denoir Offline Upload & Sell: Off
|
So, whats the point of comparing a 350 ($500) lens with a 3500 ($5000) lens? Completely pointless, right? Well, maybe but then again maybe not.
These are not any old lenses, but two 50 f/1.4s. Fast 50mm lenses were for some 50 years or so the standard SLR kit lenses. If there is something lens producers know how to do is fast 50s. Its really difficult to find a bad one. As a rule a fast 50 should be small, compact, cheap and a good performer. So its interesting to see what one of these standard lenses can do against a lens design where all stops have been pulled.
The lenses in question are the Canon EF 50/1.4 and the Leica Summilux-M 50/1.4 ASPH.
http://peltarion.eu/img/comp/lux50-4.jpg
The Canon 50/1.4 is truly an average prime. Its not Canons cheapest one and not Canons most expensive one. Its light, small (for an SLR lens) , cheap (350) and generally considered to be a solid performer. In short it is exactly your typical fast 50.
The Leica 50/1.4 Summilux ASPH can be considered to be the flag ship of modern Leica lens designs. Its big (for a rangefinder lens), heavy (almost 2x the weight of the Canon), expensive (3500) and generally considered to be a milestone in lens design. Leica set out to make the best normal lens and few people think they failed.
Now before I get to the test part, a few words on the execution of the test. The big issue is of course the use of two very different cameras the Canon 5DII for the Canon 50/1.4 and the Leica M9 for the 50 Lux ASPH. Although I have used RAW development profiles that in theory should match colors the same way the reality is quite different and youll have to take that into account.
The Canon has some focus shift and I used live view to focus (with the lens stopped down to the right aperture) so the Canon shots are definitely dead on. The M9 is a different story theres no live view. My rangefinder seems to be well calibrated and the 50 Lux ASPH should not suffer from noticeable focus shift, but both are possible sources of error when it comes to focusing the M9.
Alright, on to the tests. Open the images in separate tabs in your browser and flip between them.
Test A:
http://peltarion.eu/img/comp/50cl/A.jpg
This is a very good test subject that shows difference in sharpness across the frame, color separation, geometric distortions, vignetting etc. Its a painting that hangs over my work desk at home. The shots are closeups at about 0.8 meters. Neither of the lenses are macro designs so it should be a relatively weak point for both.
Test at f/1.4:
--Canon EF 50/1.4 @ f/1.4
--Leica Summilux 50/1.4 ASPH @ f/1.4
Comment: Interesting to observe is the difference in clarity, geometrical distortions and perhaps most of all color separation. Both lenses show pretty good resolution at the widest aperture (expected from the Leica but it was a surprise to me that the Canon was as good in terms of resolving power at f/1.4).
Test at f/5.6:
--Canon EF 50/1.4 @ f/5.6
--Leica Summilux 50/1.4 ASPH @ f/5.6
Comment: The Leica has improved marginally, mostly in corner resolution while the improvement in the Canon is dramatic. Very nice corner to corner sharpness. The Canon is no slouch at f/5.6.
Test B:
http://peltarion.eu/img/comp/50cl/B.jpg
Test of medium-close distance rendering wide open.
--Canon EF 50/1.4 @ f/1.4
--Leica Summilux 50/1.4 ASPH @ f/1.4
Comment: Here we can see the characteristically sharp sharpness-to-blur gradient of Leica glass compared to the more soft transition of the Canon. The effective DOF is significantly smaller with the Leica.
Test C:
http://peltarion.eu/img/comp/50cl/C.jpg
Test of close distance rendering wide open.
--Canon EF 50/1.4 @ f/1.4
--Leica Summilux 50/1.4 ASPH @ f/1.4
Comment: Again, we see the different sharpness-to-blur gradients. Aberrations are well controlled on the Summilux, but not all that great on the Canon.
Test D:
http://peltarion.eu/img/comp/50cl/D.jpg
Complex background test.
--Canon EF 50/1.4 @ f/1.4
--Leica Summilux 50/1.4 ASPH @ f/1.4
Comment: Some truly awful bokeh from the Canon. Unfortunately its a common problem with this lens when you have complex backgrounds.
Test E:
http://peltarion.eu/img/comp/50cl/E.jpg
Stopped down (f/5.6) performance at infinity. Just to show that the 50/1.4 is actually a good prime, Ive included two more lenses here the awful Sigma 18-200 OS zoom and the Sigma 30/1.4, both on a crop camera (7D) giving an effective equivalent focal length to a 50.
--Canon EF 50/1.4 @ f/5.6
--Leica Summilux 50/1.4 ASPH @ f/5.6
--Sigma DX 30/1.4 @ f/5.6
--Sigma 18-200 OS@ f/5.6
Comment: Not much to comment here. The 50 Lux ASPH produces excellent colors, contrast, color separation and a distortion free image. The Canon lags behind in all areas. The two Sigmas are far behind the Canon with the prime producing better colors and contrast across the frame.
A couple of 100% crops:
http://peltarion.eu/img/comp/50cl/E_Crops.jpg
Comment: This was a near center crop. Both the Canon and the Leica show outstanding performance. The difference you see in per pixel sharpness is due to the different camera sensors not due to the lenses. The Sigma 18-200 performs as awfully as I expected it to. The real shocker was the performance of the Sigma 30.
http://peltarion.eu/img/comp/50cl/E_Crops3.jpg
Comment: This is left of center, mid frame. The Canon has dropped in resolution and the Sigmas are as awful as they were in the center. The drop in Canon resolution is interesting as we did not see it in test A (close distance). Generally it seems the Canon is better at short and medium distances than it is at infinity.
http://peltarion.eu/img/comp/50cl/E_Crops2.jpg
Comment: Left edge. The Leica is still going strong while the Canon is definitely weaker now than it was in the center of the image. The Sigmas are far behind the Canon with the zoom being much worse than the prime.
Conclusions:
The interesting part here is not that the Summilux is better in every situation anything else would have been astounding. Whats interesting is in what way it is better. Leica went out of their way to produce a lens that is consistently superb at all apertures and distances. They succeeded, but there was no quick and cheap solution. For instance more than half of the cost of the lens is for a single optical element of exotic glass, produced by just one manufacturer exclusively for Leica. That glass has to cool off for over six months before it can be ground into the element.
The Canon and similar lenses control for first and second order aberrations the Summilux controls for up to fifth and sixth order. Lenses like everything else suffer from the law of diminishing returns. Getting the last 5% of the way may be ten times as difficult as the first 95%. So naturally the Canon is much better value for money. The improvement Sigma 30->Canon 50 is at least as big as the Canon->Leica improvement and the price difference is an order of magnitude smaller.
Still, if you want the perfect 50mm prime, the Summilux comes very close. Its an amazing piece of glass and a milestone in modern lens designs.
The Canon 50 is of course not in the same league, but it's still a competent performer. It's a relatively nice portrait lens that can be suitable if you wish to produce a more moody, softer look.
|