Olympus Zuiko 17mm vs Panasonic Lumix 20mm

Exposure and DoF is the same. If it looks like the first lens has less blur in the OOF area, it is primarily because that lens is sharper in the in-focus area.

The first is the 20/1.7 and the second the 17/2.8. They are part of a series where I tested the bokeh of eight MFT WA lenses (including some zooms with a WA component). In the little poll I organized before revealing which lens took which image, the 17/2.8 came in as number six, just ahead of the Panasonic 7-14 (no. 7) and Olympus 12/2 (no. 8). The bokeh winner was the Panasonic 14-42/3.5-5.6 (not the X-version) followed by the 20/1.7 (no. 2). More info here if you are interested:

http://forums.dpreview.com/forums/thread/3279903#forum-post-50023507

Here are crops that let you inspect the OOF rendering at greater detail. The first is the 17/2.8 and the second the Panasonic 14-42 at 17 mm, both shot at f/5.0 (again click "view original size" to see them properly).

Olympus 17/2.8 at f/5.0

Panasonic 14-42/3.5-5.6 (non-X version) at 17 mm and f/5.0

As you can see, the crop from the 17/2.8 is full of small rings with a bright rim and a darker core. It is as though it would have measles. The fact that the rings are significantly brighter at the edges than at the core is known as outlining and is conventionally associated with bad bokeh. If you look closely, you will also see that the outermost part of the ring has a greenish fringe and that the core tends towards magenta. This is a sign of longitudinal CA and is, just like outlining, conventionally associated with bad bokeh. The presence of these phenomena is also what makes the rings stand out. Less of both and they would have melted together into a more harmonious whole.

When the rings are sufficiently numerous as well crammed, as in the naked branches in the upper left corner, they nevertheless coalesce but not in a very palatable way according to the canons of bokeh (and my taste). The branches look very harsh since the outlining leads to what the Japanese call nisen-bokeh (double-line-bokeh).

If, instead, you look at the crop from the Panasonic 14-42, the individual rings are hardly visible at all. Instead, they for the most part melt together into a rather pleasant whole. As a result, everything looks signicantly smoother and calmer.



In general, I disagree. Most lenses have pretty palatable OOF rendering when the background is strongly blurred. It is in the transition zone, where things are no longer perfectly sharp but not strongly blurred, as in the examples above, that lenses with poor bokeh tend to have difficulties. Have a look at the samples below, the first two from the 17/2.8 and the second two from the 20/1.7. Again, it is obvious that the 20 does better in the moderately OOF shot but as shown by the shot where things are strongly OOF, both do pretty well when things get sufficiently blurry.

Regrettably, it is not easy to get a whole lot of blur with WAs (even fast WAs) on MFT unless you go closer than I ordinarily would with a WA lens. Again, the first examples I showed above are much more representative of the amount of blur you are likely to see.

Olympus 17/2.8 at f/2.8

Olympus 17/2.8 at f/2.8

Panasonic 20/1.7 at f/1.7

Panasonic 20/1.7 at f/1.7

You linked to the sample from the 20/1.7 a second time rather than that from the 17/1.8. Here's the correct one:

http://pliki.optyczne.pl/oly1817/oly17_fot13.JPG

The strong highlights that you see in the 20/1.7 image, e.g. in the upper left corner, are not present in the image from the 17/1.8 on the simple ground that the season, the weather, and the time of day is not the same. If you look at things that are reasonably comparable, I'd say that the bokeh of the 20 looks preferable to me, although I reserve final judgment until I am able to shoot the 20 and the 17/1.8 side by side as I did with the 20 and the 17/2.8 in the test I referenced above.

Based on what I know about the bokeh of earlier Oly WA primes such as the 12/2 and the 17/2.8, and based on the samples from the 17/1.8 that I have so far seen, I am afraid I am not that optimistic about the bokeh of the latter. I hope I am wrong because this is a lens I am interested in as an alternative to the 20 that I already have, but ...