Or rather, problems. The main problem is that I don't have the wherewithal to make good recordings. The First Demonstration Reel is okay as a demonstration, but the performances on them have been criticized -- and justly -- of being amateurish. What can I say? -- I'm an amateur. It would be so much nicer to have good performances done by real musicians.
There's also the issue of the poor quality of the image on videotapes. Both the resolution and the fidelity of colors is poor on videotape (especially NTSC; the nickname Never The Same Color is well-deserved). There are other problems with videotape: fast forward and rewind are slow and imprecise, it's hard to play the same section over and over -- not to mention that you're stuck with a single recording.
Working harder on the performances. One thing that editing performances on the M.A.M. has taught me: there's always room for improvement! No matter how much I've worked on a rendition of a piece, I've never felt that I had done all I could, or even all I could think of. On the contrary, the more I worked on a piece, the more I became aware of what a lousy job I was doing. So yes, I could do a better job, but I don't think I'd ever end up with something I was happy with (though I admit that it would be a good exercise for me).
Selling the M.A.M. as software only goes part way toward addressing these problems; the image quality and ease of use would be much better on a computer, but the performances would not.
Licensing existing recordings. This is very difficult. I've contacted various companies to ask about this, and have been uniformly rejected (or rather, ununiformly -- some don't even answer my letters).
More and more computers these days have CD players that are capable of playing audio CDs -- and being controlled by the computer. So, the idea was this: for a given CD, there'd be a piece of software that was specially keyed to it. The software would contain the data for the scrolling score (or other animation) corresponding to the music on the CD, and would not work unless the CD were present in the computer's disk drive. The beauty of this is that I wouldn't have to license the recordings at all -- since I wouldn't be selling or distributing them! A customer would have to buy the CD themselves (or perhaps I could arrange some kind of bundling (though the one recording company I contacted about this was pretty hostile to the idea).
There'd be a fair amount of technical work required to make this all possible. The worst part is that pitch extraction from audio data is still not something that's easy, and I don't think I have the technical chops to succeed where others have failed. So I'd probably do it a more labor-intensive way: input the scores through MIDI, then sync them to the audio recordings, with the help of software to make that easier. There'd also be the task of making the video player for these things (not to mention marketing...).
The result of this would be everything the First Demonstration Reel was not: a great performance, perfect colors and clarity, ease of control, and the flexibility to do many different views of the same piece.