mirror of
https://gitlab.science.ru.nl/mthesis-edeboone/m-thesis-introduction.git
synced 2024-12-21 19:13:32 +01:00
ZH: script to show signal_to_noise
This commit is contained in:
parent
265cb16ce2
commit
a2ed3d1fa8
1 changed files with 139 additions and 0 deletions
139
simulations/airshower_beacon_simulation/ac_show_signal_to_noise.py
Executable file
139
simulations/airshower_beacon_simulation/ac_show_signal_to_noise.py
Executable file
|
@ -0,0 +1,139 @@
|
|||
#!/usr/bin/env python3
|
||||
# vim: indent=fdm ts=4
|
||||
|
||||
"""
|
||||
Show Signal to noise for the original simulation signal,
|
||||
the beacon signal and the combined signal for each antenna
|
||||
"""
|
||||
|
||||
import numpy as np
|
||||
import h5py
|
||||
import matplotlib.pyplot as plt
|
||||
import numpy as np
|
||||
from collections import namedtuple
|
||||
|
||||
from earsim import REvent
|
||||
import aa_generate_beacon as beacon
|
||||
import lib
|
||||
|
||||
|
||||
passband = namedtuple("passband", ['low', 'high'], defaults=[0, np.inf])
|
||||
|
||||
def get_freq_spec(val,dt):
|
||||
"""From earsim/tools.py"""
|
||||
fval = np.fft.fft(val)[:len(val)//2]
|
||||
freq = np.fft.fftfreq(len(val),dt)[:len(val)//2]
|
||||
return fval, freq
|
||||
|
||||
|
||||
def bandpass_samples(samples, samplerate, band=passband()):
|
||||
"""
|
||||
Bandpass the samples with this passband.
|
||||
This is a hard filter.
|
||||
"""
|
||||
fft, freqs = get_freq_spec(samples, samplerate)
|
||||
|
||||
fft[ ~ self.freq_mask(freqs) ] = 0
|
||||
|
||||
return np.fft.irfft(fft)
|
||||
|
||||
def bandpass_mask(freqs, band=passband()):
|
||||
low_pass = abs(freqs) <= band[1]
|
||||
high_pass = abs(freqs) >= band[0]
|
||||
|
||||
return low_pass & high_pass
|
||||
|
||||
def bandpower(samples, samplerate=1, band=passband(), normalise_bandsize=True):
|
||||
fft, freqs = get_freq_spec(samples, samplerate)
|
||||
|
||||
bandmask = bandpass_mask(freqs, band=band)
|
||||
|
||||
if normalise_bandsize:
|
||||
bins = np.count_nonzero(bandmask, axis=-1)
|
||||
else:
|
||||
bins = 1
|
||||
|
||||
power = np.sum(np.abs(fft[bandmask])**2)
|
||||
|
||||
return power/bins
|
||||
|
||||
def signal_to_noise(samples, noise, samplerate=1, signal_band=passband(), noise_band=None):
|
||||
if noise_band is None:
|
||||
noise_band = signal_band
|
||||
|
||||
if noise is None:
|
||||
noise = samples
|
||||
|
||||
noise_power = bandpower(noise, samplerate, noise_band)
|
||||
|
||||
signal_power = bandpower(samples, samplerate, signal_band)
|
||||
|
||||
return (signal_power/noise_power)**0.5
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
from os import path
|
||||
import sys
|
||||
import matplotlib
|
||||
import os
|
||||
if os.name == 'posix' and "DISPLAY" not in os.environ:
|
||||
matplotlib.use('Agg')
|
||||
|
||||
f_beacon_band = (49e-3,55e-3) #GHz
|
||||
|
||||
fname = "ZH_airshower/mysim.sry"
|
||||
|
||||
fig_dir = "./figures/"
|
||||
show_plots = not False
|
||||
|
||||
####
|
||||
fname_dir = path.dirname(fname)
|
||||
antennas_fname = path.join(fname_dir, beacon.antennas_fname)
|
||||
time_diffs_fname = 'time_diffs.hdf5' if not True else antennas_fname
|
||||
|
||||
# create fig_dir
|
||||
if fig_dir:
|
||||
os.makedirs(fig_dir, exist_ok=True)
|
||||
|
||||
# Read in antennas from file
|
||||
_, tx, antennas = beacon.read_beacon_hdf5(antennas_fname)
|
||||
# Read original REvent
|
||||
ev = REvent(fname)
|
||||
|
||||
# general properties
|
||||
dt = antennas[0].t[1] - antennas[0].t[0] # ns
|
||||
pb = passband(30e-3, 80e-3) # GHz
|
||||
beacon_pb = passband(50e-3, 55e-3) # GHz
|
||||
|
||||
##
|
||||
## Beacon vs Noise SNR
|
||||
##
|
||||
if True:
|
||||
beacon_snrs = [ signal_to_noise(ant.beacon, ant.noise, samplerate=1/dt, signal_band=beacon_pb) for ant in antennas ]
|
||||
|
||||
fig, ax = plt.subplots()
|
||||
ax.set_title("Beacon SNR")
|
||||
ax.set_xlabel("Antenna")
|
||||
ax.set_ylabel("SNR")
|
||||
ax.plot([ int(ant.name) for ant in antennas], beacon_snrs, 'o', ls='none')
|
||||
|
||||
if fig_dir:
|
||||
fig.savefig(path.join(fig_dir, path.basename(__file__) + f".beacon_snr.pdf"))
|
||||
|
||||
##
|
||||
## Airshower signal vs Noise SNR
|
||||
##
|
||||
if True:
|
||||
shower_snrs = [ signal_to_noise(ant.E_AxB, ant.noise, samplerate=1/dt, signal_band=pb) for ant in antennas ]
|
||||
|
||||
fig, ax = plt.subplots()
|
||||
ax.set_title("Shower SNR")
|
||||
ax.set_xlabel("Antenna")
|
||||
ax.set_ylabel("SNR")
|
||||
ax.plot([ int(ant.name) for ant in antennas], shower_snrs, 'o', ls='none')
|
||||
|
||||
if fig_dir:
|
||||
fig.savefig(path.join(fig_dir, path.basename(__file__) + f".shower_snr.pdf"))
|
||||
|
||||
if show_plots:
|
||||
plt.show()
|
Loading…
Reference in a new issue