F80EXR vs F70EXR – Large versus Medium – Bright Sunlight

Another comparison. Very bright sunlight. Nice day.

Shot 1. A bird in a tree a hundred feet away at least. Both cams shot with my usual settings (P mode, etc) and both shot in large mode, something I have never recommended. But I am going to beat this to death today to show you why Large mode is folly with these very small sensor cameras.

It should be rather obvious that in bright sunlight, the small branches on the F80 image are really suffering. This comes from degradation caused by noise reduction. Just a few too many pixels on this sensor.

Shot 2: Branches. Shot at both M and L sizes to compare.

This is more interesting. The F80EXR is actually holding its own against the F70 with this much light. That should not surprise us, as extra light increases contrast. This defeats the selective noise reduction that Fuji uses to destroy low contrast detail, which it deems as noise.

But … it is blindingly obvious how big a difference there is between the upsized 5mp and 6mp images on the left and the 12mp and upsized 10mp images on the right. Fuji would have clearly had better image quality at the large size by binning *always* and then *upsizing* in camera when 10mp is requested. I.e. by *never* using the unbinned full-size data. Edges are better and low contrast details are *far* better.

This shows clearly the tragedy that is the EXR sensor … the technology is brilliant, but only when *binned* …. you get amazing dynamic range and you get fantastic noise characteristics – if, and only if, you keep the pixel count at the right level. 10mp works on the 1/2” sensor, and 12mp does not!!!

Image 3 … the unholy mess.

Here, we peek into the shadows a bit. You will want to click through on these to see the full extent of Fuji’s disaster. Have some tissues handy.

The F80 again acquits itself very well at M size … the detail in these shadows is just fine. But interestingly, the detail in the upsized F70 image (I did apply a small amount of sharpening to compensate for the upsize interpolation) is just about identical. So why was it we needed the extra 2mp again? Oh yeah … the marketing department thought that they could sneak one by us and advertise a 12mp camera … all the while destroying its edge integrity and noise characteristics in difficult light. Good call :-(

On the other hand, here again the F80 at large size (and why else would you want to add the extra pixels?) simply “shits the bed”, to use the crude lingo of the teenagers. My god, what a horrid mess. Did they really *never* test the camera? Or can they actually think that this kind of mess is acceptable? The F70, by the way, sucks too … but much less than the F80 at L size.

My conclusion so far is pretty simple. I like the F80 in great light at 6mp. And only then. It sucks in difficult light. It sucks at L size. Totally, I mean. And the difference between the F80 and F70 in great light is negligible, but the F70 walks all over the F80 in any other light. The F70 remains Fuji’s best compact camera ever … in my opinion of course.

*sigh*