Kermit Gain

Looking back through all the GB of Kermit data, I have three different sets of variance test data, from March 2004 through May 2005. I just re-analyzed them and I get the following estimates for the gain:

2004.03.10      2.31984 ± 0.057045

2004.06.10      2.53101 ± 0.259411

2005.05.18      2.28919 ± 0.093442

The June '04 data is noisier for some reason, and some of the data points don't seem to fall on the expected linear relationship between mean and variance; if I throw those files out, I can get a gain of 2.28 ± 0.10, which is more in line with the March 04 and May 05 data.

So empirically, it seems like the Kermit gain really is around 2.3 e-/DN ± 0.1.

This page last modified on 2005-12-07 by Marshall Perrin.