You can download the patch here.
My colleague Pinar Temiz was talking up filterbanks yesterday, and after i heard it a bit i was sold. So today i had prepared a “filterbank” subpatch which is just 10 bandpass filters with the frequency rising with a ratio. so the [pd filterbank] object has 4 inlets. Audio in, then fundamental frequency, then ratio with which is next filter is tuned to and then the q for all of them. I started of using rjlib’s drunk and noise to create a nice synthetic wind sound. But then i started to add other sounds to the mix. At that point something magical seems to happen. THis is also where the audio of the videocapture improved because pd’s writesf~ had failed on me the previous 2 times (going back to bullet proof Sony m10 tomorrow).
Anyways, back to this magic. After 45-50 seconds or so, listen closely in the noise. You will hear a distant Gregorian singing voice or voices singing something “aaaah”. These beautiful slow melodies appear out of nowhere singing a song I have never heard. Good Times!
If you’re going to try this patch out, play around with the input material. I’m sure you can come up with something better then I did. If you do, send it back. Would love to hear it!