Minutes of 2008-Nov-13 S5 QPipeline Review Teleconference
Shourov Chatterji, Jonah Kanner, Isabel Leonor, Dave Reitze
Minutes by Isabel Leonor.
* Announcement Dave: Jolien has asked for report on review at one of the upcoming Burst Review Telecons. Dave told Jolien we are ready to sign off on review, pending some things which need to be finished. Dave does not think completion of the written report should hold up things for S5y1 paper. Jolien is okay with this. * Documentation of H1H2 coherent transform pipeline Isabel: Presented and went through the documentation of the qcollocatedtransorm.m code. This was something requested by Jonah to make sure we understood what the H1H2 coherent transform pipeline was doing. The link to the documenation is: http://www.uoregon.edu/~ileonor/ligo/qreview/review/qcollocate.pdf Shourov: Turns out that in code used for year 2 search, the coefficients for the components of the null and signal streams are the same as the coefficients used in the current code. No negative energies result in the year 2 code. All: Need to absorb and think about the documenation some more. Dave: Thinks this can be checked off the to-do list. * Post-processing scripts Dave: Jonah still looking at segments and detect scripts. Shourov says segments script is the same as segwizard. Jonah: Still need to do simple reality checks. Some more details to cover. Dave: Detect is the thing that checks agains injections? Jonah: Yes. Shourov: Detect looks at list of injections and triggers. Determines which injections have been seen. There are four kinds of outputs. Dave: Is there a table of missed injections? Have you done that Shourov? Shourov: I don't think so. Dave: Is there an injection your should've seen and didn't? Shourov: This can be done. It will be interesting to do so. Dave: You can say if they affect efficiencies. Sounds like work, though. Shourov: This would be beneficial to the burst group as a whole. We can use measure of SNR relative to a baseline noise curve, and say here are high-SNR injections which were missed. The reason might be, for example, that the injection is right at the edge of a segment. Or there was a glitch. Laura's notebook link has a table with -1, 0, +1 flags: # -1: this injection was not analysed (e.g., not in a science mode stretch); # 0: this injection was analysed, but not detected (i.e., below threshold); # +1: this injection was analysed and detected (i.e., above threshold) https://www.lsc-group.phys.uwm.edu/cgi-bin/bag-enote.pl?nb=burs5allsky&action=view&page=42 Dave: (To Jonah) Check of postprocessing scripts almost done. You just need to convince yourself that you believe them. Is this doable in two weeks? Jonah: Yes. * Revised sensitivity estimate Dave: Jonah sent around revised estimate which shows worse agreement between back-of-the-envelope estimate and actual sensitivity. The discrepancy is about 20% to 30%. Isabel, this is okay with you? Isabel: Yes, this is fine with me. * Table of loud events Shourov sent out link to table of loud events: https://ldas-jobs.ligo.caltech.edu/~shourov/s5qreview/triggers/s5/box.html Dave: Now sorted by correlated energy. Check the outlier's (red dot) correlated vs. coherent energies. Shourov went through first event in table as an example. Shourov: This is low-frequency event. Null stream not powerful enough in this case. This is an H1-only glitch that the null stream was not able to get rid of. Dave: Propose we all go through these and understand. * Other remaining things to do Dave: What about the plots Isabel requested? Shourov: Possible to do this over next two weeks. Dave: Also missed injection as a function of amplitude. Dave: Review document. * Dave: Meet next week. Week after next is Thanksgiving. Will send out doodle poll to check common availability.