alundeb Offline Upload & Sell: Off
|
p.2 #16 · Canon 1D X: The best of Canon’s full-frame sensors | |
PhilDrinkwater wrote:
What surprises me all of the time is the overall ratings, not the graphs. I've always found them confusing.
Overall scores:
1dx 82
1ds3 80
Look at the graphs of these cameras. The 1dx is a lot better or a bit better at everything except a tiny reduction in DR at ISO100 and that's only worth 2 points on their overall scale?
Now add the d800 in. The 1dx is pretty much as good at everything - better at high ISO, worse at low ISO DR, small difference on low ISO colour sensitivity - and the d800 gets a score of 95? So they are massively biasing low ISO DR.
Now look at the sports scores. The D800 sports (low light ISO) *beats* the 1dx - 2853 vs 2786 - when clearly you can see that the 1dx is massively better at high ISO. The SNR graphs are the same and the 1dx has equivalent DR at ISO1600 and wins from there.
Now remove the 1ds3 and add the d3s. It has the same score (82) as the 1dx and the same profile on pretty much all the graphs (although the 1dx wins a bit on high ISO after 3200), but it has a sports score of 3253 ISO - which is more than the 1dx. (unless you look at the colour sensitivity graphs maybe)
Huh?
I'm not saying their numbers don't mean anything or they're made up and I'm sure they have a formula, but if you look at those scores and take them at face value, you'd buy a d800 since what they're effectively saying is that it's better at high ISO than the 1dx.... and it's not ... and you'd buy a d3s since it's also better at high ISO .... and it's also not unless you look at the colour sensitivity graphs.
I just don't get where their numbers come from sometimes.
The overall core should be ignored beacuse it has a weighting.
The low light ISO score should also be ignored because of the way it is calculated. A limit is defined for 3 different measurements. What ever measurement that goes below the limit first, defines the low light ISO. This doesn't match with reality, because in reality you can trade the noise vs color in processing. If the low light ISO were to mean anything, it should be calculated as a weighting between SNR, DR anc Color Sensitivity.
I am not defending the scores, they should be ignored, but just if you wonder how they get there....
For the lens tests, I have no idea how they get results that don't match other test sites or experiences. IMO, DXO ar capable of measuring sensors, but not lenses. Their horrible lens measurements should not be used as an argument against their sensor measurements.
Back to the 1DX, it really has the best high ISO Dynamic Range in the world. The read noise at high ISO can be calculated back to 1.3 e-, and that is the lowest read noise for any DSLR to date.
|