Tempo Step

[November 2020]

Background

Recently, I’ve been trying to get back into the habit of running cross-country, and found pacing to be somewhat of a challenge. So, I wanted to work on a program that analyzes music tempo and plays songs to listen to while running, based on someone’s step pace. Then, the person would simply need to match their footsteps with the music beat.

Searching up how to start

I started learning audio processing in Python. It turns out there are many possible packages one could use, which all have some pretty handy capabilities. For example:

• Scikit-learn MIR (Music Information Retrieval) algorithms
  • Detects key attributes of a song, ex. pitch, transcription, mood, genre, beat, key
  • The key pipeline the audio goes through is: Segmentation –> Feature Extraction –> Machine Learning.
  • This talk