It's interesting, sometimes a new product appears and for some reason everybody believes it is a dog. This is exactly what happened to the new Olympus M.Zuiko Digital 17mm f/1.8. There was one inital and very favorable review by Robin Wong, a positive but not completely enthusiastic review by Pekka Potka, a similar one by Ming Thein, and finally the absolutely condemning conclusion drawn by Lenstip.com. That last one nailed the coffin, now it's a dog. Is it? I've had the lens since Monday evening and all this talk about not enough sharpness made me take some test images.
The first was taken near the shop where I bought it. On the left side you see the JPEG straight out of the camera, on the right side a quick conversion in Lightroom. The thumbnails link to full sizes on Flickr.
This is f1.8, ISO 400 and 1/40s. The building is huge, thus although the focus point is far away, at f1.8 we are still not hyperfocal.
An aspect ratio of 4:3 is pretty unusual for me, but I really wanted an uncropped image to show you true corner performance. The result does not look in any way bad to me, but let’s have a look at a better example.
In this image I tried to shoot straight at the ceiling. This is a normal height office room, therefore DOF at f1.8 is very shallow. Lighting is uneven and in Lightroom I have tried to compensate for that and for lens vignetting. This and the fact that I didn’t expose to the right raises noise, and you can also see that I didn’t manage to hold the camera perfectly parallel to the ceiling. On the left side there is a slight magenta halo around the edges, in the right upper corner it is green, only somewhere in between it is exactly in focus.
Does this look wrong? Actually I don’t see a sharpness problem here. I know this is unscientific, I know this is not precise, but what I see is some sloppy technique, bad exposure, mild longitudinal CA, considerable vignetting and only good corner performance wide open. Technique and exposure are my fault, longitudinal CA is a weakness of all fast primes (with the exception of some obscenely expensive Leica lenses), and this leaves vignetting and corner performance.
The much praised Panasonic Leica DG Summilux 25/1.4 (mind you, praised by me as well!) also vignettes to a comparable degree, and actually I find the corner performance of this lens not bad at all. It’s not perfect though, but stopping down only a bit increases performance. As a worst case, I really can live with that.
In my opinion the real problem of this lens is not what it is, it is what it not is.
What this lens is not, is the kind of wonder that people seem to have expected: a 34mm lens on par with a 35mm Summilux on a Leica M. Sorry, it’s not. Of course the Leica is seven times more expensive and of course there is only a weak relation between lens performance and lens price. To those who want to focus manually on an outdated body design, using one of the best lenses in the market, well, it may be worth seven times more than the 17/1.8. Heck, they may even think it’s 20x as worthy and they may consider it a bargain. It is all relative.
Compared to its more likely competitors, the 17/1.8 fares well. It is faster than the Sigma 19/2.8 (which I never had) and it is an obvious upgrade to the old Olympus 17/2.8 pancake (which I’ve used a year ago on the E-P2). Its build is much ahead of both and obviously so is its price.
More difficult is the Panasonic 20/1.7. The Olympus focuses much faster and more silently than the 20/1.7, it is sharp, maybe not as sharp, but that’s hard to tell. I’ve sold my 20/1.7, thus I can’t compare. Overall the performance of the 17/1.8 is more even than that of the 20/1.7. For instance the 20/1.7 has a strong tendency to show excessive purple fringing. I don’t speak of CA here. CA can easily be corrected, purple fringing can not. The 17/1.8 is not completely free of purple fringing, but what I see is only mild, even under extreme conditions. I have no good example to show you, but I have tried, found it OK, and this also coincides with all the other reviews.
The Voigtländer 17.5/0.95 is a clear case. It’s twice as expensive, 4.5 times as heavy, bigger, manual focus only, much faster but with bad corners wide open and heavy CA and edge glow. If you need one you already know it, if you want one, you should consider its substantial drawbacks.
Compared to the 25/1.4 the 17/1.8 is lighter. Both are well built, personally I prefer the look and feel of the Olympus, but essentially it’s a draw. Of course the Panaleica has more bokeh, but that is due to it being slightly faster and due to the longer focal length. Sharpness? Again, both are sharp, I suspect that both are sharp enough to outresolve the OM-D’s sensor. Maybe the 25/1.4 more so, but that is speculation. We will know for sure in three years, with the next iteration of the sensor.
If all that is so, why the bad rap? Really, the 17/1.8 does not suffer from any of its few mild faults, it suffers from the fact that the 75/1.8 came as a sensation into a vacuum and that the 17/1.8 is only an excellent competitor in a crowded market.
Part of this post was taken from one of my comments in a Flickr group. The image of the “dog” was made Monday morning using the 25/1.4. Proper images taken with the 17/1.8 will follow in the next post.
The Song of the Day is “Puttin’ On The Dog” from Tom Waits’ 2006 album “Orphans”. Hear it on YouTube.