Nikon D2H full review posted, with sample images from a production-level camera!|
(Friday, December 19, 2003 - 11:03 EST)
If you've been wondering why so few review postings on the site lately, here's at least a partial explanation for it: My in-depth review of the Nikon D2H.
Given all the foment in the online community about image noise with this camera, I wanted to take particular pains to analyze exactly what was going on in the camera across a range of ISOs and light levels. The results were interesting to say the least, and one of my most surprising findings was how poorly numbers representing absolute noise levels correlate with visual perceptions of the test images.
I've about come to the conclusion that simple magnitude numbers for image noise are all but useless for comparing cameras to each other. Numerically, the D2H competes very well with its primary adversary, the Canon EOS-1D. On a purely visual basis though, the EOS-1D seems to win the noise derby most of the time under good lighting. (The D2H wins handily in low-light conditions though.) I also looked at color saturation and resolution between the two cameras in quite a bit of detail.
At the end of the day, I ended up with somewhat mixed feelings about the D2H. On the one hand, it's by far the easiest-shooting d-SLR I've used to date, with a user interface that should be the envy of the industry. It also has some really outstanding features, in its Wi-Fi capability and the amazing SB-800 flash system. Its noise levels under good lighting and its color saturation at high ISO leave it wanting though.
I personally seem to be less bothered by image noise than many folks, so perhaps don't down-rate the camera for its noise characteristics as much in my own mind as a lot of other people do. (Rob Galbraith in particular recently changed his opinion of the D2H's image noise, now calling it "unacceptable" at middle ISO values.) I don't know that I'd go so far as to call the D2H's noise levels unacceptable, but I certainly wouldn't mind seeing less of it. (As I alluded to above though, the issue isn't so much the *amount* of noise, as it is its *structure.*)
Given how personal the issue of noise seems to be, I'll refrain from drawing any stronger conclusions than this, but instead direct interested readers to my review for all the absurdly detailed discussion, complete with charts, graphs, and rollover images. Check it out! (This reviewer will now resume his "life," normally scheduled to appear in this time-slot. ;-)