ROBOD: a Real-time Online Beat and Offbeat Drummer ock 1 , Florian - - PowerPoint PPT Presentation

▶

Nov 24, 2022 513 likes •670 views

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion ROBOD: a Real-time Online Beat and Offbeat Drummer ock 1 , Florian Krebs 1 , 2 , Amaury Durand 3 , Sebastian P oll 1 , Raminta Sebastian B Balsyte 1 1Department of

SLIDE 1

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD: a Real-time Online Beat and Offbeat Drummer

Sebastian B¨

ck 1, Florian Krebs 1,2, Amaury Durand 3, Sebastian P¨
ll 1, Raminta

Balsyte 1

1Department of Computational Perception, Johannes Kepler University Linz, Austria 2Joanneum Research, Graz, Austria 3T´ el´ ecom ParisTech, Paris, France 1 / 12

SLIDE 2

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

Beat tracking

2 / 12

SLIDE 3

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

Beat tracking : online problem

Limited information : only previous data is available Cannot fix errors afterwards

3 / 12

SLIDE 4

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD

interaction

4 / 12

SLIDE 5

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD

interaction

4 / 12

SLIDE 6

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD

interaction Rhythmic companion beats downbeats rhythmic patterns

4 / 12

SLIDE 7

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD : Global description

100fps every beat

Audio

Signal pre-processing Beat tracking Downbeat & pattern tracking Pattern Position inside bar Beats

STFT Semitone filterbank Positive differences RNN DBN Beat / Features synchronisation GMMs DBN ROBOD

5 / 12

SLIDE 8

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion RNN DBN

Beat tracking : RNN

6 / 12

SLIDE 9

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion RNN DBN

Beat tracking : DBN

Before beat

7 / 12

SLIDE 10

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion RNN DBN

Beat tracking : DBN

On beat

8 / 12

SLIDE 11

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

ROBOD : Global description

100fps every beat

Audio

Signal pre-processing Beat tracking Downbeat & pattern tracking Pattern Position inside bar Beats

STFT Semitone filterbank Positive differences RNN DBN Beat / Features synchronisation GMMs DBN ROBOD

9 / 12

SLIDE 12

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

Pattern tracking

10 / 12

SLIDE 13

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

Conclusion

ROBOD is more than just a beat tracker The online beat tracking algorithm is open source and available at : https://github.com/CPJKU/madmom or pip install madmom Special thanks to : Martin Preinfalk, Rainer Kelz, Harald Frostel, Filip Korzeniowski.

11 / 12

SLIDE 14

Beat tracking ROBOD Beat tracking Pattern tracking Conclusion

Question

Questions ?

12 / 12