next up previous contents
Next: Function: qn_inject() Up: GRASP Routines: Black hole Previous: Function: qn_normalize()   Contents

Function: find_ring()

void find_ring(float *h, float *u, float *r, float *o,
               int n, int len, int safe, int *off,
               float *snr, float *mean, float *var)
This optimally filters the strain data using an input template and then finds the time at which the SNR peaks. The arguments are:
h: Input. The FFT of the strain data $\tilde{h}(f)$.
u: Input. The normalized template $\tilde{u}(f)$.
r: Input. Twice the inverse power spectrum $2/S_h(\vert f\vert)$.
o: Output. Upon return, contains the filter output.
n: Input. Defines the lengths of the arrays h[0..n-1], u[0..n-1], o[0..n-1], and r[0..n/2].
len: Input. The number of time domain bins for which the filter $u(t)$ is non-zero. Needed in order to eliminate the wrap-around ambiguity described in subsection [*].
safe: Input. The additional number of time domain bins to use as a safety margin. This number of points are ignored at the beginning of the filter output and, along with the number of points len, at the ending of the filter output. Needed in order to eliminate the wrap-around ambiguity described in subsection [*].
off: Output. The offset, in the range safe to n-len-safe-1, for which the filter output is a maximum.
snr: Output. The maximum SNR in the domain specified above.
mean: Output. The mean value of the filter output over the domain specified above.
var: Output. The variance of the filter output over the domain specified above. Would be unity if the input to the filter were Gaussian noise with a spectrum defined by $S_h$.

Author: Jolien Creighton, jolien@tapir.caltech.edu


next up previous contents
Next: Function: qn_inject() Up: GRASP Routines: Black hole Previous: Function: qn_normalize()   Contents
Bruce Allen 2000-11-19