utils: calculating gain and readnoise methods don't agree

[stharrold]

As of 2014-08-18, gain_readnoise_from_master and gain_readnoise_from_random do not agree. Test with IRAF.

[stharrold]

• for test data on UWCrb, 2014-05-30:
  • from master bias, flat:
  • gain: 0.076 electrons/ADU
  • readnoise: 0.28 electrons
  • from random bias, flat:
  • gain: 0.067 electrons/ADU
  • readnoise: 0.78 electrons

[stharrold]

from ProEM certificates of performance (traditional amplifier, 1 MHz readout speed, high gain), should be:

• gain: 0.84 +/- 0.05 electrons/ADU
• readnoise: 6.1 +/- 0.05 electrons RMS

[stharrold]

use numpy.ravel before calculating?