Sigma 10 mm f/2.8 EX DC FISHEYE HSM
4. Image resolution
Additionally, the proximity of the lens and the chart might be favorable for the tested instrument. The lens is so close that you can hardly expect any kind of distortion or even hot air movements. Short focal length additionally marginalizes any minimal vibrations of the body, which can’t be eliminated even using the best tripod stands, a cable release or an initial mirror lifting.
Let’s get down to facts. Similarly to the circular Sigma 2.8/4.5 we made the test on RAW files from the Canon 20D. Also in this case the frame center performance can be acclaimed as simply outstanding and right from the maximum relative aperture! There, the lens reaches a very high result of 37 lpmm; what’s more, by f/5.6 it breaks the Canon 20D record by achieving as much as 44.7 lpmm.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Like in the case of the Sigma 4.5 mm we have suspected that the results are a bit overstated. Apart from the resolution record, there’s also a very high value (32.7 lpmm) reached by f/16. For normal lenses with the focal lengths from more than a dozen to above 100 mm the MTF50 values by f/16 are a tad above 30 lpmm on the Canon 20D. In the case of the Sigma 10 mm the result is better by several percentage points. Even if we corrected that result downwards we wouldn’t change our overall frame center assessment – the performance here is brilliant.
Much worse situation can be observed in the frame edge. Once again, though, we must repeat the Sigma 4.5 mm test arguments. In the case of a “fisheye” the frame edge resolution areas move significantly from the center and are very distorted so the result is understated. You can assess the lens in the best way by having a look at the sample shots with some far away objects. The frames are nice, with most of the area sharp, which clearly proves that flat charts tests with the “fisheye” lenses are very difficult and the results don’t show the true frame edge performance.
At the end of the chapter we present the excerpts of our test chart for the frame center, saved in the JPEG format (standard sharpening).