Why would we need a sound tracking device?
Sound tracking is pretty simple; Everyone used or at least knows applications such as Shazaam or Soundhound that are able to listen to a song, recognize it and give you its title.
The technology they are using is called “audio fingerprinting” and it works well to identify music and voices.
Another thing that they could recognize is an environment; Even if our brains are able to filter the ambient noises (Except for people suffering of LLI), we live in very noisy places: The noise of your fridge from the kitchen, the birds outside, the traffic in the street…
All these noises are potentially clues that can be used to guess the place you are and eventually the activity you are engaged in. For instance, eating, watching TV, working out, sleeping…
The purpose of that project is to build a smart listening box that listens to the surroundings, and with some prior training and continuous learning can identify changes of places and states in an environment.
How we want to realize it?
Because we care about inter-operability and transparency and decided to release the software and hardware parts as open-source. In order to facilitate development and community development, we chose to go for Python at least for the prototype (Python has huge libraries for audio processing compiled in C).
We favorize organization and tractability in our projects, but we also care about flexibility and adaptability so we adopted the Agile methodology.
Currently we are gathering the team and evaluate the feasibility of creating a sound tracking by implementing a POC (Proof of concept). Any help, documentation, research paper is welcome.
Possible features of our sound tracking device
- Activities, events and day phases recognition
- Voice tracking (presence detection, identification)
- Environment recognition (moving device)
Feel free to make any comment or join us if you too think activities tracking would largely improve automation.