Presented as a challenge by Logitech, during Lauzhack 2018, the title of the challenge was : "Music powers our passion". We were invited to develop a music app for running, with the opportunity to deal with smart watches , in order to tailor the experience by collecting physiologic data from the user.
What it does
mYOUsic RUN is a game changer in the world of sport music. Instead of having a predefined playlist, which is shuffled and played depending on the tempo or the genre, mYOUsic RUN creates the music on the go. While you run, in order to help you achieve the best experience possible, it uses physiological data to generate the perfect melody! Depending on your objective, it may detect if you need a boost or on the contrary of you need to slow down in order to keep your pace.
In the default setting you can just open the app and go for a jogging:
We monitor your heartbeat and modulate the power of the bass in order to keep you around the default objective in term of cadence, which you can modify from the interface during the run. The music will adapt to help you achieve your goal!
We use your number of steps per minute, in order to match the tempo of the music with your rhythm. This helps us modulate your speed through the music and to make the running experience as pleasing as possible when the tempo of the music you are listening to is exactly in synch with yours.
You only need your smartwatch, your phone and earplugs !
How we built it
We use android studio and created an app from scratch which collects data from a smartwatch (due to too few smartwatches form Logitech, we simulated the physiologic input data in real time for this implementation), and then adapt the music we create depending on these parameters, and the type of run you selected.
The music given by the app is based on electro music, from which we separated three tracks: the bass, the medium and the high frequencies. We then use this decomposition to be able to modulate the different part of the song independently in frequency and amplitude while keeping the playback rate speed constant among them.
We modulate the bass, the tempo and the high frequency in order to get the user to its goal: the bass is used as a motivation when your heartbeat is high, meaning you are exercising hard. The tremble and the tempo are used to motivate you to go faster if needed, our to push the user to slow down.
Challenges we ran into
During the Hackathon, we could not have access to any smart watch during the whole 24 hours, and so we decided to simulate the data, based on our experience in running and some of our personal jogging data. This is why the current implementation may seem simplistic, but with real users data and the proper training (regress the parameters of the modulation of the different tracks (dB,rate of playback,...) versus the influence in the rhythm and the overall performance of the user), this could improve.
For the moment, the implementation has only a restricted sound library. This is due to the time constraint of the hackathon and the non-experience we had in the matter of creating music. But this could be easily overcome with the help of trained people.
This is also related to the third issue, which is the quality of the modulation of the music.
What's next for mYOUsic RUN
We have the following ideas for the future:
- Get a smart watch and work with its specific API in order to have a live feed of physiologic data from the user
- Work in collaboration with artists in order to create more modular songs of better qualities and of different types
- Propose multiple training profiles: interval trainings, long run, progressive run, fartlek, hills repeat...
- Also we did not use the specific sound equalizer of the Jaybird from Logitech, a next step would be to use both our modulation program and the built in equalizer of the headphone to achieve optimal sound quality
- Improve the user interface with a a more user friendly design, more statistics displayed...