Saturation point seems to vary depending on the subject/lighting



I've calculated the headroom of my highlights based on the tutorial on your website.

I'm getting inconsistent results though i.e. sometimes I'm clipping the highlights.

Several times I have done the headroom test and have measured my saturation point to be 13,284 (Green channel after the black point has been subtracted).

On a canon 5D3 I have done a few tests which measure around 3.5 - 3.7 EV headroom so I am using metering for + 3 1/3.

Today I redid the headroom test to try to figure out the problem (my subject was a super bright white painted timber wall in direct sun). I metered +3 1/3 stops and it still was clipped.

Max value in green channel is 13,672. I've also done some other tests where the saturation point is different again.

I'm using fixed ISO100 and WB5500. I've also tried with Uni WB.

Indoor tests seem to work as expected so perhaps flare is influencing things? I'm often shooting outdoors in bright conditions....

Some cameras may change

Some cameras may change saturation (clipping) point based on in-camera (temperature?) calibration. This change is not big.

13672 vs 13284  is only 1.03 times or 0.04 photographic stops, so you may just ignore that difference (and assume camera maximum to be 13000).

Also, please make sure, that max value is really 'hit the wall' on the histogram, not several pixels above the wall. Newest Canon 5DS may have some pixels 'above the wall' (I've seen it on some samples), not sure about 5D Mk3.


Adaptive noise filtering?


Some questions regarding noise filtering:

Is the effect of adaptive noise filtering applied to the originally captured data before writing the data to the card in the camera detectable, e.g., by a characteristic signature?

Would it be possible to find out whether a camera manufacturer is applying adaptive noise filtering by comparing images taken by different cameras that otherwise have the same sensor (but perhaps with different CFAs) under standardized conditions?

Do you know of any known camera where adaptive noise filtering is applied before writing the raw data?

Add new comment