Sound Collage Proposal

*note: completely different from original list

On SoundCloud, users are able to comment with timestamps, placing their comment at a particular time in the track. This allows them to comment about a particular section of the music. Often times, the comments crowd around the climaxes of the music. Given that this is the internet, the comments are occasionally ridiculous.

My sound collage will be dynamically generated from the comments of a given track (either user-selected or a random popular one) on SoundCloud. Ideally, the program would convert the text of the comments into speech with the appropriate tone of voice and place them at the time the comment was timestamped, with some random panning/spatialization. This will use the SoundCloud API, an HTML5 text-to-speech synth, and WebAudio.

What will end up being played back is, not the music itself (or maybe it’ll play quietly), but the reactions to the music. As a result, the structure of the piece will loosely follow the structure of the given track. If the comments are sparse, the sound collage will appear to contain random commentary over the track. Listeners will be able to easily use semantic listening to understand the commentary. If the comments are dense, the sound collage will sound like the crowd at a concert, reacting to the intensity of the music. Listeners will end up using causal listening to assign the source to a crowd, and reduced listening in following the intensity of the music. By listening to the reactions, we will be able to listen to music from a different perspective.

I expect people to listen to this like they would listen to ordinary music, since the sound collage is generated in the browser using SoundCloud. If the synthesized voices are convincing enough, densely commented tracks may give the listener the impression of being at a concert. However, it’s more likely that the synthesized voice will be ironically expressionless and the listener will be busy LOLing 😐


One thought on “Sound Collage Proposal

  1. Hi David,

    I’m looking forward to trying this out. I imagine you’ll need to experiment with a number of test files to see how it works out in different scenarios, which may help you tune things like the synthesized voice(s), speed, and etc. I also wonder whether you might need to provide some guidance as to the best files to choose — if the project is equally successful with any file, perhaps not. But if files that are heavily commented (for instance) are far more successful than files with light or no commenting, it might be frustrating to have to poke around for the best files.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s