lens which I do think is silly, especially for the price point...

FYI, this lens gets a lot of use in my arsenal.

I'm going to suggest there is too much emphasis here on synthetic benchmarks data vs practical use. I typically use my standard zooms a lot on the wide end, so performance at the 24-28mm matter more than 70mm. Likewise, I typically use my telephoto zooms a lot on the long end, so performance at 200mm or 400mm matters more than 70mm or 100mm on those. And for landscapes? It depends what's mounted, what I have on me and how I'm feeling.
 
Last edited:
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.

The issue here, between this top of the range 16-55 aps-c 2.8 is the rendering, I'm surprised nobody else has picked up on it. Everyone is focused on sharpness but looks at the plain of focus, how can a 2.8 FF have perceived greater dof within the closest focus elements, clearly have shallower bokeh at the same time. The Fuji lens is clearly tilted or curved in some way, which I didn't believe was a thing at Tele photo but maybe it is. Currently, I'm thinking this RF 28-70 is outstanding!
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range. But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens and that, setting aside the price, means that it misses the point of APS-C outside the telephoto ranges. Just as the EF-S 17-55mm IS did. (That's my rant done with.)
 
Last edited:
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range.
But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens
exactly, defeats the purpose of small apsc in the normal ranges
and that, setting aside the price, means that it misses the point of APS-C
incredible that canon developed a lighter FF lens that has more range and is stabilized
outside the telephoto ranges.
yes for reach, wildlife, etc
Just as the EF-S 17-55mm IS did. (That's my rant done with.)
I never bought that one either
 
I've read and watched conflicting information about this lens's performance.

Sometimes it is shown/described as being weaker at 28mm in the corners of the image and being great at 70mm.

The Digital Picture sample images support this claim (link to the comparison between 28 and 70). The corners look a lot better at 70mm than they do at 28mm.

Other times it's the exact opposite: great corners at 28mm and weak in the corners at 70mm.

The video review by Christopher Frost shows this behaviour (link to the video review), with the corners looking much better at 28mm than at 70mm.

So which is it? It's a weird situation.
I took a series of landscape images and didn't really find anything untoward with the lens. The sharpness is excellent throughout. My issue is the short range, which is why might have to consider using/buying a 24-105 or using the 24-240 more.

What I found fascinating is the rendering between these two lenses. With this being the second lens from Fuji it seems very similar, maybe it's a bit sharper but the rendering imo is weird and not flat, some kind of curvature is going on, I can't believe two lenses would be tilted!
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range.

But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens
exactly, defeats the purpose of small apsc in the normal ranges
and that, setting aside the price, means that it misses the point of APS-C
incredible that canon developed a lighter FF lens that has more range and is stabilized
outside the telephoto ranges.
yes for reach, wildlife, etc
Just as the EF-S 17-55mm IS did. (That's my rant done with.)
I never bought that one either
My honest opinion is that the RF 28-70 is an incredibly good lens considering it's size, weight and speciation but it's range, starting at 28mm can be a bit limited and will require lens swapping, which I don't find convenient a lot of the time. So, it's either 24-105 or 24-240 for me unless Canon can introduce a new and interesting zoom offering.
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range. But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens and that, setting aside the price, means that it misses the point of APS-C
While true, it can be pretty redundant to have both an aps-c and a full frame camera. If you have both compact telephoto AND low light indoor needs a lens like this allows you to go with one aps-c camera only, in stead of adding a full frame camera for low light indoor purposes. The good thing of ILCS: nothing stops you from leaving the indoors low light lens at home when you want to travel light.
outside the telephoto ranges. Just as the EF-S 17-55mm IS did. (That's my rant done with.)
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range.

But that Sigma is a bigger, heavier unstabilised
IBIS is the answer. Outside the big R7 not Canons answer, but that's not Sigmas fault. In the range up to 70mm the IBIS of my A7RV is as effective as ILIS, and in stead of lagging behind and leaving out IBIS of anything but flag ship models Canon has some homework to do here, not Sigma.
lens than the equivalent stabilised Canon full-frame lens
exactly, defeats the purpose of small apsc in the normal ranges
and that, setting aside the price, means that it misses the point of APS-C
incredible that canon developed a lighter FF lens that has more range and is stabilized
The Sigma being bigger than effective is partly due to Canons choice to come up with a smaller crop sensor size than the Sigma lens is capable to feed with light.
outside the telephoto ranges.
yes for reach, wildlife, etc
Just as the EF-S 17-55mm IS did. (That's my rant done with.)
I never bought that one either
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range.

But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens
exactly, defeats the purpose of small apsc in the normal ranges
and that, setting aside the price, means that it misses the point of APS-C
incredible that canon developed a lighter FF lens that has more range and is stabilized
outside the telephoto ranges.
yes for reach, wildlife, etc
Just as the EF-S 17-55mm IS did. (That's my rant done with.)
I never bought that one either
My honest opinion is that the RF 28-70 is an incredibly good lens considering it's size, weight and speciation but it's range, starting at 28mm can be a bit limited and will require lens swapping, which I don't find convenient a lot of the time. So, it's either 24-105 or 24-240 for me unless Canon can introduce a new and interesting zoom offering.
starting at 28 mm is issue for me also - I have the RF 24 -105 F4 L IS on R8
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range. But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens and that, setting aside the price, means that it misses the point of APS-C
While true, it can be pretty redundant to have both an aps-c and a full frame camera. If you have both compact telephoto AND low light indoor needs a lens like this allows you to go with one aps-c camera only, in stead of adding a full frame camera for low light indoor purposes. The good thing of ILCS: nothing stops you from leaving the indoors low light lens at home when you want to travel light.
Perhaps, but I've already got the full-frame cameras & lenses, and neither the 17-40mm not the 28-70mm lenses will fit my EOS M cameras which I keep for portability.
outside the telephoto ranges. Just as the EF-S 17-55mm IS did. (That's my rant done with.)
--
R5 & RV
EF & FE
 
I've read and watched conflicting information about this lens's performance.

Sometimes it is shown/described as being weaker at 28mm in the corners of the image and being great at 70mm.

The Digital Picture sample images support this claim (link to the comparison between 28 and 70). The corners look a lot better at 70mm than they do at 28mm.

Other times it's the exact opposite: great corners at 28mm and weak in the corners at 70mm.

The video review by Christopher Frost shows this behaviour (link to the video review), with the corners looking much better at 28mm than at 70mm.

So which is it? It's a weird situation.
Both my RF28-70 f/2.8 IS STM and RF28-70 f/2L USM, are sharper at 28mm, than they are at 70mm, and especially wide open.

There is copy variance, sure, but this one is straightforward. The wide end is crispy, the long end, it's crisp only the in center if shooting f/2.8, and that's a good compromise for the reduction in size and weight, in my book for being an f/2.8 lens.

I'll go one further: the RF28-70 f/2.8 IS STM is in fact very sharp, for what it is.

Now if you want to pick on this lens? The bokeh isn't bad, but I prefer the bokeh of my RF 28-70 f/2L USM in terms of bokeh quality. And, I'll say the lack of nano-coating, it does impact shooting in hostile light. There's a reason for the L, no compromises. But, there's a reason for the RF 28-70 (not-L), some compromise in exchange for reduced size and weight (and price). It's a good compromise I might add. I make the compromise often and leave the L's at home when I feel like giving myself a break.

I look forward to seeing Canon complete the unholy f/2.8 not L trio with some form of RF70-200 f/2.8 like offering as I enjoy my RF100-400, but there's times I'd like more than f/5.6 on the long end, but not have to bring a brick along.
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range. But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens and that, setting aside the price, means that it misses the point of APS-C
While true, it can be pretty redundant to have both an aps-c and a full frame camera. If you have both compact telephoto AND low light indoor needs a lens like this allows you to go with one aps-c camera only, in stead of adding a full frame camera for low light indoor purposes. The good thing of ILCS: nothing stops you from leaving the indoors low light lens at home when you want to travel light.
Perhaps, but I've already got the full-frame cameras & lenses, and neither the 17-40mm not the 28-70mm lenses will fit my EOS M cameras which I keep for portability.
I'm currently sorting out some pics I made with my M50, and looking at the results I'm just so happy what the old Sigma EF 18-35mm f/1.8 did for me. Eye AF, f/2.9 equivalent in a zoom, made before the first Canon full frame mirrorless camera was released.
outside the telephoto ranges. Just as the EF-S 17-55mm IS did. (That's my rant done with.)
 
you should be comparing the new siggy 17-40 f1.8 on your fuji to the RF 28-70 f2.8 IS on canon
The Soggy is too short range, it doesn't really move things forward for me.
It's 28-65mm f/2.9 FF equivalent on a Canon so I wouldn't miss the 5mm (<10% difference) in the (for me) not very useful 50-90mm range. But that Sigma is a bigger, heavier unstabilised lens than the equivalent stabilised Canon full-frame lens and that, setting aside the price, means that it misses the point of APS-C
While true, it can be pretty redundant to have both an aps-c and a full frame camera. If you have both compact telephoto AND low light indoor needs a lens like this allows you to go with one aps-c camera only, in stead of adding a full frame camera for low light indoor purposes. The good thing of ILCS: nothing stops you from leaving the indoors low light lens at home when you want to travel light.
Perhaps, but I've already got the full-frame cameras & lenses, and neither the 17-40mm not the 28-70mm lenses will fit my EOS M cameras which I keep for portability.
I'm currently sorting out some pics I made with my M50, and looking at the results I'm just so happy what the old Sigma EF 18-35mm f/1.8 did for me. Eye AF, f/2.9 equivalent in a zoom, made before the first Canon full frame mirrorless camera was released.
Yes but with the adapter that lens was 2⅓× the weight of the R50 body and nearly twice the weight of the full-frame R8 body. Quite ridiculous for an APS-C standard zoom (28-57mm equivalent). It's like hitching a caravan to the back of a Smart car. (Though I'm pleased with results I got from mounting a TS-E 24mm f/3.5L II on an Extender EF 1.4X III on my M100, which was nearly as bad so I can't really talk.)
outside the telephoto ranges. Just as the EF-S 17-55mm IS did. (That's my rant done with.)
--
R5 & RV
EF & FE
 
Interesting test.

Are these SOOC JPEGs or processed from RAW.

I think what you are seeing on Fuji is a lens that is not good enough for the Fuji's 40mp pixel pitch. Maybe try DXO and see if it gets any better but my guess is it won't because your test is exhibiting weakness in the central region of both lenses.

OR, the other possibility is that the JPEG processing is being heavy handed with the contrast. The lack of contrast on Fuji image is what I am seeing stand out. Could be the lens itself or the way JPEG was rendered. If however these are both JPEGs produced by the same software then it is quite likely the lens.

R5 II is also one hell of a camera. More pixels than X-T5, and not only that .... more bigger pixels. And the metering is more segmented on R5 II. And I am pretty sure Canon's engineers had all this in mind when working on this lens.

R5 II - Pixel Pitch 4.39um and 45mp

X-T5 - Pixel Pitch 3.04um and ~40mp

Quite possibly R5 II + the lens is blowing the smaller sensor (and smaller pixels and weaker signal capture due to the both) out of the water.

I have the RF 28-70 f/2.8 and in general I align with Optical Limits' findings (and also same from Christopher Frost's testing) but even with the known "shortcomings" it is still a great lens. At 70mm I know from use that bokeh is nervous but most people will only see it if they know what they are looking for. Clinical testing only shows what the testing was made to show.

I think Optical Limits gave it 3.5 stars due to the price. It is not cheap but in my use it is well worth it.

And while we are at it even the 16-28 f/2.8 comparisons with its peers gives me no reason to buy any other UWA zoom when I need one. Also eagerly waiting for a 70-100+ f/2.8 zoom to complete the focal range.

--C
 
Something many reviews ignore is the distance of the test shots. Field curvature can change a lot relative to focal distance, it usually appears worse closer to the camera. Of course depth of field increases at further distances too. This alone may increase the impression of more sharpness near the edges as there will be a lot less falloff. Usually what review sites call "softness" is not strictly lens softness but field curvature. Since the world is 3 dimensional it tends to average out for most shooting.
 

Keyboard shortcuts

Back
Top