As it is it might be possible to mask a bad file as an .mp3, so we should be doing some sort of validation on the /play endpoint to verify what we're going to broadcast to the socket is valid.
How to actually validate that it is a valid MP3 or not is up in the air. And if we can get around it (headers? I'm not sure), we'd also want to avoid the time, and bandwidth, consuming part of it by not downloading the MP3 to the server running soundspaces.