I just did an 18 sub ImageIntegration and see this warning logged:
No convergence in MRS noise evaluation routine - using k-sigma noise estimate.
Below that in the log I see:
Gaussian noise estimates:
?s = 2.205e-004
I ran Script>NoiseEvaluation on the integration and get this:
* Channel #0
?K = 9.520e-005, N = 65045 (3.12%), J = 3
The integration visually appears to be OK.
II parameters were what I usually use for lights: Average, Additive, Noise evaluation, Winzorized Sigma Clipping, Scale + zero offset.
What is the meaning of the different ?s and ?K measurements and what is the significance of their differences? I would like to learn more about this.
Is there any way of telling why convergence failed? Note: when I integrate only 7 subs (using Percentile Clipping) rather than all 18 no warning appears.
I attached a copy of the log.
Thanks,
Mike
