Originally posted by Adam Makes sense: (approximately) the same pixel density + newer technology.
That's not exactly what I meant, you shouldn't jump to conclusions so quickly (e.g., you ignore sensor size and don't appreciate that pixel density is irrelevant):
If I scale ((*)sensor size normalization) the D7000 and K-5 results to the D800 result (assuming they all use the same pixel architecture), I get:
Measure: D800 / K-5(*) / D7000(*)
Color depth: 25.3 / 25.5 / 25.3
Dynamic Range: 14.4 / 14.7 / 14.5
Low light ISO: 2853 / 2685 / 2696
So, you see that the pixel technology in the D800 is almost identical, matching the D7000 overall within a tiny +/- 0.1EV error margin. That's not a newer technology, it is exactly the same deployed on a larger sensor.
So, what DxO is telling us is simple: the D800 has 36 million of the exact same pixels the K-5 has 16 million of. In retrospect, the D800 score was trivial to predict because full frame adds 15 to an APSC score and the D7000 was at 80, the D800 is 95 indeed.
Whatever Pentax did to the D7000 sensor, a Pentax D800 would probably have scored 14.6 EV in the Dynamic Range measurement.
That was what my comment was about, not this trivial stuff ... :ugh: