Hacking for the hack of it. It is a great mantra, and one that we often take to heart. While there is significant value in hackathon projects that offer aid in difficult and demanding tasks, sometimes the most interesting hacks are those that exist for their own sake. Soundscape takes a novel approach to an activity that many of us love — discovering music. Instead of letting the user simply respond "yea" or "nay" to an ever increasing list of songs, Soundscape places you in midst of the action and shows you a world of music right under your feet. Users can then pursue avenues they find interesting, search for new or exciting pieces, or merely wander through a selection of dynamically curated music. With Soundscape, you have a hack-of-a-lot of power.
Soundscape is a Virtual Reality application based on the Google Daydream platform. It curates data by crawling Soundcloud and building a relationship model of songs in their repository. From there, it uses advanced graph search techniques to identify songs that are similar to each other, so that users can start with one set a long, and shift the genre and style until they find something new that they enjoy.
Soundscape is built on top of Google's yet unreleased platform for high quality mobile virtual reality—Daydream. Developing most of the application’s front end in Unity, we make use of this framework in conjunction with the existing Google Cardboard technology to help power a virtual experience that has high fidelity, low stutter, and intuitive input. The application itself is built in Unity, with custom hooks built into the Daydream infrastructure to allow for a high quality user interface.
The core functionality of Soundscape lies in our backend aggregation server which runs a node, mongodb, and express.js stack on top of Linode. This server fetches song, user, and playlist data through the SoundCloud API to generate similarity scores between songs — calculated through user comments and track favorites. This conglomerated data is then queried by the Unity application, alongside the standard SoundCloud data and audio stream. Search functionality within the app is also enabled through voice recognition powered by IBM’s Watson Developer Cloud service for speech to text. All of this works seamlessly together to power one versatile and unique music visualization and exploration app.
We are excited about Soundscape, and look forward to perfecting this for the final release of Google Daydream. Until then, we have exciting ideas about better search, and ways to incorporate other APIs