Upload & Sell: Off
| Re: Got the 1DX: Shadow recovery/DR sample |
[Come of think of it: Can pattern noise be more pronounced when using very fast glass as F1.2/F1.4?
Yes. When using fast glass, there is a light loss due to oblique rays that do not hit the photodiode. Since a few years ago, it is known that the camera compensates for this behind the scenes by amplifying the signal. This will bring out the read noise and FPN. There is an article about it over at dxomark.com
I didn't know that, but then please keep in mind that the test shot here was done with a 35mm F1.4L shot wide open @ F1.4. Does this influence the result?
Was it just the actual shot that used the 35 1.4 or the black frame too?
Did the black frame have the body cap on and no lens mounted or did it have a lens capped 35 1.4?
The last. Oooops?
OK, that improves read noise a tad, it now measures an actual 5.29 (instead of 5.79 with the filthy lens on! SERIOUSLY how could you THINK to take a snap with a.... LENS?!!?? mounted on the camera?!! ) for about 11.87 instead of 11.75 stops DR, not a huge, difference, but hey. So call it about 11.9, maybe 11.8 is the well number is lower than expected or 12.0 if higher than expected. So this brings it to a measured level about same as the 5D2, almost same as the 1Ds3 and a little bit better than the 5D3, although all within 1/3 of a stop probably means they wouldn't be that easy to tell apart from a practical standpoint, well maybe my low outlier at 11.6 for my particular 5D3 copy would just. The real story is in the banding. I need to look at the 1Ds3 again, the true body cap black frame made the banding seem a little worse than I thought for the 1DX but it's still better than any of the recent cams, absolutely, especially than the 5D2/50D/7D and even 5D3/1D4, but maybe not quite as good as the 1Ds3.
I'd still say the 1Ds3 wins as best ISO100 from a Canon DSLR. 1DX takes second.
1DX should easily 1Ds3 around ISO320 and up and easily so ISO800 and up.
(why the difference with the lens? it was a fast lens, faster than f/2.8 and most makers apply secret extra levels of ISO gain beyond what the claimed levels are when shooting with fast lenses to make up from sensor micro lenses not grabbing as much light as film would from super fast lenses, so instead of shooting ISO100 maybe you are really doing ISO118 or 130 or something, etc. the first black frame had a 1.4 lens on the second one did not and thus lower read noise, but OTOH signal wasn't lifted quite as high above the banding and the lesser read noise let it show more so the banding seemed a little worse than original thought, but it's still better than the recent heavy banding junk they have released for the past few years)