audeering/opensmile-python

Python package for openSMILE

What it solves

This project provides a Python interface for the openSMILE toolkit, allowing users to extract standardized audio features from audio files for use in machine learning from audio-visual signals.

How it works

The library allows users to specify a feature set (such as ComParE 2016, GeMAPS, or eGeMAPS) and a feature level (Low-level descriptors or Functionals). It then processes audio files to extract these specific acoustic features, which can be used as input for ML models.

Who it’s for

Researchers and developers working in audio analysis and machine learning from audio-visual signals who need standardized, high-dimensional audio feature extraction.

Highlights

  • Standardized Feature Sets: Supports ComParE 2016 (over 6,000 features), GeMAPS, and eGeMAPS.
  • Multiple Extraction Levels: Supports Low-level descriptors (LDD) and Functionals, as well as LLD deltas for ComParE 2016.
  • Simple Python API: Enables feature extraction from audio files with just a few lines of code.
  • Community Extensions: Additional feature sets have been added by the community.

Related

  • Project
    jcvasquezc/DisVoice

    A Python framework for extracting acoustic and linguistic features from speech to detect speech disorders and recognize emotions.

  • Project
    openml/openml-python

    A Python API for OpenML that allows users to download and upload datasets, machine learning experiment results, and curated benchmarking tasks.

  • Project
    Uberi/speech_recognition

    A Python library that provides a unified interface for speech recognition, supporting multiple online and offline engines like OpenAI Whisper, Google Speech, and Vosk.

  • Project
    librosa/librosa

    A Python library for audio and music signal processing that provides the foundational tools for building music information retrieval systems.

  • Project
    haifengl/smile

    A comprehensive, high-performance machine learning framework for the JVM that provides tools for traditional ML, deep learning, LLMs, and an agentic data science IDE.