LSC-Virgo Burst Analysis Working Group


Burst Group Home
ExtTrig Home
CBC Group Home
LSC, LIGO, Virgo


'How-to' docs
[2008, 2007, earlier]
Talks/Posters [pre-wiki]
Paper plans
White papers


Analysis projects
Old notebook [General, S2, S3, S4, S5]
Virgo workarea
External collabs


Main review page

S5 Qpipeline High-Frequency Review Telecon 24th of July 2008 11:00 Eastern

Agenda and minutes of the S5 Qpipeline High-Frequency Review Telecon for Thursday the 24th of July at 11:00 Eastern time.

Agenda and Contact Info


Action items
  1. update on astro waveforms
  2. Table of review items
  3. AOB

Contact Info

  InterCall telecon service:

    USA (toll-free): 1-866-616-1738
    UK (toll-free): 0800 073 8914
    Italy (toll-free): 800-906-494
    Germany (toll-free): 0800-1014-907

    participant code: 251 288 9495#


Minutes by Jonah

July 24, 2008

Michele walks us through his matlab script
for processing astrophysical waveforms.

Issues include:
Converting sampling rate
Working out right amplitude scale

Conversion specifications are comments 
at top of this m file

Waveforms have "inspiral phase" and "burst" phase
NS has some angular velocity - when it collapses,
a different radius leads to a different angular velocity -
this is the "burst".  This part of the signal should be
isolated, since that is what our search is sensitive to.

* loads the data
* resamples the data from 203 kHz to 16kHz
* apply a low-pass filter with a 6000 Hz cut-off

The 6000 Hz filter is used to keep energy in the 
search space

The second and third blocks are to study features of the 
filter - phase shift, numerical errors, etc.  They 
show consistent results.

* Remove the mean value of the waveform to
get rid of step function at ends
* Apply hanning window, which gives minimal 
distortion of injection spectrum

Siong:  Why window now?  Doesn't the pipeline
window the signal at some point?

Michele:  If waveform does not go smoothly to zero,
then there is non-physiscal high frequency content
at ends

Siong:  So, the Hanning window is just to remove
this discontinuity

Michele:  Yes.

Jonah:  Maybe we should see the waveforms before and after processing

Michele:  Sure, we can do this.

* Apply zero-padding to be immune to edge effects
of filter

Block at "Plot the hrss at 10 K-parsec"
* "norm" function computes hrss of waveform
* "nor_phy" is the physical hrss at 10 kpc
* "nor" is the hrss that went into mdc

The code then zero pads the waveform to match
the mdc length standard

* Finally, the vector is written to a file

Code includes a check to make sure the hrss is not
adjusted by the application of the hanning window

Early mistake was made by directly computing hrss of 
waveforms without applying windowing and then to use 
THAT hrss to make conversions.  This was a problem 
because hrss we are sensitive to is part of burst that
is high frequency.  

Jonah:  Where is the low-freq. content that you 
are getting rid of?

Michele:  We are only looking at a piece of the
waveform in the TIME domain - the "burst" piece.
Our hanning window is centered around this burst

Michele:  It would be good for you guys
to run this script, and plot results block by block

Siong:  What is the mass of the proenitor?  Where is the cut?

Michele:  It is a cut in the time domain.  There is 
wobbling in the progenitor, and then it collapses.

Siong:  I think it is worth knowing what the 
nature of the initial wiggles is.

Michele:  The initial conditions are described
in the paper.  The masses are 1.8 solar masses.
They have a model that is, by itself, dynamically

Siong: Are the wiggles you throw away physically 

Michele:  What I throw out are 6 or 7 periodic 
cycles.  I suspect that is what they call "close
to stationary solution".  In the paper, I did not
find the timescale for this.

Siong:  If what we are getting rid of is not
physical, I think we don't need more discussion. 
If it IS physically motivated, it is not clear 
to me that we should just chop it away because 
it is not in our band.  

Michele:  I will ask them this question.
What they pass to me is a few milliseconds.
If they tell me the pre-cursor can be 
in place for a few seconds, what should we
do?  Ask them to re-run the numerical 

Siong:  At 10 kpc, the inspiral part could be
detected.  If there is a big sigal in our band, 
we want to know about it.

Michele:  Ok, I'll ask it.

Siong:  I think it is worth getting your hands on the 
earlier part of the waveform.

Michele:  I need to go at this time

11:55 am Michele leaves

Brennan:  The hyper-massive neutron 
star wave forms have some issues.

Siong:  So, these get made into 
MDC frames

Brennan:  Yes, we've done that.  

Brennan:  Let's look at my page:

Brennan walks us through this page.

Preliminary results show D1 and D4 waveforms with 50% detection 
around 100 kpc

Siong:  It would be great to see some text in the paper about
these results.

Siong:  Let's look at the table of review items.

Siong:  I think these are the items that should be there.  
Perhaps we should add issues related to using corr power 
in high freq to this list.  Please have a look at the table, 
and let me know if there is anything missing.  


Brennan Hughley, Jonnah Kanner, Matt Pitkin, Michele Zanolin, Ik Siong Heng