SNIFA

SNIFA is a tiny Python library for computing an instantaneous frequency approximation numerically using a very simple algorithm. This algorithm was originally created for the purpose of analysing the gyrotropic dynamics of vortex based nano-oscillators but it can be used for any kind of varying monocomponent-frequency signal (only the main frequency is captured by the algorithm).

Installation

Use the package manager pip to install snifa.

pip install snifa

This package is also available in this private GitLab instance and can be installed using the --extra-index-url parameter.

pip install snifa --extra-index-url https://gitlab.flavio.be/api/v4/projects/9/packages/pypi/simple

Usage

import snifa
import numpy as np

ti, tf, nb_pts = 0, 1, 400
fs = nb_pts/(tf-ti) # sampling frequency
t = np.linspace(ti, tf, nb_pts)

f = 20+((100*t)**1)/2 # my shirp :-)
x = np.sin(2*np.pi*f*t)
y = np.cos(2*np.pi*f*t)

x *= (1.0 + 0.5 * np.sin(2*np.pi*3.0*t))
y *= (1.0 + 0.5 * np.sin(2*np.pi*3.0*t))

# Uncomment if only x-component is available
# and comment the other line with filt_freq
# t_, f_ = snifa.filt_freq(t, x, w_filt=0)
t_, f_ = snifa.filt_freq(t, x, y, w_filt=0)

# w_filt is a multiplyer of the number of cycles over which to
# perform a moving avereage. The default value is w_filt=1
# (w_filt=0 means no filtering)

import pylab as pp
pp.plot(t_, f_)
pp.xlabel('Time (s)')
pp.ylabel('Frequency (Hz)')
pp.show()

Contributing

Pull requests are welcome.

License

MIT