Daaaaaaaamn! Check out the D5200 sensor!

Started Jan 24, 2013 | Discussions thread
Jack Hogan Veteran Member • Posts: 6,507
Re: Aptina?

Iliah Borg wrote:

PRNU (1/SNRtot^2-1/SNRshot^2)

Maybe we should analyse the applicability of the formula and underlaying assumptions.

If you mean how DxO calculates SNR data, I still am not sure, but if they are consistent hopefully it will not cause issues when comparing values at different ISOs.

Otherwise there are three main assumptions, I think:

1) the 3dB/doubling line is tangent to the curve in a spot where the noise is 100% shot noise
2) At 100% grayscale total noise is due entirely to shot noise and PRNU
3) At 100% grayscale Ntot^2 = Nshot^2 + Nprnu^2

The formula for PRNU from the other post is simply 3) rewritten keeping in mind that S/SNR=N, with S=FWC at 100%.  If one reads off DxO's graph above at 100% grayscale, SNRtot=41.9dB, SNRshot=44.2dB, converts to linear units and plugs them into the formula one gets PRNU=0.52%

Jack

Post (hide subjects) Posted by
cpw
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow