[gorjuss] The story of "The Sound of the Internet"

[Giles Turnbull]

My latest for The Morning News is "The Sound of the Internet", which
you can read / listen to here:

http://www.themorningnews.org/article/the-sound-of-the-internet

I knew quite early on what sort of thing I wanted to write. I knew I
could extract texts (potentially very long ones) that would at least
*represent* my use of the internet - pings, traceroutes, browser
histories, that sort of thing. And I was very taken with the idea of
converting images into ASCII-art, then running the ASCII thorough
music-making software. What would photos sound like? Anyway, I had a
plan. What I stumbled over was the practical reality of producing the
sounds.

I knew about the work of Slub: http://slub.org/

... and I knew that while I might not be able to code my own sounds, I
might be able to find software that would help me. Thus began a
time-consuming search for the right sort of thing, concentrating on
the idea of generative music software.

En route I found Karlheinz Essl:

http://www.essl.at/

... and his various software projects, such as Lexikon-Sonate:

http://www.essl.at/works/Lexikon-Sonate.html

... which were all fascinating and fun to play with, but I couldn't
see a way of doing a conversion from text into music, and that was the
format I'd settled on early on in the process.

More discoveries. Petri:

http://www.sineqube.com/blog/?page_id=153

... which looks amazing, and would be something I'd love to try, but
wasn't what I needed for this.

Plogue: http://www.plogue.com/ (didn't understand what was going on
here, to be honest)

Tapper: http://www.musanim.com/tapper/

cgMusic: http://codeminion.com/blogs/maciek/2008/05/cgmusic-computers-create-music/
(Windows only)

There's a list of loads more Windows options here:

http://conceptual-algorithmic-music.blogspot.co.uk/p/algorithmic-software-list.html

Finally, after much hunting around and experimenting with different
search terms, I stumbled upon Danny Ayers' Web Beeps:

http://webbeep.it/

I knew immediately that this was precisely the sort of tool I'd been
looking for. Something into which I could paste my various text
sources, and get music back. Unfortunately, there was a limit of 63
characters per session, and some of the text files I'd prepared were
many thousands of characters long. I contacted Danny and he was
wonderful - immediately offering to see if he could code a work-around
that might let me input more than 63 chars. (He did, as well: if you
enter more than that number into the text field on his web page, it
will now break your text into 63-char strings, process each one, then
assemble them into one mp3 with a short gap between each string. You
can hear the result in the track titled "The first web page":
http://soundcloud.com/andrewwomack/first-web-page .)

(Meta note: the sound files were uploaded to SoundCloud by my patient
Morning News editor Andrew Womack, which is why they appear there in
an account with his name.)

Danny was eager to help out, and did so despite being busy with normal
life. Thanks Danny.

I had lots of help from others, too. Oliver Humpage at Bristol's
Watershed arts centre was happy to spend some of his valuable time
taking me on a guided tour of the building's hidden IT underworld. We
lurked in server rooms and I recorded their hummings and purrings.
Oliver even found a server that wasn't actually doing anything, so we
had a chance to reboot it without upsetting anyone, and record the
whoosh of its restart. Thanks Oliver.

http://www.watershed.co.uk

Right at the start of the project, I'd mentioned what I was after on
[underscore], Bristol's mailing list for web folk and geeks. Oliver
Kohll saw that message and was inspired to start his own coding. The
result was TextSound:

https://github.com/okohll/TextSound

Again, Oliver was incredibly helpful, assisting me with the task of
getting TextSound installed and working on my computer, which involved
various bits of messing around on the command line and fiddling with
Java Preferences and whatnot. He was very patient. Thanks Oliver.

TextSound outputs MIDI, so my final task was seeking out MIDI player
apps for OS X. I found a few in the Mac App Store, but the one that
really struck me was MidiTrail:

https://github.com/okohll/TextSound  (Yes, it's on SourceForge *and*
the App Store.)

It's a MIDI visualiser, which means you can fly through your music as
its playing. I was able to navigate the output from TextSound,
changing the playback speed, skipping silent sections, and generally
messing about, using the keyboard shortcuts in MidiTrail. In effect,
it allowed me to "perform" the output in a way. The result was the
final track, "Trace Route":

http://soundcloud.com/andrewwomack/trace-route (the final couple of
minutes are the best bit)

The original TextSound output was 35 minutes long; I got this down to
a more manageable 9 minutes, although it still has long silences in
it.