I wrote a short piece recently to explain why I decided upon this lens rather than its competitors, so I will now run through the first phase of testing and explain my first impressions. This lens has made plenty…
Starting with the review basics:
This is at a very high standard. While it lacks the metal alloy outer construction of some other L series lenses, it is made from a tough plastic that mimics the same finish. With a flick of the finger you can detect the difference, but Canon has executed it very well and perhaps to an even higher standard than on the likes of the 70-200 f4 L. It feels solid and a quality product through and through. Design-wise, it matches the legendary 85mm f1.2 L II (which I love) and you can clearly see the two lenses are form the same stable. The 50mm is just over half the weight though (545g vs. 1025g), so perhaps we should be glad Canon dropped the metal. The mount itself is very sturdy, the usual L weather seal is in place and it exudes quality in a functional DSLR way, rather than a Leica M ‘made to survive a nuclear apocalypse’ way.
These are just fine. The AF/MF switch is easy to activate intentionally, but difficult to move by accident. The lens narrows towards the mount nicely, so that the release button can be easily depressed and the lens removed. The hood is a usual twist on and click into place design and has no lock, which I like. The locking hood on my 24-70 f2.8 L II does my head in, because none of my other L lenses have one! It has a nice flocking lining on the inside and looks functional, if perhaps not quite as well made and robust as it might be. Its just a bit below the standard of the rest of the lens.
Auto and Manual Focus
It is reasonably fast, virtually silent and very accurate indeed. It is faster than the 85mm f1.2 L II, but still feels like it is running on a similar and very robust bearing system. On the 5D III it does not hunt, but once on focus that really is it: done. I have many faster to focus EF lenses, but I find this lens plenty fast enough for its likely application: people shots and portraits, rather than action. You can detect that it is smooth and swift, rather than lightning quick, as it is on the likes of the 24mm f1.2 L II, 24-70 L II etc. The manual focus ring is much better than on the 85mm, however. It has a really good weighting and resistance, rather than using the ultra light fly-by-wire of the 85mm f1.2 I & II. There is more to say on the subject of achieving correct focus, so please read on…
Basics out of the way, let’s move onto actual performance. Before we move onto the images I will present in this review, please note that I have already tested the lens on inanimate objects. During these tests, several points became clear:
1. It resolves surprisingly well on centre at f1.2. Yes, it does. Really. My copy did not need any focus compensation dialing in. It is smack on, wide open. End of. However, off centre, the lens does degrade significantly and this is in keeping with some of the commentary on the web. This is particularly acute at long distances and seems to be reduced closer in. While I could not personally care less how sharp the lens is edge to edge at wider apertures at infinity, I do need to know how large the zone of central sharpness is at various apertures at people distances. This is where I will be using f1.2 to f2.8 (and possibly smaller) and so one has to ask, ‘how usable is f1.2 for portraiture?’ The issue is this: how often are the subject’s eyes, or the point of focus smack in the centre of the frame? ‘Not often’ is the answer (which is a source of much frustration for Leica M photographers wanting to shoot wide open up close, because the rangefinder patch is in the middle of the viewfinder). We therefore need to know that a face placed a third off centre can still be rendered sharply and this is not going to be possible if the zone of good resolution is too small and central at f1.2.
2. The lens + camera + Tom did not always manage to nail perfect focus, hand-held, even on centre. However, ‘we’ did become much more consistent as distances increased beyond a metre or so. This will be familiar to anyone used to the 85mm f1.2, who knows how much tiny changes in subject camera distances can affect resolution due to limited depth of field. Even with the shorter 50mm, it is clear that at close distances (say, 70cm), even a sway back or forth of 1cm can completely screw up the shot. You cannot even perceive such changes, which is why I did some tripod testing and was able to see that much of the error was mine and not down the lens. Poor subject contrast definitely saw the camera and lens combo start to struggle to deliver accurate AF. Such are the demands of f1.2!
3. At close to minimum focus distance, the lens does not deliver consistently sharp results using AF. This is something I will look into in more detail, perhaps in part 2.
4. The lens has enormous vignetting wide open, so much so, that the entire frame appear underexposed!
5. The lens does indeed produce a distinctive combination of ‘soft sharpness’ at f1.2-f1.6. By this I mean that line details are well recorded – surprisingly so – but delivery is remarkably gentle. This seems to be a combination of the following:
- Focus melts away from the point of focus so smoothly that you cannot see it happening. This is not always the case: some lenses have a very abrupt change in focus, such that the resultant image can feel jarring.
- There a slight glow. This is probably spherical aberration.
- Contrast is lower wide open, along with colour saturation.
In order to take things further, it was clear that I needed to photograph a human subject. People move and their eyes are a given size relative to a camera’s focus points, which can prove important (at least with AF rather than live view). While choosing a woman would have appealed to some, there is a problem: few women want their photo plastered over the internet ‘avec wrinkles’, which means they will insist on make up and/or post processing, which defeats the object of this test, which is to show you what you actually get! I therefore found an excellent compromise and enlisted the services of the lovely Pauline, an aspiring actress and award-winning bearded lady from Wales.
Here are the shots, which I have tried to shoot at realistic ’50mm’ distances i.e. how one might apply this lens in anger so to speak.
Please note that all images below are 1500 pixels on the long side.
As you can see, the lens produces a very gentle, but clearly sharp result at f1.2. In this case the subject’s eyes are 1/3 of the way down from the top of the frame, which I thought was a reasonable ‘outer test zone’ (resolution smack on centre is not in doubt, for me). Getting sharp results was not always easy, for reasons I have mentioned, but its hardly impossible either, even hand-held and in a hurry.
By f1.8 the lens is developing more bite and the glow is largely gone. By f2.0 it is clearly extremely sharp, even well of centre (at these distances) and by f2.8 it is absolutely biting. Note that at these apertures, you need to go easy on sharpening, because the images contain so much acuity and contrast that they are easily pushed into unreality. Personally, I find the results at f2.8 or so more pleasing for even male portraits with default sharpening, or at least very mild further sharpening.
Note how the 50mm is starting to distort perspective in the above two frames, due to proximity. I would normally use a longer lens at this sort of distance, but I have included it for your reference.
Click below for next page.