jwheare’s web of music and the Media URI spec

Towards a web of music:

Playdar isn’t really a browser. It’s more of a search engine. But can you imagine using a web browser today that didn’t have Google built in? The idea of a browser goes hand in hand with that of a search engine; in my efforts to relate the two I may have blurred the waters.

I’d like to talk a bit more about what the web looks like today, and how we can make it friendlier to the idea of a music browser.

So far, this has led to a web of control, with content centralised in the hands of publishers. The distribution of creative works is stifled, not to ensure the protection of rights, but due to a careless muddling of formats. And this sloppy integration of multimedia into the browser ensures we’ll be mired in codec hell for years to come.

Can we do better? Can we create an ecosystem for browsing, subscribing, sharing, discussing, listing and rating the stuff of the web that’s separate from HTML? Music is an obvious area of opportunity, and we’ve already got music browsers that have escaped the gravitational pull of the browser. The problem is, they’re still locked into the publisher’s web of control. The iTunes Store and Spotify represent bold new ways to access a wealth of music, but they’re essentially blind to a world of sound outside their borders.

(via James Wheare’s Blog | jouire.com)

Audio itself isn’t hypertext. I wonder what it would be like to have music audio be hypertext. Would you click on a Led Zeppelin lyric to go to the blues tune that it came from? Maybe you could navigate from a chord progression in one song to the same chord progression in other songs.

When you share a set of songs by putting them all in the same MP3, you can’t address them individually. When you use a playlist, the playlist is a hypertext container for each song.

And then there’s Media Fragments URI 1.0, which is a specification for pointing inside of multimedia files. That’s full-fledge music hypertext, even though it is used for audio bytes rather than musical songs.

Which makes me speculate about metafiles that attach semantics to audio files. The metafile would link to a range within an audio file in such a way that high level concepts were communicated. Like:

<div class="song-meta-map">
<a href="example.mp3#t=0s,15s" class="intro">intro</a>
<a href="example.mp3#t=16s,45s" class="verse">first verse</a>
<a href="example.mp3#t=46s,75s" class="verse">second verse</a>
<a href="example.mp3#t=76s,106s" class="chorus">chorus</a>
<a href="example.mp3#t=107s,137s" class="solo guitar">guitar solo</a>
<a href="example.mp3#t=138s,178s" class="chorus">chorus</a>
<a href="example.mp3#t=179s,210s" class="outro">outro</a>

My Web of Songs deck is different that jwheare’s concept in that it doesn’t conceive of audio-specific forms of hypertext, but thinks about general systemic issues preventing music from being a first-class citizen of the web.