Level-0 frames spanning 2 to 4 day windows from Dec 13, 2013 are used to 
construct time-dependent pixel-calibration products. About 6000 to 8000
framesets per window satisfying |b| > 15°, SAAdist > 5°, and MOONsep > 35° 
are combined to make various calibration products (see below).
NOTE: L1b (processed frame) performance trending can be found here.


* Cal-image montages (same stretch, labelled with dates and scan ranges)

  Relative responsivity maps (using the background-gradient method):
  Pixel range shown is ~ 0.97 to 1.3.
    FlatMeth1W1_montage.pdf
    FlatMeth1W2_montage.pdf
  Fractional delta-flats relative to first time interval, i.e. (t - t0)/t0:
  Pixel range shown is ~ -0.03 to 0.03.
    DeltaFlatMeth1W1_montage.pdf
    DeltaFlatMeth1W2_montage.pdf

  Relative dark/bias maps (trimmed stack-average of "L0 - med{L0}" images):
  Pixel range shown is ~ -10 to 500 DN.
    DarkBiasW1_montage.pdf
    DarkBiasW2_montage.pdf
  Fractional delta-darks relative to first time interval, i.e. (t - t0)/t0:
  Pixel range shown is ~ -0.2 to 0.2.
    DeltaDarkBiasW1_montage.pdf
    DeltaDarkBiasW2_montage.pdf

  Standard-deviation stack maps (trimmed stack-RMS of "L0 - med{L0}" images):
  Pixel range shown is ~ 2 to 6 DN.
    StdDevW1_montage.pdf
    StdDevW2_montage.pdf
  Fractional delta-StdDevs relative to first time interval, i.e. (t - t0)/t0:
  Pixel range shown is ~ -0.03 to 0.03.
    DeltaStdDevW1_montage.pdf
    DeltaStdDevW2_montage.pdf


NOTE: the following histograms are not updated on a regular basis.

* Relative responsivity map histograms
    FlathistosW1.pdf
    FlathistosW2.pdf

* Relative dark/bias histograms and zoom-in on tails
    DarkhistosW1_peak.pdf
    DarkhistosW2_peak.pdf 
    DarkhistosW1.pdf
    DarkhistosW2.pdf

* StdDev-stack histograms and zoom-in on tails
    StdDevhistosW1_peak.pdf
    StdDevhistosW2_peak.pdf
    StdDevhistosW1.pdf
    StdDevhistosW2.pdf
    StdDevhistosW1_peak w/ 12/5/13 "firstlight" scan
    StdDevhistosW2_peak w/ 12/5/13 "firstlight" scan

* Histograms of difference between StdDev-stack map used to flag noisy pixels
  in long-term static pixel mask and subsequent StdDev-stack maps
    rmsdiffhistosW1.pdf
    rmsdiffhistosW2.pdf
  Related plots showing same StdDev-stack differences vs dark signal map
  for each pixel
    rmsdiff_darksignalW1n.pdf
    rmsdiff_darksignalW2n.pdf
  Related plots showing actual StdDev-stack map pixel values vs dark signal 
  map pixel values for subsequent windows. Horizontal dotted lines mark
  25 DN -- the threshold used to create the pipeline static masks
    stddev_darksignalW1n.pdf
    stddev_darksignalW2n.pdf


L0 frame metrics from YEAR 1: 12/13/13 - 12/13/14

NOTE: the following (yearly-based) plots are not updated on a regular basis.

* L0 backgrounds (with 128 DN bias level removed) vs MJD. These are mode values
  within "clean" regions on each array. The horizontal lines are medians from 
  the end of original post-cryo and the red dots are binned medians. 
    L0bck_vs_mjdW1_yr1.jpg
    L0bck_vs_mjdW2_yr1.jpg

* L0 robust pixel-RMS values vs MJD. These are computed within "clean" regions
  on each array. The horizontal lines are medians from the end of original 
  post-cryo and the red dots are binned medians. 
    L0rms_vs_mjdW1_yr1.jpg
    L0rms_vs_mjdW2_yr1.jpg

* The number of hard/soft saturated pixels in L0 frames inferred from the DEB
  encoding vs MJD. The horizontal lines are medians from the end of original 
  post-cryo.
    L0sat_vs_mjdW1_yr1.jpg
    L0sat_vs_mjdW2_yr1.jpg


* L0 backgrounds (with 128 DN bias level removed) vs MJD. These are mode values
  within "clean" regions on each array. The horizontal lines are medians from
  the end of original post-cryo, the red dots are binned medians and the
  vertical dashed line is at the start of the survey-parameter change
  (~noon 01/08/2015 UT).
    bck_vs_mjdW1.pdf
    bck_vs_mjdW2.pdf

* L0 RMS vs MJD. These are robust pixel-sigma values (from a trimmed low-tail
  StdDev from the pixel mode) within "clean" regions on each array. The 
  horizontal lines are medians from the end of original post-cryo and the 
  red dots are binned medians.
    rms_vs_mjdW1.pdf
    rms_vs_mjdW2.pdf

* The number of hard/soft saturated pixels in L0 frames inferred from the DEB
  encoding vs MJD. The horizontal lines are medians from the end of original
  post-cryo.
    sat_vs_mjdW1.pdf
    sat_vs_mjdW2.pdf

* The number of hard/soft saturated pixels in L0 frames inferred from the DEB
  encoding vs L0 background level (in "clean" regions).
    sat_vs_bckW1.pdf
    sat_vs_bckW2.pdf


* Hot (high-dark) and noisy pixel trends using the above relative dark and
  StdDev-stack map histograms. Thresholds: "hot > 20 DN" in L0 trimmed stack
  average and "noise > 30 DN" in trimmed stack-StdDev. NOTE: these thresholds
  are not those used to create masks for production. The noisy pixel fractions
  are closer to what one would mask (i.e., declared "unusable" for all time).
  A large fraction of the hot (high-dark) pixels subtract-out nicely using
  either the time-dependent relative dark maps above, or the dynamic
  sky-offset calibrations. The hot pixel trends are normalized to the
  number of hot pixels in original post-cryo (at Feb. 1, 2011): 132,930
  for W1 and 174,976 for W2.

    hotpixtrend.pdf
    noisypixtrend.pdf
    hotpixtemp.pdf
    noisypixtemp.pdf


F. Masci, 01/27/2016