coherent WaveBurst Review Meeting 30 July 2009 09:00 Pacific / 12:00 Eastern
Agenda: Thursday 30 July 12:00 Eastern
- Updates on action items:
Dial-in number: 1 800 704 9896 International dial-in: +1 404 920 6472 Conference code: 5374 2349 #
AttendanceMarco Drago, Patrick Sutton (minutes), Igor Yakushin, Virginia Re, Romain Gouaty.
Segment lists and livetimes for LFS
- PJS showed page comparing stated livetimes to those estimated from counting injections, tests by IY. Satisfied with segments and livetimes; need to do the same for HFS.
Missed Loud Injections in HFS: Injection detailsLoud missed injection follow-ups for HFS: Looking at MD and RG's pages; MD's first.
- "Not reconstructed" means there was not trigger above trigger-generation threshold in analysis job (as opposed to having a trigger that is cut in post-processing).
- H1H2L1: Not reconstructed events (0.3% of total) are correlated with low network antenna alignment factor "epsilon". Also correlated with
low overall antenna response "g".
PJS: Are epsilon, g computed using noise spectra? MD: No, only using antenna responses.
- Scatterplot shows SNR in H1 very low wrt L1. Consistent with failure of consistency test.
- H1H2L1V1: Much fewer "not reconstructed" events (0.04%). Again,
correlated with low epsilon. Not clear if correlated with low g.
PJS: Speculate that these injections are aligned with cross polarization by bad luck.
- RG: Asked about 4-detector scatterplots.
PJS: 8-10 injections with SNR in H1 > 200, L1 > 100, V1 < 20-ish. At what point in the pipeline are these cut?
MD: Maybe cc cut.
- RG: Give GPS times of these and I can check to see if the data was bad at these times.
Action item: RG to run omega/q scan of these times from MD.
Missed Loud Injections in HFS: Q scansLooked at RG's page of q-scan and data quality flag checks for loud missed injections (LMI).
- H1H2L1V1: Total of 26 LMI in 4X time; all checked by RG.
- All scans are on LIGO data only; unable to get V1 data run. Also missing some channels for LIGO, such as seismic channels.
- Data is not especially glitchy. No clear data-quality related reason for most of these injections to have been lost. Could be related to Virgo data since not available for scanning.
- 5 of the 26 LMI had the H2 bad month flag.
- Yellow: flag precedes injection.
- Red: flag coincides with injection.
- Green: flag lags injection.
- Will rerun with higher frequency range where possible, since comparing to HFS triggers.
- H1H2L1: About 300 LMI. Scanned all, regardless of antenna responses. Most may be explained by antenna responses.
- Some LMI due to glitches, but not the majority. E.g.: 849704762 -- no red flags but H1 glitch. Also 873026063 -- L1 glitches.
- RG: Want to add seismic and other auxiliary channels to these scans.