tsm-realtime 0.1.1

Real-time audio time-scale modification with look-up approximation and full computation methods

Realtime Time-Stretching Service

This project provides real-time audio time-stretching capabilities using two classic DSP techniques:

• Phase Vocoder (PV)
• Overlap-Add (OLA)

It also supports hybrid approaches that includes a look-up method in the aim to reduce runtime for the Phase Vocoder system.

Authors: Sayema Lubis, Clark Peng, Jared Carreno

---

Features

• Real-time processing and streaming with PyAudio
• Multiple time-stretching algorithms:
  • Baseline OLA
  • Baseline and Look-up methods for Phase Vocoder
  • Combines into two seperate Hybrid methods
• Harmonic/Percussive source separation using librosa
• Interactive keyboard controls to adjust playback rate (alpha) in real-time
• Modular implementation with reusable TSM functions

---

Requirements

Install dependencies with:

pip install numpy scipy librosa pyaudio keyboard

Usage

Run the service directly: