I spend a lot of time preparing data to make my animated graphical scores:
People often ask "can't all that stuff be done automatically?"
The short answer is "no."
A longer answer is "yes, but not well enough to save any time."
It would be great if software could automatically extract a score from an audio recording,
They've been working on this for years; let me show you where they are now ...
Here's a recording I made of a simple two-part piece:
I took put that recording through pitch recognition software; here's the result:
This is a pretty simple test: only two notes, single instrument, etc.
Sometimes, if there's no score for a piece and it's relatively simple,
That's what I did for this video:
But that's a lot of work: nearly every note needed some sort of hand-adjustment.