mYOUsic RUN revolutionizes the concept of playlist: instead of having a predefined set of music in which the app shuffled depending on the tempo, the genre and so on, mYOUsic RUN creates the music on the go, while you run, in order to help you achieve the best experience possible, based on your physiology! 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.
Here is the link to the Devpost submission for the hackathon Lauzack 2018.
you will need to install android-studio, and then create a virtual device using AWD with the following specificties: emulate an Nexus 5X, with android 8.0, API 26
folder structure:
.
├── mobile app for the mobile phone
├── src
├── main
├── java contain the java files necessary
| to actually do the computations, modulate the music
└── res contain the files for the layout
└── build
├── wear in relation with the smartwatch
├── src
└── build
└── README.md
In mobile/src/java, in the mainActivity.java are the detalied implementation of our sound modulating algorithms.
We propose here an algorithm to "create" an appropriate music to each type of run and tailored for the usere, withouth user input apart from its biological data from the smart watch (see next section about this part).
We split a song into three tracks, using bypass filter: bass are between 0 and 60 Hz, medium are between 60 Hz and 4K Hz and finally high contains all the frequencies above.
We then combine them with modified amplitute between the tracks (bass,medium,high). While keeping the speed of playback constant between the tracks, we use it at our advantage.
we suppose the following effect of the music on the runner (motivated by our own experience):
- The stronger the
bass, the more motivated he will be to do an effort - The synchronisation between the tempo of the music and his "running tempo" helps him keep his cadence, and he will sligthly modify his own cadence if necessary to match the tempo of the music (if the desynchronisation is small enough)
- More melody might encourage some "heroic" effect and helps him go faster.
Hence, depending on the goal of the user (defined in step per minute), we modulate the different tracks mentionned above in order to keep him as closed as possible to his goal (motivating him by boosting the bass and the tempo, or calming him by moderating the tempo and the volume of the bass).
Due to a technical issue during this 24 hours Hackathon, we were not able to get access to a smart watch to collect the physiologic data we need to adapt the music. We simulated the behaviour of a dummy utilisator, with a simple response to the new objective: he will always follow the new pace and will never get tired. We simulated the heart rate as a function of step per minutes where we added a momentum to help mimick the process of recuperation of a true human.