Tuesday, December 20, 2011

Fuji X10 vs Olympus XZ-1 vs Panasonic LX5 vs Canon G12 – How close are they at 1600 ISO?

The previous post analyzed these cameras at base ISO. The XZ-1 the easy winner. The X10 the surprise loser. Moiré and lens softness plague this sample. We’ll stay tuned for Fuji to send DPReview a hand picked sample with a better lens (hopefully.)

Meanwhile, a quick repeat of the analysis at 1600 ISO. This is an ISO that allows shooting in low-ish light with reasonable detail retention on recent small sensors. Not great, but certainly better than a few years ago.

So let’s see how they fare … the larger sensor of the X10 should make it the winner this time round, as the moiré will vanish in the sea of noise these sensors generate.

Crop 1 – Coin and Dog

image

The Fuji gets a first again on this crop, but this time the moiré is much more subtle (but not completely gone you should note) and the detail retention on the dog’s nose and the little puff balls is sufficient for the easy win.

Order of goodness: X10, LX5, G12, XZ-1

Crop 2 – Face and Wool

image

The Fuji still can’t render the lines on the face, which both the LX5 and the G12 manage to almost do here. But the wool is clearly the class of the field in the X10 image. And the face is decently clear. So the X10 gets another lone first.

Order of goodness: X10, G12, LX5, XZ-1

Crop 3 – The Watch Face

image

The Fuji jpeg engine is smearing the ticks in the face of the watch. The others are worse, except for the G12, which does a great job of balancing the NR against the detail. Canon wins this crop.

Order of goodness: G12, X10, LX5, XZ-1

Crop 4 – The Globe

image

This one is interesting. The softness shows again. I thought it might be masked by the general softness of 1600 ISO, but no … that’s not how this works. It is obviously additive.

Order of goodness: G12, LX5, X10, XZ-1

Crop 5 – Paper Clips

image

As I’ve noted in a past article, the XZ-1 has a strange affinity for these paper clips. The Fuji is again plagued by that contrast loss, which provokes the jpeg engine’s NR to get medieval on this image.

Order of goodness: XZ-1 (!), LX-5, G12, X10

Result Table – no ties this time, so the numbers will add up …

  X10 XZ-1 G12 LX5
Firsts 2 1 2  
Seconds 1   1 3
Thirds 1   2 2
Fourths 1 4    

Overall winner was the G12. I thought, as we went along, that the X10 would sweep to an easy victory. But the G12 has a great jpeg engine and the X10 a weak lens. So again, this copy had trouble. But judging by the demosaic issues with the X10, the G12 will continue to compete.

Of course, the X10 fans will quickly denounce this test because the X10 has a faster lens. You can shoot at lower ISO they will say. And that is true … up to about 112mm. And then you have to start cropping, and with the detail issues in the demosaic algorithm, that’s not gonna be pretty. There is not much room for cropping with these sharpness issues.

This is all, of course, a jpeg issue. I’m sure that RAW, once the files are available, will tell a slightly different story?