[argyllcms] Re: Estimating Gamma
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Wed, 15 Jan 2020 12:31:20 +1100
graxx@xxxxxxxxxxxx wrote:
It was suggested to me that, in order to estimate (monitor) gamma, I simply
needed to take
the “Log” of the both the Input and Output and it would yield a straight
line. But that’s
not quite what I’m finding, in the following Excel sheet, see my measurement
data and
gamma (transfer function) graph in Log form. I don’t get a “straight line” :
It's only a straight line if the curve is exactly a power curve.
A more typical plot is log(Y)/log(X) vs. X though, which ideally
would be a horizontal line at the gamma value.
(Note of course that you have to omit the 0.0 and 1.0 values).
Non power curve shapes generate log/log graphs that are not single value.
So to estimate gamma, you need to pick a point on the graph or
average it in some way.
[ This is not how I like to estimate gamma. The way I like to do it is
to find the power that has the same 50% response as the curve you are
estimating. This is because it gives an overall impression of brightness
that is similar. ]
Cheers,
Graeme Gill.
Other related posts: