cgarcia Offline Upload & Sell: Off
|
RobDickinson wrote:
hm that sux, not sure that image is clipped both ends but not sure you need that.
can provide more raws if it helps?
Yes, clipping is useful to determine the white point. Numerically it seems 16383, which is the maximum possible 14-bit value (a bit rare at ISO 100).
Anyway slight variations on the white point doesn't affect a lot the measured dynamic range, e.g.:
For 16363 -> LOG2((16383-512)/7.4878) = 11.04 EV at 26MP
For 15360 -> LOG2((15360-512)/7.4878) = 10.95 EV at 26MP
Being 512 the black point and 7.4878 the read noise (to measure it we don't need a dark photo with the lens cap on because the raw files internally provide the black pixels).
However, a ISO series of photos (one at each ISO level) would be useful to measure the high ISO performance. In case it is not stellar, we would assume that Canon has crippled the 6D2 to further differentiate it from the 5D4.
Edited to punctualize:
I previusly mentioned a read noise of 7.53556, which was referred to both green channels combined.
The 7.4878 is a green channel alone (the other green gets 7.58229, red is 7.52778 and blue 7.49653).
So all channels behave nearly the same (only the 5D4 was a bit odd on this matter).
|