Daaaaaaaamn! Check out the D5200 sensor!

Started Jan 24, 2013 | Discussions thread
Jack Hogan Veteran Member • Posts: 6,333
Re: Daaaaaaaamn! Check out the D5200 sensor!

dethis2 wrote:

D5200 has a tiny bit higher SNR18% from D7000 so it should have a tiny bit higher QE.

Hm, gotta think about that, because it's not true at base.

D5200 has a much lower FWC than D7000

D5200 -> 27517

D7000 -> 49058 which normalized to 24Mp is 16/24*19058 = 32705

Except that the Sensorgen D7k data is a bit of an oddball: I get an umodified FWC of 38300 from the DxO SNR curves, bclaff, who averages results from several cameras, gets 37200 (Sensorgen's DR is also correspondingly off and it looks wrong).

Using my extrapolations from DxO data for the D5200 and D7000 I get 24MP equivalent FWCs of about 26500 and 25800 respectively. However, as Steen Bay mentioned, the D5200 is less sensitive, so its Average Absolute QE suffers.

I inspected the raw histograms with rawdigger. At all ISOs >=200 the green histograms (which are not affected from WB preconditioning) have gaps every 5-6 levels. The only solid histogram is the one for ISO100. Counting the gaps I find a repeatable pattern of 4 gaps every 23 raw levels so we have a division 23/19 = 1.21

It looks like ISO100 is ISO121 compressed ..

I would look at it the other way around: base is base and the data is not scaled. On the other hand, in order to achieve higher ISOs part of the 'gain' is analog, and part is digital. It shouldn't really make any difference to our calculations, however, because with digital scaling the statistics stay the same.

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