Warning: this is much more conceptual than most OF projects, and requires a bit of introduction
it all looks very fascinating!
so what comes first, the glitchy images which are then translated into mp3?
the other way around or are they separate all together?
i don’t think the stream is abrasive at all.
there seems to be a pattern and continuity in the sound which is somewhat rhythmic.
also, since the stream “contains every sound we can distinguish as humans”, will the random noise eventually (between now and infinity) be able to produce a coherent sounds such as someone speaking for example?
The glitchy images came first, they helped me find an enumeration that had more variation than general binary or gray code enumeration. They weren’t quite “translated” to MP3 so much as “interlaced” or “dropped in” to an MP3 skeleton.
The rhythm you hear comes from the regularity of the transformed binary.
It will eventually produce every moment (26 ms snippet) of someone speaking, but I don’t think those moments will appear next to each other. The original idea I had for this composition was “every song” – in which case coherent speech would eventually be produced, but it would be so much longer you would never really get to that part
really nice - so its a bit like the Library of Babel but as an audio work? ( http://en.wikipedia.org/wiki/The-Library-of-Babel )
I enjoyed listening to it - the rhythmic pulse from the bits definitely lends something to it.
For it to be able to produce any possible sound -like “good morning” with a heavy french accent- the pulse would have to disappear ? Is that possible ? Are you archiving really interesting sounds / moments you find?
Is it pure generated or is it working with a seed noise? It seems generated but I think when you were talking with me a while back you mentioned analysis and re-encoding on the fly.
anyway very nice work!
I’m super glad you shared the code as I can’t wait to look at the mp3 encoding stuff you wrote!
Yeah, definitely Library of Babel-y. Also see “The-Aleph”, which is a bit more poetic approach to the idea of “everything” instead of via enumeration.
The pulsing would have to disappear for something like “good morning” to appear, yes… but more importantly, after 10^450 years it would have to start working on every combination of two 26 ms sounds… then after that’s over, three 26 ms sounds… I decided not to go this route because it gets increasingly harder to enumerate longer sounds in an interesting way.
There is no archiving going on, as I have no metric for which sounds are more interesting than others
c++ and a reference to borges in the same discussion; how could someone ask for more?
in the same mood there was also a short story of Rudy Rucker where the protagonist wanted to reach by walking a place that was distant aleph1 steps from him…
that’s great! didn’t knew that mp3 is based in something similar to audio grains, it’s really interesting.
there’s a weird effect: sometimes when the sequence of low/high frequencies happens faster it seems to be polyphonic, like you were hearing both at the same time.
i suppose you’re just generating every possible binary combination for that 26ms, right? what bitrate are you using?
also some times the lower frequencies produce some interesting patterns but you can’t almost hear them cause the high ones that are coming next will destroy your ears if you turn up the volume. can there be a way of assigning a relative amplitude depending on the frequency? i suppose there’s no way cause that will make it a different sound and you won’t be going through all the possible combinations?
@naus3a: I’ll post my thesis on this thread when I have my final revision, hopefully I’ll be fitting all these things together smoothly
MP3 has something called “scale factor bands” that allow you to do a sort of EQ on the current sound – but you’re right, if you controlled for this, you wouldn’t be iterating through everything
I think the short tale was from “Infinity and the Mind”. If I manage to find the book I’ll send you a more precise reference
Very interesting work, and intriguing as well. It is very cool that a purely mathematical setup like this one actually produces a very interesting audible output. It is animal like, ‘outer space-ish’, and musical at the same time …
@Rabbagast: Thanks! I think it’s interesting what we associate with “outer space” sounds…
My written thesis was accepted by my school, so there goes another OF+Processing driven degree
Here’s the final version: http://oelf.googlecode.com/files/mcdonald-thesis.pdf
I also posted four 1 hour excerpts on the project page: http://kylemcdonald.net/oelf/ In which you can hear some more of the variety over the 10^450 years of the composition.
And it got posted to Create Digital Music! http://createdigitalmusic.com/2010/05/12/your-hearing-according-to-mp3-sounds-for-humans-played-for-10450-years/
Thanks for posting up your essay, I’m looking forward to reading through it. It’s exciting to see more conceptually sophisticated work done with OF; I like interactive projection as much as the next person but it’s always nice to have something that gives you something meatier to think about in those quiet philosophical moments that sneak up on one every so often.
Very cool thesis Kyle. Neat that you referenced Evan Meaney, I went to undergrad with him
Thanks Joshua + Tim
@Tim: very cool! I heard about his work through Rosa Menkman’s blog. He doesn’t have much on his site, but what he does have I think is fairly advanced in terms of how he is conceptualizing it.