python_speech_features

What it is and what it does

python_speech_features is a pure-Python library for extracting acoustic features from raw audio signals, primarily for automatic speech recognition (ASR) tasks. It computes Mel Frequency Cepstral Coefficients (MFCCs), filterbank energies, log filterbank energies, and spectral subband centroids—standard representations used in speech processing pipelines. The library accepts audio as a numeric array and returns feature matrices suitable for downstream machine learning or acoustic analysis.

The package provides configurable parameters for window length, step size, FFT size, frequency bands, and other signal-processing details, with sensible defaults for 16 kHz audio. However, the project has been abandoned since 2017 with no active maintenance, making it a legacy choice. If you need speech feature extraction and cannot find an actively maintained alternative, this library remains functional for basic MFCC and filterbank tasks, but you should verify compatibility with your Python version and consider whether a maintained fork or newer tool better suits your needs.

Use it for:

• Extract MFCCs from audio files for training automatic speech recognition models.
• Compute filterbank energies as input features for acoustic analysis or speaker verification systems.
• Preprocess raw audio signals into standardized feature representations for machine learning pipelines.
• Generate spectral subband centroids for audio classification or speech quality assessment tasks.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Extracts speech features from audio signals for automatic speech recognition, including MFCCs, filterbank energies, and spectral subband centroids.

No, unless you have a specific legacy requirement or cannot find an actively maintained alternative. The package is abandoned (last release 2017-08-16, 3285 days ago), has high install friction (source-only distribution), and lacks specified Python version support. For new projects, seek a maintained speech-processing library; for legacy systems already using this package, it remains functional but unsupported.

Install

python-speech-features on PyPI

pip install python-speech-features

uv add python-speech-features

poetry add python-speech-features

Installing python_speech_features

Before you install

High install friction due to source-only distribution (tar.gz). Package is abandoned—last release was 2017-08-16, over 3285 days ago, with no recent maintenance despite 2423 GitHub stars. Use only if no active alternative exists.

License in practice

MIT license is permissive; you may use, modify, and distribute this package freely with minimal restrictions, though you must include the license notice.

Quickstart

pip install python_speech_features

import python_speech_features

signal = [...]  # audio samples as array
mfcc_features = python_speech_features.mfcc(signal, samplerate=16000)

Audio signal must be provided as a numeric array; exact runtime dependencies are not specified in the package metadata.

Verify before relying

• Whether the package works with modern Python versions (requires_python is unspecified).
• What the implicit runtime dependencies are beyond what is declared.
• Active maintenance status or community forks that may have addressed abandonment.
• Whether the package requires external system libraries or compiled dependencies.

Package facts

Evidence: python_speech_features-0.6.tar.gz

Tags

More Information Analysis packages