Synesthesia
Synesthesia will be a learning music composition/remastering tool, improving its own abilities through user/community interaction. Users can listen to randomly generated samples and rate them based on a variety of metrics, which will be used to finely tune the composition algorithms, potentially leading to the ability to create masterpieces. The primary piece of this project would be a web based interface through which an open community can generate and rate dynamically generated sounds/samples/pieces. Based on the feedback from the user, Synesthesia would use genetic algorithms to create mutations in the sounds, trying to discover what sounds good to a user and what doesn’t, something identical to an optometrist asking ‘Number 1? Or number 2?’. Highly rated mutations would be identified and improved upon, while low rated mutations would be avoided. From the knowledge base that Synesthesia builds, the next piece would be generating larger, more elaborate compositions, possibly to act as a turing test versus classic compositions.