We can be thankful it's not $2390.
The DXO score apparently uses centre resolution only, which explains why the Sigma 50 and the Samyang 85 have DXO scores that are higher than better lenses. Mostly you can't see the difference between the resolution of different lenses in the centre, but you sure can in the corners! Of course, the corners don't matter for portraits, but neither does the centre.
RichRMA: The odd thing is, when I look at the RAW images from the test set-up, the Fuji doesn't seem to resolve as well as some of the competition, at least until very high ISO where its low-noise helps.
The ACR Raw output is more blurred than the out-of-camera jpeg, which is itself not as sharp as the NEX7 or EM5.
If it's not possible for a raw converter to produce sharper pictures than Fuji's out-of-camera jpegs, I'd say this camera and the XPro1 are for people who especially hate moiré (or for shots that the photographer knows will be badly affected by it) and are prepared to sacrifice resolution. The AA filters of the Bayer-sensor competition are never strong enough to eliminate it. One strong enough would probably produce pictures like these. A better solution for good light is the 15 Mp Foveon. That leaves the Fujis for moiré-prone pictures in poor light.
A waste of bandwidth, money. and power. There's no point in having pixels any smaller than the angular resolution of the human eye. There's no point in a huge screen either. Positioned 50 cm in front of a 150 cm-wide screen, I'd have to get out of my seat to examine the corners in the detail provided. It's more convenient to position the appropriate part of the picture in front of me with a mouse, in which case a screen 50 cm wide or a little larger will do.
(unknown member): Dpreview why do you post jpgs for samples for reviews of lenses?!
It's bad enough for camera reviews but for lens reviews too?!
And how about standardizing some shots that will help the reader see for him or herself the edge to edge performance of the lens at different apertures and focal lengths for zooms?
In other words, make the reviews much more useful to your readers instead of trying to impress people with technical charts and diagrams!!!
After shooting my mouth off I read the "Image Quality Tests" section of DPR's 5DIII review, which indicated that its out-of-camera jpegs were poor (either mushy or over-sharpened), that good results can be obtained with Adobe Camera Raw, and suggesting appropriate sharpening settings. Having worked all that out, I think it would be better for DPR to process the image samples that way for for lens tests using the 5DIII. The final output is a jpeg, Basalite, but processed better than the 5DIII is able to do by itself. The problem is not the jpeg format, but the jpegs that come out of the 5DIII.
Hi Barney,Sorry to be rude about those sample pictures, but if that's the Canon 5DIII standard sharpening setting, it's far too aggressive, and producing sharpening halos strong enough to make a mess of the detail. One or two sharpening levels down would give a better picture for testing a lens, I think. Another alternative is to use no sharpening at all and allow us to do it to taste, as (only) Lenstip does. This has the big disadvantage of assuming that we understand why the pictures are blurry, have software to sharpen jpegs, and can be bothered to use it. I wish you had used a D800E for the samples!
Unsharpened jpegs (as per Lenstip) are OK for lens test samples, but the over-sharpened ones here are horrible and useless.
pixel_peeper: There's no bite to any of the pictures I have seen from the XPro-1, with the same random-array sensor as the X-E1s. They all look as if they've been shot at f22. I put it down to the sensor, which appears to give significantly inferior resolution to equivalent Bayer-array cameras.
If you all think the latest Fuji sensor is sharp, have a look at the last picture on this page. http://www.sigma-dp.com/DP2Merrill/samplephoto.htmlThe Fuji idea was a flash in the pan.
There's no bite to any of the pictures I have seen from the XPro-1, with the same random-array sensor as the X-E1s. They all look as if they've been shot at f22. I put it down to the sensor, which appears to give significantly inferior resolution to equivalent Bayer-array cameras.