What it does
Based on a hearing test, Spotify listening preferences, and custom device profiles, Mage mixes you an ideal version of the song you are playing to match the conditions.
How I built it
Both Anthony and I have a background in audio, but as our resident audio engineer, Anthony took the reigns on most of the planning and system design for our mixing and recommendation systems. After the planning process, we split off onto individual tasks, me building out the database and Anthony coming up with a sample dataset based on his hearing test and listening history.
Challenges I ran into
We spent arguably too much time planning out the nature of the interactions between each level of our mixing chain. Because the system we designed relies on a ton of inputs, we wanted to make sure the changes that we were making to the audio were both noticeable and beneficial to the users so we very deliberately and intricately planned out the structure. However, that came at the cost of time and we weren't able to build in many of the features that we wanted to because of it.
Accomplishments that I'm proud of
Getting smooth multitrack audio working in the browser, and developing a