Lens review

Canon EF 50 mm f/1.2L USM

10 August 2010
Szymon Starczewski

4. Image resolution

The Canon EF 50 mm f/1.2L USM was tested by us in the transitional period when we used the APS-C sensor Canon 50D and the full frame 1Ds MkIII. The review of the test results we will start with the analysis of RAW files from the former camera. The graph, presenting the performance in the centre and at the edge of the 15 megapixels APS-C sensor, is shown below.

Canon EF 50 mm f/1.2L USM - Image resolution

At the beginning it’s worth stressing out that in tests performed with the 50D the decency level is placed near 34-35 lpmm and the best fixed focal length lenses reach the results of 51-52 lpmm. How the tested lens compares here? We must honestly say that by f/1.2 and f/1.4 the performance is weak. The decency level in the frame centre is reached only near f/1.8. On further stopping down the MTF50 values increase very quickly; their maximum- a bit over 50 lpmm- is achieved by f/4.0 and f/5.6. These values are very high, although not exactly record-breaking and such should be expected from an expensive L series prime lens.

- - - - - - - - - - - - - - - - - advertisement - - - - - - - - - - - - - - - - -

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

The edge of the APS-C sensor is even less interesting. The decency level is reached here only after stopping down to more than f/2.8. Such a performance would be more appropriate for a device costing 300 – 600 $; in the case of an expensive 50L we definitely expect a better performance.

Now let’s consult the photos from the EOS 1Ds MkIII. It’s worth reminding that, because of smaller pixels density, the MTFs are lower here. The decency level is situated near 31 lpmm and the best lenses reach about 46 lpmm. The graph below shows the MTF50 values for the centre of the full frame sensor corresponding to the edge of the APS-C sensor and at the edge of full frame.

Canon EF 50 mm f/1.2L USM - Image resolution

The conclusions which can be drawn for the frame centre are, in fact, identical as in the case of 50D results. The images we got by f/1.2 and f/1.4 are “soft” and the good level of focus we reach only after stopping down the lens to near f/1.8.

At the edge of the APS-C sensor the results are in full accordance to those from the 50D – to approach sensible MTF50 values we must stop down the lens to more than f/2.8. The worst situation can be observed at the edge of the full frame sensor. Even after a significant stopping down the lens barely brushes against the decency level there.

It’s worth noticing an interesting performance at the maximum relative aperture. The plane of focus and off-axis aberration make the sharpness graph rather unusual. After you move from the frame centre the image quality starts to decline very slowly but after a while it starts to increase again and at the very edge of the full frame it reaches the same values as in the centre within the margin of error.

How does the EF 1.2/50 compares with other fast prime lenses? The Nikkor AF-S 50 mm f/1.4G, launched in 2008, is slightly worse than a tad faster Canon. What’s interesting, it fares a bit better at the edge of the full frame sensor for a change, achieving even almost 36 lpmm there. Taking the price into account, though, it would be better to invest in a Sigma 1.4/50 or a Zeiss Planar 1.4/50 instead of buying the Canon 1.2/50 – these two, being two-three times cheaper, have similar results by f/1.4 and f/2.0 than the pricey Canon. Does paying extra a lot of money for just a slightly faster device make sense? We will try to discuss that in the summary, having already more knowledge about the Canon’s performance in other testing categories.

Canon EF 50 mm f/1.2L USM - Image resolution