Originally posted by DonDouglas I tend to rely on Imaging Resource Imatest results
The two dynamic range measurements aren't directly comparable. DxO has a better justified procedure.
If at all, the DxO figures would correspond to the Imatest figure labelled "Low":
1.0 (imatest low): SNR=1 (0 dB)
0.1 (imatest high): SNR=10 (20 dB)
0.03 (dxo high): SNR= 30 dB
So, subtract something like "4" from DxO to have figures in the same ballpark as imatest.
However, the imatest figues have 2 major flaws which is why the DxO results are more trustworthy:
1. Imaging Resource applies ACR with default settings which applies demosaicing, tone mapping and some NR. Not having an own raw converter, it's the best you can do. I do the same at lumolabs (except I don't use default ACR settings. I rather use flat ACR settings which I have verified to be much closer to raw). But DxO has a grip on the data prior to demosaicing. Therefore, the numbers don't match and the DR ranking at Imaging Resource depends on the version of ACR/LR. DxO figures are software-independent.
2. Imaging Resource doesn't correct for amount of pixels. DxO gets this trivial prerequisite correct which tells me they know better what they do. Imaging Resource seems to run the Imatest software w/o really understanding what's going on.