“Memory Mosaic” is a generative audio mashup tool for iOS devices developed by Mital as part of his doctoral thesis research on Audiovisual Scene Synthesis. The app listens to and learns sounds, either from your microphone or your music library. It saves unique sounds as “memories” and plays them back when it hears something similar.
Each of the “memories” represent an event of sound the app has heard and decided to store into its database. Once the app hears a new event of sound, it will be matched to its existing memories, triggering playback of the nearest memory. This new event of sound will also be learned, if it is distinctive enough.