|
I'm not knocking Canon at all - I own a couple and the 6D is undoubtedly a fine camera.
I am just saying that you cannot really be sure how low the noise of any DSLR chip is by just looking at some stretched RAW darks - the data may well have already been processed in a variety of ways that make it look better. The Stark paper suggests that both gain and offset modification are employed in the 450D - who knows what they do in the 6D or any other modern DSLR? After all, the new processors can implement a wide range of operators and would seem to have enough power to run almost any of the standard noise reducing algorithms.
It is probably more meaningful to compare dark variance and hope that no adaptive smoothing has been applied in low signal regions. Or maybe just take some side-by-side astro images with identical pixel scales and compare the results.
Last edited by Shiraz; 17-12-2012 at 06:09 PM.
Reason: change camera type
|