/dance_detect

detect / predict dance event from audio in real time

Primary LanguagePythonMIT LicenseMIT

Dance Detect

This project is designed to be used with Baxter Boogie. It uses the Essentia library to extract information from music for a robot to respond to. It also plays the song. Requires python 2.7

Information

  • Beats: It extracts beat information from the current fram of music and makes a prediction of when a beat will occur in the future.
  • Energy: It extracts the amount of energy in the current frame buffer.
  • Onset: It extracts the number of new music onsets on the frame buffer.

Usage

  1. Place a wav file in a subdirectory called "samples". (So dance_detect/samples/my_music.wav)
  2. Set the IP of robot running Baxter Boogie and a ports to use. These live in the network.ini file. Ip is address of computer on the robot. The ports are for each part of baxter boogie. So one for each arm plus the base.
  3. Run with python dance_detect.py /samples/some_music.wav
  4. Close with ctrl-c

Installation

This project depends on:

  • Essentia for music information algorithms. Install instructions here
  • PyAudio to play the audio.
  • Clone this repo then places music in samples sub folder. Make sure they are 8 or 16 bit wav files.