Our team was busy hacking at HackRPI this weekend. They had a hilarious opening ceremony with lots of programmer jokes.
A database admin walks into a bar, he goes up to two tables and asks, "Can I join you?"Anyway, we went to this hackathon with a plan (how unusual?). Our idea was to create a speech synthesis engine that could listen to someone speak, and then generate speech in that person's voice.
Our plan was to record a user speaking, break the recording up into individual words by indexing it against Google's speech recognition API. After that each word could be further divided into individual phonemes that could later be reassembled into new word sounds.
Unfortunately, we ran into a few pit falls along the way while attempting to normalize the recorded sound parts so that the resulting speech sounded realistic and not choppy. Our final presentation for the judges ended up being less than spectacular since our table ended up being placed right next to a team that had used a large set of Bose speakers for their noise-making project (this made speech recognition a bit difficult).
We didn't win anything, but this was still a great hackathon to participate in. I really enjoined working on this project and learned a lot about processing audio in Python.