 |
RF 16mm f2.8 at f3.5 on EOS R5 Manual focus at 5m |
With advent of the RF mount Canon appears to have gained newfound enthusiasm for innovative approaches to the range of products they design and make.
The RF 16mm f2.8 is a very good example of this. Traditionally ultrawide lenses have been big heavy expensive things. But here we have an ultrawide which is very compact, light and inexpensive.
So, is it any good? Spoiler alert, yes it is, read on for the details.
By chance the RF 16mm f2.8 and the RF 100mm f2.8 L macro lenses arrived in my household within a few days of each other. Each is an excellent example of Canon’s thoughtful approach to lens concepts and design in the mirrorless era.
In terms of outright optical and mechanical capability they occupy a different place in the catalogue.
The 100mm f2.8 L macro is a very high end product with exceptional sharpness right across the frame at all apertures and focus distances.
The RF 16mm f2.8 is somewhat at the other end of the spectrum with more modest and limited optical and mechanical capabilities. It does however fill a definite niche and as such will be appreciated by plenty of Canon RF mount camera users.
Here is the picture which explains the reason for the existence of the RF 16mm f2.8.
 |
As many users already have a standard zoom starting at 24mm the 16mm prime makes sense as an ultrawide companion to the standard zoom.Photo courtesy of camerasize.com |
And a little data table:
Lens model | Length mm | Mass grams | Price AUD retail GST paid |
RF 16mm f2.8 | 42 | 165 | 489 |
RF 14-35mm f4 | 100 | 540 | 2378 |
RF 15-35mm f2.8 | 127 | 890 | 3704 |
There are people like me who infrequently need an ultra wide lens and have no interest in buying or carrying either of the ultrawide zooms.
The RF16mm f2.8 is perfectly suited to this buyer group. It is very small, relatively inexpensive and of good optical quality. It can be used to make excellent pictures provided you understand and work with its limitations.
It can be used for interior architecture and real estate, some exterior architecture, landscape and some types of street and documentary work where one is working close to the main subject(s). Some users on forums have indicated they want to use it for Astro. I have not yet tested the lens for Astro so I await feedback from other users about this use case.
As far as I am aware no other lens maker currently offers a comparable product.
Physical
As with all RF mount lenses the 16mm f2.8 is compatible only with Canon RF mount camera bodies.
The RF 16mm f2.8 appears to use the same housing as the RF 50mm f1.8. When placed side by side the two lenses appear identical apart from the front and rear elements and the fact that one is labelled 50 and the other 16.
Both have several cost cutting features. They do not come with a lens hood (none of Canon’s non-L lenses does) they lack weather sealing, do not have AF/MF or IS On/Off switches, utilise front focussing, have a single smoothy ring which can be used for manual focus or another function as set in the menus.
They both have a 43mm filter thread, utilise an STM (stepper) motor and can focus quite close.
At closest focus distance the front element of the 16mm is only about 40mm from the subject, giving a maximum reproduction ratio of 0.26x.
AF/MF and Stabiliser On/Off have to be set in the menus. I allocate these items to My Menu for reasonably quick access.
Autofocus speed is decent but not as quick as that on lenses which utilise USM focus drive.
Stabiliser effectiveness
There is no stabiliser in the lens. The R3,5 and 6 have in body stabilisation, the R and RP do not.
On my standard test I can get sharp shots hand held down to around 1/20 sec without the stabiliser and around 1/5 sec with the stabiliser in the R5.
This represents a modest but useful 2 EV step benefit with the stabiliser.
Optics
The RF 16mm has 9 elements in 7 groups with 1 aspheric element. The front element extends about 12 mm at closest focus. The effective focal length increases with close focus as is usual with front focussing lenses like this.
 |
The front of the lens is on the left. |
The appearance of uncorrected Raw files indicates that the actual focal length of this lens is less than 16mm. The output JPGs or Raw files converted in Canon DPP have been corrected for distortion and substantially cropped.
The lens is designed with obligatory reliance on post capture correction of distortion, peripheral shading and color fringing. Some purists on user forums complain about this as if the practice were somehow immoral but the reality is that every lens produced these days does require post capture correction of some kind. Even the otherwise amazing and almost perfect RF 100mm f2.8 L macro needs a bit of post processing help with peripheral shading at f2.8.
 |
The MTF chart indicates expected performance at f2.8 only. This matches real world experience which sees very good central sharpness at f2.8 but soft periphery which improves at smaller apertures. |
Uncorrected Raw files from the RF 16mm f2.8 exhibit massive barrel distortion (which is the same at all apertures) with considerable peripheral shading and color fringing at f2.8.
Geometry is fully corrected in JPGs and Raw files converted in Canon Digital Photo Professional.
The extent to which peripheral shading, color fringing and diffraction are corrected in JPG files depends on settings applied in the Camera (red) menu, [lens aberration correction]. On the R5 this is on screen 4.
 |
Uncorrected Raw file opened in Adobe |
At the time of writing Adobe does not yet have a correction profile for the RF 16mm (or the RF50mm f1.8 by the way) but one is needed ASAP.
In the meantime the profile for the RF 24-240mm lens can be used as a starting point for adjustments in Camera Raw or Lightroom. I add distortion correction 130 (in the Optics tab in Camera Raw) and Vignette correction dependent on aperture.
 |
Appearance in the viewfinder or a JPG or Raw converted in Canon DPP |
Sharpness
Very good to excellent sharpness is available in a substantial center circle right from f2.8.
The diameter of the circle of very good sharpness increases as the lens aperture is closed down to f4 and f5.6. The little diagram below is my attempt to illustrate this graphically.
Basically sharpness extends to the corners by f5.6 which on my tests is also the optimum aperture for sharpness across the frame. F8 is also very good but Some decline in sharpness due to diffraction is perceptible from f11.
The lens is perfectly usable at f2.8 for subjects which do not require critical sharpness in the corners.
The definition of “sharpness” here is subjective of course. One person’s “good enough” sharpness may be another’s “unsatisfactory”. Subject type is an important factor. For instance sharpness sufficient for an architectural subject might not be good enough for foliage.
Notwithstanding this the basic formula remains true. Sharpness is very good to excellent in the center at any aperture and improves in the periphery as the lens aperture is closed down to f5.6 or f8. Further reduction in the aperture to f11 or f16 reduces definition due to diffraction but may be necessary to achieve adequate depth of field for close-ups.
 |
Landscape at f5.6 |
Color fringing
In Raw files this can be quite prominent towards the corners with some subjects especially foliage against a hot (white) sky. It is generally correctable in Camera Raw. I find that using the eyedropper in the Optics > Defringe tab in ACR mostly works well.
Flare
I find the lens to be quite resistant to flare even when I try to induce it with the sun at or near the frame edge. This is a pleasant surprise as ultra wide lenses can sometimes be very flare prone.
Some conditions can produce flare blobs near the edge of the frame but these are generally manageable in post processing.
 |
Foliage f8 Purple fringing in the corners removed in Adobe camera Raw |
Bokeh
The out of focus rendition is quite pleasing to my eye with no sign of unpleasant nisen or other nastiness seen in testing so far.
3D-sharpness distribution as the aperture is closed
For this test I used my usual protocol which you can readabout here. This is simple and easy to do yet reveals considerable useful information about the 3 dimensional distribution of sharpness of a lens. This is especially important with wide and ultra wide lenses which often exhibit complex sharpness distribution.
The photo below shows sharpness distribution at f5.6 which will be a frequently used aperture with this lens. You can see that more sharpness is distributed towards the camera from the focus point (on the stick) than away from the camera. I have found it is essential to run tests on each lens separately as very few of them behave in the way reference books would have you believe. This particularly applies to wide and ultrawide types.
 |
Sharpness distribution at f5.6 |
This tells me that with a typical landscape type scene I need to place the focus point higher up the frame than (equivalent to further back from) the midpoint.
On further testing I have found that for many subjects simply setting focus manually at 5 meters and using f5.6 or f8 gives a very satisfactory result with good distribution of sharpness front-to-back and into the corners.
On my R5 focus is at 5 meters when the analogue manual focus indicator is at the middle of the infinity symbol, see photo below.
Unfortunately the analogue manual focus scale in the R5 is all on one line. The little Powershot G1X.3 divides the scale into two parts, close-up range and normal range which provides a much more user friendly indication of focus distance in the normal range.
Aperture for use case
These are just suggestions which can be used as the start point for experiment.
F2.8 Low light, indoors, peripheral sharpness not important.
F4 Most indoor and low light work with moving subjects. Corners still soft.
F5.6 Basic aperture for outdoor work. Good all round image quality and depth of field.
F8 Basic aperture for landscape.
F11 Landscape if foreground is very close to camera. (less than a meter).
F16 Basic aperture for close-ups when the front element of the lens is very close to the subject. (4-10cm).
 |
Close-ups with an ultrawide lens are a bit of a novelty |
Alternatives
Apart from the Canon RF zooms already mentioned there are several third party lens makers which produce ultrawides for a range of lens mounts including Canon RF.
Some of these are reported to be very good optically but most have no electrical contacts so all manual exposure and focus is always required.
In addition these lenses are generally larger, heavier and more expensive than the RF 16mm. Several have a bulbous front element precluding the use of a filter.
Summary
The RF 16mm f2.8 is a welcome addition to the growing RF lens catalogue. It is small, light, and inexpensive yet can deliver very good results if used thoughtfully. It is a genuine Canon with a 5 year warranty and ongoing firmware updates as and if required for new model camera bodies.
Highly recommended.
.