# Photon Transfer Curves (PTCs) at PhotonsToPhotos

Started 4 months ago | Discussions thread
Re: Calculating digital gain from histo gaps

Jack Hogan wrote:

Ah, I see. I had missed the fact that both gapCount and span are defined differently than how I used them in the formula/question above - and I can see that Jim's formula would be more accurate than my approximation if one knew the start and end of the sequence.

@ Jim .. BTW .. I think a more compact and faster calculation is

gain = span / (span - gapCount) .. one division less .. result in our example .. gain = 10/(10-1) = 1.111

And given the fact that often gaps are not evenly spaced and we do not have access to data over the entire histogram, shouldn't the span simply be an arbitrarily chosen roi where we are sure that gain remains constant?

So for instance in the case of the D850 at ISO1250 there are about 30 gaps between level 601 to 800 in the red channel so its gain should be approximately 1 + 30/200 = 1.150? If more precision is required increase the span?

Jack

We have to calculate on full cycles .. by choosing an arbitrary ROI we can fall on cases where our ROI just misses gaps before after or just includes them

Yes, but how do you know when a full cycle begins (or ends)? Gaps can occur at multiple frequencies. So my suggestion is to choose a large enough span for our purposes and simply count gaps within it. Being mindful of not crossing into an area of the histogram with a different sequence of gaps.

Jack

A full cycle begins or ends at gaps' locations .. no ?. Jim locates all gaps and defines "span = zeroLocations(end) - zeroLocations(1)" .

Fair enough, I was thinking how to figure out the start/end of a fairly typical sequence such as this for the D850 at ISO64, but the error should be acceptable:

6 levels, gap, 6, gap, 5, gap, 6, gap, 6, gap, 5, gap, 6...

Jack

I was reading the new Hasselblad article - they seem to calibrate each camera individually. Maybe those gaps are due to calibration (I think that there is agreement that it is due on some rescaling)? Are they the same on another camera, same model?

Complain
Post ()
Keyboard shortcuts: