Signal Processing Rockstars Imagine the Future

L-R: Angus Keatinge, Jeremy Bell, James Wagner Beats on the Barbie kick off the TA Reception at IEEE Sections Congress 2017. The team provided the evening’s entertainment, with several people dancing up through the last beats. Credit: Jessica Perry.

Music and signal processing – combining these two passions resulted in a hit for Beats on the Barbie, 2017 Signal Processing Cup champions. The team of undergraduate students used their knowledge of signal processing to create a robotic drum kit that plays with their band.

The IEEE Signal Processing Society’s 2017 Signal Processing Cup “Real-Time Beat Tracking Challenge, asked participants to implement a real-time beat tracker on an embedded platform and to demonstrate the performance with a creative output. The winning team, Beats on the Barbie, consists of University of New South Wales students, Jeremy Bell, Angus Keatinge, James Wagner, and Max Fisher. The IEEE Technical Community Spotlight spoke with Beats on the Barbie about their band’s technology, music, and performance at Sections Congress.

How did Beats on the Barbie form and how did you come up with the idea for the robotic drummer?

We were buds from uni, we all had an interest in signal processing and music so we decided to enter the Signal Processing Cup. We spent ages trying to think up what we were going to do for the creative part of the competition. We thought about having strobe lights, a flamethrower that pulsed with the beat, a mechanical boot that forced your foot to tap along with the music. But in the end, we decided that a drum kit that played along with you would be awesome and wouldn’t be terminator-esque.

How does the drummer work?

The drummer uses an electro-mechanical system of linear actuators and springs to provide a striking force. We call this configuration a ‘whacker’. The drummer has seven whackers across a full kit providing the ability to play any drum pattern or fill.

The overall system contains two embedded devices:

  1. A Raspberry Pi 3 which runs the real-time beat tracking algorithm we built, and
  2. An Arduino which provides the brains behind the drummer.

When a beat in the music is detected, (A) outputs a signal to (B). (B) then activates the relevant whackers based on the programmed drum pattern. Timing needs to be extremely precise for the drummer to sound good. We spent a long time refining both the beat tracking algorithm and drummer interface which accounts for timing offsets in the physical drum strikes.

We have two options to tell the drummer what to play:

  1. You can control the drum sequence real-time using an interface board we built, or
  2. You can load in a midi file and have the drums play along. This can be done with both live instruments and music files.

Do you mainly focus on covers or do you have your own songs?

So far we have mainly focused on covers. Playing our own songs with the drum kit would definitely add another dimension to the usefulness of the system though it’s certainly something that we hope will be tried in the future.

 

2017 Sections Congress attendees experiment with the robotic drummer. Credit: Jessica Perry.

What was it like to attend and play at Section Congress?

Sections Congress was fantastic. But boy, being rockstars is a lot of work. Performing there was some of the most fun we’ve had as a group, the crowd was really into it during the performance and lots of people came by and had a go on the drum sequencer. Attending the rest of Sections Congress was very interesting, even though the workshops mostly weren’t targeted at us, we still met a number of cool people and had some really good chats.

What is the future of Beats on the Barbie?

Hopefully a bright one. We all have jobs now but we still manage to catch up pretty regularly.

Some more work still needs to be done to make the drum kit more user friendly, but other than that, the drum kit is pretty much complete. Before too long we’ll be building some other crazy device that has a very niche application and realistically isn’t very practical.

How has this experience influenced your plans for the future?

We learned a lot about digital signal processing (DSP), embedded systems, and product design while working on the Signal Processing Cup. All of the members of Beats on the Barbie are starting careers in engineering design and signal processing, so these skills will serve us beyond university as well.

L-R: James Wagner, Jeremy Bell, Angus Keatinge
Beats on the Barbie pose at the Signal Processing Society booth at IEEE Sections Congress 2017. Credit: Jessica Perry.

The project was funded by UNSW’s school of Electrical Engineering and Telecommunications, who will be looking after the drum kit into the future. We still demonstrate the system whenever the school is having an open day or high school engagement program, with the aim of spreading more awareness and interest in digital signal processing to future engineering students.

 

 

 

View performances by Beats on the Barbie:

  1. Beats on the Barbie Compete at the 2017 Signal Processing Cup
  2. Watch Beats on the Barbie perform Weezer’s “Island In the Sun” at IEEE Sections Congress 2017

Comments

Leave a Reply