Sounds cool, Gary. I agree with Brooke: the MusicIR community (J. Stephen
Downie & many others) would be a good starting point.
The work I'm aware of goes the other way -- starting from scores. For
instance, here's the Bodelain's crowdsourced digital annotation project:
Andrew Bullen also did a talk in 2008 and Code4Lib journal article about
Optical Music Recognition:
You can get to a video from:
On the crowdsourcing front, you might look at the nichesourcing approach to
get stuff to the right people:
Let us know how it goes!
On Thu, Jul 10, 2014 at 8:07 PM, Gary McGath <[log in to unmask]> wrote:
> I'm doing some volunteer work on a project for a basement sound engineer
> and publisher who has a large collection of recordings from conventions,
> and wants to crowdsource identification of them with the people who
> attend those conventions.
> The idea is that people would be able to log in to a website, see a list
> of short clips they can play, and enter identifying information such as
> song title, performers, and instruments. It occurs to me that this might
> have broader usefulness among libraries, archives, and researchers. If
> so, it may be worth expanding this into a bigger open-source project and
> seeking crowdfunding for developing it further, getting nice graphics,
> etc. (Crowdfunding for crowdsourcing just seems right.)
> On the other hand, the most common result when I come up with a great
> idea is discovering that someone else has already done it better than I
> Obviously there are copyright issues. That's why only logged-in users
> with a legitimate need can play the clips for identification purposes.
> There may be other protections as well, such as limited-time availability.
> So I'm interested in thoughts on this. Does it duplicate something that
> already exists? Would it be generally useful?
> Gary McGath, Professional Software Developer