The project aims at improving the audio track analysis features to compute information such as downbeat, time signature, measures, and phrases.

Downbeat detection will pave the way for bar detection and time signature computation. This feature is necessary since a song can be in a time signature other than 4/4, which will cause most beat and time-related features such as looping and quantization inconsistent with the track. Overall smarter quantization can be achieved with the added information.

Organization

Student

Harshit Maurya

Mentors

  • Uwe Klotz
  • Be Wilson
  • Jan Holthuis
close

2020