Social Music Analysis Widget
The Transforming Musicology project includes three musicological work packages intended as case studies to demonstrate how music information retrieval and related technologies may be applied in musicological research. These are complemented by four technical work packages in which an infrastructure for supporting computational musicology will be developed drawing on a lot of existing technology.
One of the three musicological work packages is called Musicology of the Social Media. In it we propose to explore the activities—especially discussion—carried out in online music communities. Although much of this is unlikely to be considered musicology by those involved, we are interested in interpreting it as such and understanding the kinds of discourse that are going on.
That's the background. One of my main responsibilities on the project is this web site. Keeping in mind our interest in engaging the musicological community, I designed a template with a login box thinking that we may allow people to comment on pages, for example. My train of thinking then progressed: this should actually allow logging in using profiles from other sites such as Twitter or LinkedIn; and then continued: this could in fact be part of our social media work package; and resulted in this idea: if we could develop some Web widget that allows people to make comments on pieces of music (e.g. selecting a part of a piece which is your favourite, comment on why, share it with others, or selecting a part of a piece which makes it similar to another piece), we can build up a database of how people understand and talk about music and eventually mine it for musicological insights.
This social music analysis widget may become one of our first technical outputs. There will be a number of significant challenges in developing it:
- it needs an easy to use user interface that allows selection of parts of pieces of music, e.g. selecting a range in a waveform or from a score;
- it would be great if we could allow searching within a piece to find a range to comment on;
- it will require a large database of music which is free from copyright restrictions, that has reliable metadata, is encoded in ways that will work with the user interface, and that is comprehensive enough to satisfy the discussion requirements of diverse musical communities;
- we will need to think carefully about how to attract a user base;
Of these, the data acquisition is probably the hardest problem to overcome. For now, we will need to work with a very restricted collection in order to get the software infrastructure in place.