Analysis of RAW samples at Fred Miranda show weak DR

[Mikehit]

I stand by the Bill Claff comment my point was his methods are consistent.

And what's the value of consistent testing of inconsistent subject matter, exactly?

As I understand it, each frame (maybe this I unique to Canon, I don't know) has a small border where no image is recorded but is used as a noise reference point to generate the image, and Bill takes data from that area which would mean that image content is irrelevant to how he is taking his measurements.

I wait to be corrected but if so it may inspire someone to give an idiot's guise as to how he does it.