Music Theory, beautiful instruments, high quality studio samples...

Even perfectly synthetized sounds, out of Mathematics or even Machine Learning...

Are not as great, as they sound.

What sounds great, are complex, noisy, and surprising sounds.

A song with predictable sounds, will only be interesting for a few days.

But a song with unheard of madness, can remain interesting for weeks.

To be specific, I am thinking about workout songs, songs that help to manage workout duration and tempo.

Surprisingly workout songs, are a really good judge, of how long a song will stay interesting.

At the end of the poem I will play, a very basic beat, little over a minute long.

It consists of fraction of a second short samples, taken from the 1929 hit by Jimmie Rodgers entitled Waiting for a Train

Here is a quick listen to show you how different the final, the test beat is.

(play sample song)

So, the song I sampled these snippets from, is extremely noisy, and relied on very old technology.

When the random snippets come together, any noise that can be picked up, only adds to the creation.

Now allow me to play the snippets, I picked out of Jimmie's song, I'll repeat them three times.

(sample 3x)

I think it is safe to say, any complex sounding thing can be a great source of instruments.

If you need a drum, just use a low-pass filter, on something somewhat loud.

Aside from microscopic snippets of public domain songs, street noise can work, a collection of cans...

Even tapping the microphone a bit, can give you a source of instruments.

The thing that is especially interesting here, is the randomness that makes each song...

Not only madly unique, but also a game like challenge.

(play the hokey little song)