A Python wrapper for calculating PESQ (Perceptual Evaluation of Speech Quality) scores from audio files.
PyPESQ is a Python package that calculates PESQ (Perceptual Evaluation of Speech Quality) scores, which measure the perceptual quality of speech signals. It wraps the official PESQ C routine to provide an easy-to-use Python interface for comparing reference and degraded audio files. This helps developers and researchers objectively evaluate speech quality in applications like codec testing, network transmission analysis, and audio enhancement.
Audio engineers, researchers, and developers working on speech processing, telecommunications, or audio quality assessment who need standardized metrics for evaluating speech degradation.
It provides direct access to the industry-standard PESQ algorithm from Python, eliminating the need for manual C integration. The package is simple, focused on evaluation accuracy, and integrates seamlessly with common Python audio libraries like `soundfile`.
A python package for calculating the PESQ.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Directly wraps the ITU-T P.862 C routine, ensuring accurate NB PESQ MOS calculation as per the standard.
Minimal code required to compute scores with NumPy arrays, as shown in the README's example using soundfile.
Accepts audio data as NumPy arrays, integrating seamlessly with Python's scientific computing ecosystem.
Designed solely for assessment purposes, avoiding feature bloat and maintaining adherence to PESQ specifications.
Only returns NB PESQ MOS, missing wideband and other score types, which restricts use for modern audio codecs.
Marked as WIP with a TODO list for more tests and uncertain features like wav normalization, indicating potential bugs.
The NOTICE section highlights PESQ ownership by BT and KPN, raising licensing concerns for commercial or distribution use.
Relies on external libraries like soundfile for audio loading, with no built-in preprocessing, limiting out-of-the-box usability.