Changes between Version 8 and Version 9 of MagicSummaryFeb2010

Timestamp:

Feb 12, 2010, 6:40:32 PM (16 years ago)

Author:

eugene

Comment:

--

MagicSummaryFeb2010

@@ -9 +9 @@
 We have also attempted to address other sources of magic false positives, in both the IPP where possible and in the magic software itself.  Over the past two weeks, we have run three iterations of a single test suite to validate magic.  The test suite included data representing all of the survey modes.  Below are plotted histograms of the number of magic streaks found for each of the three test runs, also broken down by the survey type and the filter.  Note that the intermediate analysis used only a subset of roughly 300 exposures, while the first and last iteration used the complete set of roughly 750 exposures.  Also note that in these plots, the horizontal axis uses bins of '''log10(streaks)'''. 
 
+Notice that the streak numbers depend greatly on the survey type.  For the most part, the lower the sky background and the higher the density of stars, the more sensitive will Magic be to the instrumental artifacts and persistence structures.  In addition, the area near the Geo belt (-3.5 Dec as seen from Hawaii) will have a higher density of real satellites.  For the full set of data in this test suite, the final configuration yielded a median of 49 streaks per exposure and a maximum of 279.  
+
+'''After the last test set, we missed a streak.  The Magic team adjusted the detection threshold just slightly (to 2.35 sigma from 2.4 sigma) to enable that detection.  As a result, the expected streak rates are ~15% larger than shown here.'''
 
 magicverify.%.3.% = first test