PALA – Python Audio Loudness Analysis

PALA is a modular Python package for measuring audio loudness according to the ITU-R BS.1770-5 standard.
It provides LUFS (I/S/M), True Peak, RMS, DR and LRA calculations with K-weighting and gating.

---

🍇 Purpose

The aim of PALA is to offer a unified, open-source loudness analysis tool for:

• mastering and post-production engineers
• researchers in psychoacoustics and signal processing
• developers building automated workflows

---

🏢 Installation

pip install pala

Or from source:

git clone https://github.com/HBB-ThinkTank/PALA.git
cd PALA
pip install .

---

🧠 Features (v0.2.0)

• ✅ LUFS calculation (Integrated, Short-Term, Momentary)
• ✅ True Peak (with oversampling), Sample Peak
• ✅ RMS (AES and ITU-R variants)
• ✅ DR (Dynamic Range), Crest Factor, Headroom
• ✅ Loudness Range (LRA using 10–95 percentile)
• ✅ Chunked processing for large audio files
• ✅ Modular structure with lufs, dynamics, utils, io

---

📊 Development Roadmap

Equal-Loudness Curves (Fletcher-Munson)

• 🛠️ Integration into frequency weighting
• 🛠️ Application to LUFS evaluation

Frequency Analysis

• 🛠️ FFT or Bark/octave band analysis
• 🛠️ Loudness per frequency band

LRA (Loudness Range)

• ✅ LRA based on LUFS segments (10–95 percentile)
• 🛠️ Interchannel gating / hysteresis

LUFS Calculation (ITU-R BS.1770)

• ✅ Momentary LUFS (400ms)
• ✅ Short-Term LUFS (3s)
• ✅ Integrated LUFS
• ✅ Gating: Absolute (-70 LUFS)
• ✅ Gating: Relative (-10 LU below average level)
• ✅ Multichannel weighting (e.g. 1.0, 1.41)
• 🛠️ Comparison with reference tools (e.g. pyloudnorm, ffmpeg)
• 🛠️ Validation using ITU-R BS.2217 material

Package Structure & Interfaces

• ✅ Modular package with analysis/lufs/dynamics/utils/io
• ✅ Chunking support for large files
• ✅ CLI support for single file input
• 🛠️ CLI batch mode
• 🛠️ Plot output (e.g. LUFS-M time curve)

True Peak / Peak / RMS / DR

• ✅ True Peak with oversampling
• ✅ Sample Peak (maximum sample value)
• ✅ AES RMS (unweighted)
• ✅ ITU-R RMS (K-weighted)
• ✅ Crest Factor / Headroom
• ✅ DR (True Peak - RMS)

Publishing & Distribution

• ✅ setup.cfg / pyproject.toml
• ✅ GitHub with README
• 🛠️ Upload to PyPI (production release)

---

🔮 Status

This project is in alpha stage. Results are close to professional tools (e.g. ffmpeg, pyloudnorm, APU plugins)
but not yet fully validated against ITU-R BS.2217 reference material.

Expect minor deviations (especially for edge cases or extreme signals).

---

🛠️ Requirements

• Python 3.8+
• numpy, scipy, soundfile, soxr, numba

---

📄 License

MIT License – free to use, modify and redistribute.

---

🙌 Contributions Welcome

You’re invited to test, extend or improve the code – feel free to open issues or pull requests.