numpy-rms: a fast function for calculating a series of Root Mean Square (RMS) values

• Written in C and takes advantage of AVX (on x86-64) or NEON (on ARM) for speed
• The fast implementation is tailored for C-contiguous 1-dimensional and 2-dimensional float32 arrays

Installation

$ pip install numpy-rms

Usage

import numpy_rms
import numpy as np

arr = np.arange(40, dtype=np.float32)
rms_series = numpy_rms.rms(arr, window_size=10)
print(rms_series.shape)  # (4,)

Changelog

[0.5.0] - 2025-03-14

Changed

• Bump numpy to >=2,<3. If you need compatibility with numpy 1.*, you can use numpy-rms==0.4.2

Removed

• Remove support for Python 3.8
• Remove support for PyPy on Windows

For the complete changelog, go to CHANGELOG.md

Development

• Install dev/build/test dependencies as denoted in pyproject.toml
• CC=clang pip install -e .
• pytest

Acknowledgements

This library is maintained/backed by Nomono, a Norwegian audio AI startup.