It's obvious that it's just a very soft lens at large apertures.
I dunno, there seems to be a fair bit of resolution there so it could just be the usual DPReview sample suck (IE they've got the jpeg sharpening set too low). I'd wait until we see some samples from a source which doesn't specialize in posting awful samples before making a conclusion.
Of course, it could be soft, but this is a $200 wide-angle so low expectations are encouraged.
Yeah my expectations are suitably low. But look at the brick wall shot (16 of 20), you can't screw that shot up, the centre is high contrast (and thus appearing sharp, full res image shows low resolution) but look at those corners! They're plain mush.
Anyway, I have been turned off the NEX5 because of the sheer lack of video options. I think I'll get a GF1 for similar money and use the brilliant 20/1.7 panasonic lens.
That might be a good idea - I didn't like the 3/5 interface at all either. And the lenses don't look that hot. Since I just bought an 850 and some lenses i bought a "better quality compact" to keep as the "to-go" cam until the seven surfaces. The 5 was a bit to "oversimplified" for my taste. If you want a toycam with the potential (but now, no practical way to realize it) for really great PQ, the the 3/5 cameras are really good for you. To be honest, I'd guess that meets about 90% of the customer demand which probably is what Sony cares about.
Regarding the Biogon you mentioned earlier, it has a quite extended rear pupil exit / focal plane distance - I would not hesitate to call it a lens with normal to moderate RF properties even if the optical recipe is labeled as "symmetrical". It's an elongated compression symmetrical type. The lens has a focal length of 25mm, and a exit pupil distance of almost 40mm at infinity focus. This makes it about as "hard" on the sensor as a truly symmetrical 40mm F/4.0 lens - not very hard at all.
