Skip to main content

Time-frequency analysis through Stockwell transform

Project description


Python package for time-frequency analysis through Stockwell transform.

Based on original code from NIMH MEG Core Facility.

cf-badge PyPI-badge license-badge


Using Anaconda

If you use Anaconda, the latest release of Stockwell is available via conda-forge.

To install, simply run:

conda install -c conda-forge stockwell

Using pip and PyPI

The latest release of Stockwell is available on the Python Package Index.

You can install it easily through pip:

pip install stockwell

Installation from source

If no precompiled package is available for you architecture on PyPI, or if you want to work on the source code, you will need to compile this package from source.

To obtain the source code, download the latest release from the releases page, or clone the GitHub project.

C compiler

Part of Stockwell is written in C, so you will need a C compiler.

On Linux (Debian or Ubuntu), install the build-essential package:

sudo apt install build-essential

On macOS, install the XCode Command Line Tools:

xcode-select --install

On Windows, install the Microsoft C++ Build Tools.


To compile Stockwell, you will need to have FFTW installed.

If you use Anaconda (Linux, macOS, Windows):

conda install fftw

If you use Homebrew (macOS)

brew install fftw

If you use apt (Debian or Ubuntu)

sudo apt install libfftw3-dev

Install the Python package from source

Finally, install this Python package using pip:

pip install .

Or, alternatively, in "editable" mode:

pip install -e .


Example usage:

import numpy as np
from scipy.signal import chirp
import matplotlib.pyplot as plt
from stockwell import st

t = np.linspace(0, 10, 5001)
w = chirp(t, f0=12.5, f1=2.5, t1=10, method='linear')

fmin = 0  # Hz
fmax = 25  # Hz
df = 1./(t[-1]-t[0])  # sampling step in frequency domain (Hz)
fmin_samples = int(fmin/df)
fmax_samples = int(fmax/df)
stock =, fmin_samples, fmax_samples)
extent = (t[0], t[-1], fmin, fmax)

fig, ax = plt.subplots(2, 1, sharex=True)
ax[0].plot(t, w)
ax[1].imshow(np.abs(stock), origin='lower', extent=extent)
ax[1].set(xlabel='time (s)', ylabel='frequency (Hz)')

You should get the following output:


You can also compute the inverse Stockwell transform, ex:

inv_stock =, fmin_samples, fmax_samples)
fig, ax = plt.subplots(2, 1, sharex=True)
ax[0].plot(t, w, label='original signal')
ax[0].plot(t, inv_stock, label='inverse Stockwell')
ax[0].legend(loc='upper right')
ax[1].plot(t, w - inv_stock)
ax[1].set_xlim(0, 10)
ax[1].set(xlabel='time (s)', ylabel='amplitude difference')



Stockwell, R.G., Mansinha, L. & Lowe, R.P., 1996. Localization of the complex spectrum: the S transform, IEEE Trans. Signal Process., 44(4), 998–1001, doi:10.1109/78.492555

S transform on Wikipedia.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

stockwell-1.1.tar.gz (40.2 kB view hashes)

Uploaded source

Built Distributions

stockwell-1.1-cp310-cp310-win_amd64.whl (473.5 kB view hashes)

Uploaded cp310

stockwell-1.1-cp39-cp39-win_amd64.whl (474.1 kB view hashes)

Uploaded cp39

stockwell-1.1-cp38-cp38-win_amd64.whl (474.1 kB view hashes)

Uploaded cp38

stockwell-1.1-cp37-cp37m-win_amd64.whl (474.1 kB view hashes)

Uploaded cp37

stockwell-1.1-cp36-cp36m-win_amd64.whl (474.1 kB view hashes)

Uploaded cp36

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page