Intro

The Canon RF 100mm f2.8L IS USM is a high-end macro lens designed for Canon’s EOS R mirrorless system. While pitched as a macro lens for close-up work, it’s also a highly-corrected short telephoto that’ll deliver great-looking portraits and landscapes too. Like other RF lenses, it’ll work on EOS R bodies with full or cropped-frame sensors, but it’s not compatible with EOS-M bodies or Canon DSLRs.

The latest 100mm Macro comes almost 12 years after the EF version for DSLRs (which can work on EOS R mirrorless cameras using the EF to EOS R adapter), and at first glance they appear to share many specs including the same focal length, aperture, inclusion of optical stabilisation, USM focusing and L-branding. But while the EF model delivered a maximum magnification of 1:1 or 1x, the newer RF version boasts 1.4:1 or 1.4x magnification for greater reproduction, as well as an adjustable spherical aberration control for interesting soft-focus effects. As a native RF lens it should also provide more reliable focusing as well as enhanced stabilisation.

The older EF model has been creeping up in price but remains a little cheaper and it’s still highly regarded, so the big question is whether the new RF version is really that much better and deserving of the higher asking price. To find out I tested both lenses side-by-side on an EOS R5 body and in this review I’ll help you make the right choice. My full report is in the video below, but if you prefer to read a written version, keep scrolling! Oh, and just to reconfirm, the EF Macro lens I tested here was the high-end f2.8L model, so when I simply refer to it as the EF 100, know that I’m talking about the f2.8L version.

Place the old EF 100 2.8L Macro alongside the new RF 100 2.8L Macro and right-away you can see the latter is longer: 148 vs 123mm, not to mention being 4mm wider, and at 685g, it’s 60g heavier too. But if you’re adapting the EF version for use on an EOS R mirrorless camera, the combination weighs a little more and becomes essentially the same length. Overall if you’re adapting the EF lens for mirrorless, there’s little to choose between them in overall size and weight, but in terms of controls, they are quite different.

Starting from the mount side of the RF lens, there’s three switches including a focus limiter with three ranges. Roughly mid-way along the barrel is the first big difference from the older EF model, and that’s a dedicated SA ring to adjust the spherical aberration. A switch on the other side locks it in the neutral position to avoid accidental use, but when unlocked you can smoothly turn it in a scale of +/- 2 steps. While the subject and background rendering look different in each direction, it’s clear to see how you can use this adjustment for a surreal in-camera dreamy effect. It’s all very 1980’s.

Would I use it personally? Probably not, and that’s also coming from someone who previously owned the EF 135mm f2.8 Soft Focus lens, but you might love it.

Returning to the controls are a free-spinning motor-assisted manual focusing ring without physical markings, and finally a clicky customisable RF control ring. At the end of the barrel is a 67mm filter thread, and here’s how the lens looks fitted with its supplied hood.

For comparison, the old EF 100 2.8L Macro has three control switches including a focus limiter with the same three options as the RF version, albeit not quite focusing as close. Meanwhile there’s just one ring on the EF lens, dedicated to manual focusing and like other models of this era, it’s mechanically-linked with a small window indicating the focusing distance. In contrast, RF lenses with their motor-assisted manual focusing, display the distance on-screen. And finally there’s the same 67mm filter thread as the RF version, although interestingly the supplied lens hood is longer.

As members of the L-series, both lenses are sealed against dust and moisture with rubber grommets at their mounts. Note due to the optical design, the RF 100 Macro is not compatible with teleconverters whereas the EF version is. Both lenses will work with native extender tubes, but right now Canon only makes these in the EF mount, so if you want to use tubes with the RF lens you’ll need third party models designed for the RF mount, like those from Kenko.

Ok now for focusing, starting with the RF 100 at f2.8 on an EOS R5 and in Single Autofocus mode using a central AF area. Here you can see the camera and lens snapping into focus almost instantly with no hesitation or overshooting – and this was with the full focusing range too.

Now let’s switch to the older EF 100mm, opened again to f2.8 and adapted onto the EOS R5 using the same AF modes. This time the focus pulling back and forth is a tad slower, but I’d still say it’s pretty swift and confident; certainly no complaints in this particular test. Note if either lens is struggling to lock-onto a subject, or overshoots, try reducing the range with the focus limiter switch.

Now for the same test but while filming 4k video on the R5, starting with the RF 100 at f2.8. This time there was a little hesitation between each focus-pull, but each was still performed smoothly and confidently without overshooting.

Next for the older EF 100 at f2.8 where again there’s some hesitation between focus-pulls and I’d say the refocusing is also a little slower, but not in a detrimental way. Note the EF lens was very slightly audible while focusing, whereas the RF version was almost silent.

Let’s now try face tracking with the RF 100 at f2.8 on the R5, allowing the camera to find me anywhere on the frame. There’s a slight pause as I enter the frame, but once it locks onto me, the camera and lens combo keeps me in focus, at least while I’m facing them.

Switching to the EF 100 at f2.8 again starts with a brief pause before successfully tracking me as I move back and forth. Overall I’d say the adapted EF lens is working well for video autofocus on the R5, although I did experience a little less consistency when shooting still portraits as I’ll show you in a moment.

Now for focus breathing with the RF 100 at f32, manually focusing from infinity to the closest distance and back again. Since these lenses are designed to focus closer than non-macro models, you may be surprised by the amount of breathing here, but the RF is focusing down to 26cm here where it can deliver that magnification of 1.4x actual size.

For comparison, here’s the EF 100 at f32, manually focusing from infinity to its closest focusing distance of 30cm and back again. There’s significant breathing again here, but notice how it continues breathing beyond where the newer RF lens stopped, even though the minimum distance isn’t as close and the maximum magnification at that point is lower at 1x actual size. I’ll show you how they actually perform at their closest distances in a moment.

But first for stabilisation starting with the RF 100 filming video on the R5 handheld with all stabilisation disabled. Flicking the switch on the side of the barrel activates the optical stabilisation built-into the lens which works alongside the sensor shift IBIS in the R5 whether you like it or not. As the system settles-down, you can see it’s possible to handhold video reasonably well, especially if you’re less shaky than I am.

And for comparison here’s the EF 100, again starting with no stabilisation before flicking the switch on the barrel to engage both the optical stabilisation in the lens as well as the sensor-shift IBIS in the body. Again Canon’s system has both working together where available. From these clips I’d say the RF version has a minor edge, but for video the EF isn’t far behind.

For stills, Canon claims five stops of stabilisation from the optical system alone, or up to eight when combined with bodies featuring IBIS. 

Here’s two shots taken with the RF 100 on the EOS R5 at the same shutter speed of 1/13, on the left with optical and IBIS and on the right without any stabilisation. 1/13 was the slowest shutter I could handhold for a perfectly sharp result with stabilisation enabled, and to match that result without stabilisation on the day required 1/200, so that’s about four stops of compensation. Just for comparison, I’ll switch out the unstabilised result on the right for one at 0.6 seconds with stabilisation – sure it’s a little wobbly, but not terrible.

And for comparison, here’s two shots taken with the EF 100 on the EOS R5, again on the left with optical and IBIS, and on the right without any stabilisation, but this time at the shutter speed of 1 /6. On the day, this was the slowest speed I could handhold a sharp result with all stabilisation enabled, which works out at five stops of compensation in total, making the older lens a stop better than my previous result from the RF lens. 

Now the effectiveness of different stabilising systems can vary between people and situations, but from this single test the newer RF lens certainly didn’t take the lead I expected from the specs, and actually worked out a little worse. Of course your mileage will almost certainly vary.

Ok next for my optical quality tests starting with my standard distant landscape view of Brighton Pier, angled so that fine details run into the corners. You’re looking at the RF 100 f2.8L here on the EOS R5 with the aperture wide-open at f2.8. Taking a closer look in the middle reveals very high sharpness and contrast with a tremendous amount of detail, and when I stop the lens down there’s little to no benefit in sharpness from this distance, proving this lens is already performing very well at the maximum aperture.

For comparison, here’s a close look at the RF 100 2.8L on the left and the older EF 100 2.8L on the right, both at their maximum apertures of f2.8, where you can see the newer RF version on the left is delivering a crisper result with higher contrast. If you view the EF version in isolation it still looks good, but the newer RF model is simply out-performing it at their maximum apertures as you’d expect for a lens that’s over a decade younger.

As you close the apertures on both lenses, the older EF version enjoys a boost in contrast and ultimate detail, coming very close to the RF model at f5.6 and especially by f8. So while the RF is a clear leader wide-open at f2.8, if you can close your lens down, there’s little to choose between them at this point.

Returning to the f2.8 sample and heading into the corner tells much the same story. The newer RF lens on the left remains very crisp, while the older EF version on the right shows a little softness, but is still delivering a great result – it’s just that the newer optics are even better. 

Closing down the aperture on the RF lens again makes little difference to the overall contrast and sharpness, but provides improvements to the EF image which again by f5.6 and f8 becomes very close to the newer RF model. It’s also impressive how little vignetting there was in the corner of either lens even wide-open.

As for coverage, both lenses are almost identical, as their quoted focal lengths would suggest, although as noted earlier, the older EF lens can be used with tele-converters to extend the reach whereas the newer RF version cannot.

Ok, now let’s look at portraits, starting with the RF 100 at f2.8 and using eye-detection on the R5 to focus. Telephoto macros actually make great portrait lenses, with flattering focal lengths, bright apertures for blurry backgrounds and generally well-corrected optics, and the RF 100 is no exception, delivering razor sharp details around my eyes. Like other native RF lenses on EOS R bodies, the eye-detection also proves very consistent and reliable. Meanwhile the background rendering is nice and smooth, and while you won’t get the ultimate blurring of brighter-aperture options, it’s still attractive and not distracting.

Placing the RF 100 on the left and the EF 100 on the right, both from the same distance and wide-open at f2.8 shows once again that the older EF version – or at least my sample – looks a little soft in comparison. I reshot the EF portrait multiple times using different modes and manual focus too, but this was the best result. Again, viewed in isolation, it’s not bad at all, but the newer RF version was simply sharper in my tests, and more consistent with autofocus too.

Close both lenses to f4 though and my EF sample improved a great deal, coming very close to the crispness of the newer RF model. And as for the background rendering, both lenses shared a similar style in this test.

To better-evaluate their bokeh, I photographed this ornament against LED fairy lights at each of their apertures. Here’s the RF 100 at f2.8, and now here’s the EF 100 also at f2.8.

Placing them side-by-side shows both share a lot of the same rendering characteristics, with well-behaved bokeh blobs lacking the distracting onion ring textures or outlines of lesser lenses. As the blobs become less circular towards the edges though, notice how their asymmetric shapes become flipped to each other – not a problem, just an observation.

As you close the aperture down on both lenses, the shapes become more uniformly circular across the frame, albeit influenced by their respective aperture blade systems. I don’t have a preference between them here, do you?

Which finally brings me to the most important aspect of any macro lens, the close-up performance, especially at the minimum focusing distance. To measure their maximum reproduction without accessories, I photographed a ruler as close as both lenses would focus, and this time I’m starting with the EF 100 at its closest focusing distance of around 30cm where it’s capturing a subject width of 36mm on an EOS R5. This confirms its true 1:1 reproduction capabilities or 1x magnification.

But now here’s the RF 100 from its closest focusing distance of around 26cm, where it’s capturing a smaller subject width of 25mm on an EOS R5, confirming its greater 1.4:1 reproduction, or 1.4x magnification.

Greater reproduction is always desirable for macro enthusiasts and this is the major benefit of the newer RF lens over its predecessor, although I should note the EF 100 can achieve 1.37x magnification when coupled with the optional EF25 II extension tube accessory. You could of course also use extension tubes on the RF lens, but Canon doesn’t make any in the native RF mount as yet, so you’d need to with third parties like Kenko. 

To compare their quality at 1:1 magnification, I photographed a UK pound coin, which measures approximately 23mm between points, thereby almost filling the vertical height of a full-frame sensor at 1x. So here you’re looking at the full height of the photo, uncropped, starting with the RF 100, before switching to the EF 100, both at f5.6. 

Placing them side-by-side shows a very similar degree of fine detail when shot with the aperture closed-down. Looking closer perhaps reveals the RF version on the left to be a tad crisper, but there’s barely anything in it for an unstacked macro shot at 1:1 with the apertures closed from the maximum.

But the RF lens can deliver greater reproduction without accessories, so before moving on, here’s the coin with the RF 100 on the left and the EF 100 on the right, both operating at their maximum magnifications of 1.4x and 1x respectively.

Note the RF 100 does appear to suffer a little from focus shifting when operating at very close range where the plane of focus can shift a tad backwards as you close the aperture. It wasn’t an issue in most of my tests, but I wanted to mention it.

Next here’s a sequence of 100 images I shot with the RF 100 at f4 using focus bracketing on the EOS R5 to finely adjust the focus on each shot. The goal here is to use stacking software afterwards to combine all the images into a single file where more or even all of the subject can become completely sharp. Focus stacking is a popular technique used by macro photographers and Canon claims to have reduced the breathing on the RF lens compared to the EF version, but you can still see a magnification change here as the camera focuses from one side of the coin to the other.

And here’s the final result, with 100 frames assembled using Helicon Focus software, available for Macs and PCs, and using the default render settings here. The total depth of field here is far greater than you could achieve even with the aperture closed to f32, and by shooting each frame at larger apertures like f4, you’ll also avoid the softening effect of diffraction. Helicon has also dealt with the magnification change due to breathing.

Here’s another example, showing the focus stacking process in action. I’d highly recommend giving it a try if your camera has focus bracketing capabilities, and Helicon Focus has always produced great results for me; there’s a free trial on their website.