Find a Hardware Injection: Step 1
IntroductionThe LIGO data contains simulated signals known as hardware injections. For documenation on hardware injections, see links on the data quality page. This tutorial is based on work by Ashley Disbrow, and will focus on S5 CBC hardware injections, However, many of the techniques shown here can be applied to other types of injections as well.
This tutorial will show how to implement a matched-filter search to find inspiral signals in LIGO data. For a more complete description of matched-filter searches, see Allen et al. (2011).
Using the documentationLet's start by using the available documentation to pick a hardware injection we want to recover. From the S5 CBC hardware injection page, follow the link to the table of injections for H1. We'll pick an injection with a relatively high SNR for the tutorial. Scroll down until you see GPS time 817064899. You should see a line in the table that looks like this:
817064899 H1 10 10 25 Successful 28.16 26.55This is a simulated black hole/black hole merger, each 10 solar masses, placed at a distance of 25 Mpc from earth. The injection seems to have worked, and was recovered with a signal-to-noise ratio (SNR) of 26.5 using a low-resolution PSD. We're going to emulate that result. However, this tutorial will use a higher resolution PSD, so we should expect to recover the signal with a higher SNR.
We can also look at this time on a timeline, to confirm that the injection is marked in the segment information: (link to timeline). You should be able to see that the injection is present in H1 and H2, but not in L1.
Download the data fileIf you do not already know how to download and read a LIGO data file, you may want to start with the Introduction to LIGO Data Files. As a reminder, to download this data file, follow the menu link to Data & Catalogs to find the S5 Data Archive.
Find the injection segmentWe'll use readligo.py to open this file, a sample API described in the introductory tutorials.
Notice that one of the data quality channels is
Notice that all the time around the injection passes the default data quality flags (DEFAULT is 1), so we should have no problems analyzing this time.
Finally, let's grab the segment of data containing the injection. We'll also
identify a segment of data just before the injection,
to use for a PSD estimation. We'll call this the "noise" segment, and
it will be 8 times as long as the segment containing the injection.
You can download all of the code from this tutorial as find_inj.py. You can also download the code used to make the template as template.py.
What's next?Go on to step 2 of this tutorial.