Voigtländer Color Skopar 20/3.5 SLII
The optimum competitor for this lens would be the Pentax DA21. Unfortunately I do not own the lens anymore. I now have the Sigma 20/1.8. But I tested the DA21 extensively against the Sigma. My conclusion was that the DA21 is very good and fully comparable to the Sigma. Only disadvantage of the DA21 was a slightly weaker resolution in the corners. I kept the Sigma because of its f1.8 aperture.
So, in the following I compare the Color Skopar with the Sigma 20/1.8 EX DG Asph.
The Color Skopar is very sharp.
The Sigma comes out slightly worse at the corners at infinity. This is probably owing to the large aperture of the Sigma.
Otherwise the lenses show the same behavior.
The Color Skopar shows very good contrast already wide open. I have tested the contrast against the Sigma. At f3.5 the Sigma has slightly better contrast at high frequencies, whereas the Color Skopar has slightly better contrast at low frequencies. The contrast increases somewhat at f8, most notably at at higher frequencies. The Sigma shows comparatively low contrast at f1.8, which was to be expected.
The barrel distortion is noticable at close focussing distances. It is no problem at all at larger distances. The degree of distortion is about the same for the Color Skopar, the Sigma 20 and the DA21.
Good, normal behavior for a wide-angle lens.
Chromatic aberrations are noticable at the corners of the image in strong contrast stiuations, but they are weak. Nothing to worry about.
Out of focus color fringes are basically not existent.
No flare detectable at all with the Color Skopar! Very good!
Difficult to judge. The circles of confusion show the concentric rings typical for an aspherical lens.
The circles are quite small. Bokeh is certainly not a speciality of this (type of) lens.
Normal behavior. As a full frame lens, vignetting is no problem at all for the Color Skopar on APS-C format. It shows a fall-off of about 25% in the extreme corners. This is hardly noticeable in the field. The DA21 is of course worse here (ca. 50%).
Resolution has been tested using the USAF test sheet, tripod, MLU and manual focussing, using a focus array of 20-30 images each.
Of each lot the sharpest image has been selected by visual inspection.
Resolution at a distance of 2m: 100% Crops
The resolution of all lenses is very good. They outresolve the sensor.
Below is a 400% crop from the inner part of the USAF sheet.
What one can see is that at the number 4 it appears that the orientation of the bars is changing. Here the resolution is so high that the sensor cannot cope with it anymore correctly and artificial, "spurious" patterns appear. I think this has to do with the Nyquist theorem.
Resolution at infinity: 100% Crops
(remember, this is the best out of ca. 20 shots)
In the center they are basically the same, but the Color Skopar comes out slightly better in the corners. Note the steps of the ladder of the middle antenna. They are not visible anymore in the Sigma shot.
Contrast has been measured using the method described here, using Image-J.
Do not give much on the absolute numbers (which depend on lighting) and not even much on the exact position of the points. The values have been obtained from only one image each and the error is probably quite high.
I have tested the contrast wide open, at f3.5 and f8 for the Color Skopar and the Sigma.
The test pattern:
Color Skopar versus Sigma, both at f3.5
Color Skopar f3.5 versus Sigma f1.8
Color Skopar f3.5 versus f8
Important for a wide angle lens is the distortion.
The strongest distortion is visible at the minimum focussing distance.
Below are the images of the Color Skopar and the Sigma. The distortion is virtually identical.
In the Photoshop "lens correction" filter I have to use a factor of about 3.4 to correct this.
I think this is a pretty normal value.
DA21: a friend provided me with this image. Thank you!
If you go backwards a few steps, distortion decreases.
Below is a photo taken from about 1.5 m distance.
In Photoshop it takes about a factor of 1 to fully correct it. This is very little.
Chromatic aberrations are no problem with these lenses. CAs are there, but comparatively weak. Below is the worst example I could produce.
100 % crops:
Some also wanted to know about the out of focus color fringes. No problem at all:
Compare DA55 as a contrast.
Purple fringing is weak.
Below is a 100% crop of the top left corner from a shot into a treetop:
This is a surprise. I wasn't able to produce any typical colored flare with the Color Skopar!
The Sigma produced very beautiful flare, but nothing with the Cosina. Both lenses have been used without hood.
Flare with the Sigma:
No flare with the Color Skopar:
Here is another shot directly into the sun with EV +2:
100% Crop from the above image:
The only thing I noticed was haze at times, an overall reduction of contrast, when you shoot directly into the light.
Click here for an example. The darker one is taken with EV = - 1.
Well, the Color Skopar was probably not designed for good bokeh. Nevertheless one can look at it.
The circle of confusion shows the typical concentric rings of a lens with an aspherical element.
Below is a somewhat enhanced image of the circle of confusion of the Color Skopar:
Turning the focussing ring from infinity to the minimum focussing distance:
The circle of the Sigma is larger. All at f3.5
Overall the circles look similar.
This has been produced again by turning the focus ring in small steps, starting from infinity (left).
Please note that it is impossible to do this stepping 100% identical for two lenses. Slight differences in the focus distance are unavoidable.
Note also that the contrast has been enhanced to make the effect better noticeable.
For some real life Bokeh examples please continue to the Samples section.
I must admit that I have not much experience with vignetting, since I never considered this a serious problem.
I shot a white wall slightly blurred at f3.5 and took the profile along the diagonal.
The fall-off in the corners is about 20-25%.
I have found a diagram of the Zeiss Distagon, for comparison:
The Zeiss is slightly better, but comparable in general: