Using process console info to compare integration methods

[astrodoc71]

Let's say you integrate with linear fit and windsorized sigma and compare. In the process console there is this parameter "SNR estimates". If WSC yields a smaller number than linear fit then that result has better signal to noise? Or is this an oversimplification?
Thanks
Dave

[RickS]

If you're looking at the SNR number then a bigger number is better.  It's signal to noise ratio and we want more signal and less noise    But, yes, this is correct.  You can use the SNR estimate or the median noise reduction to compare integrations of the same data.

[Juan Conejero]

Hi Dave,

SNR estimates are orientative, but the most reliable way to evaluate integration results is to compare effective noise reduction (ENR) estimates. For a description of ENR see the section on Quality Assessment of the reference documentation for ImageIntegration.

So in essence, you should try to get the highest possible ENR estimates with the required rejection of spurious data (plane trails, cosmics, etc). When comparing the results achieved with different pixel rejection algorithms, make sure to set rejection parameters so that the amounts of high and low rejected pixels are as similar as possible.

[astrodoc71]

Thanks Guys! Very cool stuff indeed
Regards
Dave