I’m currently preparing a presentation to discuss progress made so far in my 4th Year Engineering project. The project is to try and build a music transcription system. In essence, you pump a sound file into a computer, turn the handle, and the score comes out the other end. These sorts of systems have many potential uses.
Great jazz (if there is such a thing) often includes lots of improvisation and original work. How is the ordinary man in the street supposed to be able to produce similar sounds without spending years learning the art of improvisation? How much simpler it would be if a recording could be transcribed quickly and easily and the results published for anyone to play?
Sound files are massive, using lots of disk space. It would be far simpler to store a transcribed score of a recording. This could be played back using a digital synthesiser. Inevitably there would be a quality loss, but this could be made up by storing additional data such as vibrato and dynamics.
Finally imagine the potential for searching on the internet. Imagine if you could whistle a melody into a microphone and Google could find a recording of the piece it came from. This will only become possible when it is simple to transcribe recordings.
In order to illustrate this final point I produced the following picture:
The fact remains that Internet Explorer 7 running in Vista looks very, very cool.