I should probably clear up the whole "math" thing - what I did was much more akin to DATA-BENDING, and parameter mapping, not math per se. Essentially, 10000 digits of Pi were used as a giant data set to drive things. In fact, I'm not sure how different it would sound with random numbers as opposed to the PI data set - though the PI version is determinate and always plays the same, I believe. The fibonacci sequence structure was there though, and possibly the only overt display of math, as that sequence controlled when different parts began to fade in or out, within a set plan I'd laid out on paper.
Unfortunately, although the sound at the start is synthetic, it is from a VST, and not synthesized in Pure Data. I'm not convinced I could even run a big, polyphonic subtractive synth built in PD currently, I fear with so much running it would most likely eat my entire CPU on WinXP, so using VST's in PD has been my compromise. Actually, because of that and time constraints, it was either use VST's, or ditch Pure Data altogether for the project. The piece as a whole did push my CPU pretty much to the limit, especially with the reverb added.
The quacking = simple fm chirps, that wasn't how I originally intended to realized that part of the composition, but the deadline for the piece was today (on the microsound list). I do intend to make a new version next week though, and if I'm lucky I can synthesize something closer to my original intention in PD. Also, as I've said before, low-level DSP isn't really my thing, I'm more into the composition side of things and at least some higher level modular components.
Anyway, this piece realizes only about 1/3 of my original plan, though now that the important abstractions are built I could realize new versions more quickly.
I'll have a look at those abstractions, thanks! I must mention that none of this stuff would be possible without ZEXY, Frank's list-abs collection (and especially, I must mention the [line-interp] which I only just discovered after working on a different solution to the same problem, and Grill's [vst~]. So a big thanks to all those who contributed in those projects!
~David
On 3/24/07, padawan12 padawan12@obiwannabe.co.uk wrote:
Thanks for sharing that I enjoyed it. Some lovely sounds. I wasn't much taken with Mr quacky at the start :), but once that was over I enjoyed the textures and decelerating rythms. I don't hear how the maths works, but it works for me.
If you like textures based on dilating/warping events the attached patches might inspire some ideas. The first is for an elastic object that gives up a little kinetic energy on each bounce, the other is a fragmentation model for something "falling apart" that uses bifurcation. They're for bouncing balls and breaking glass for me, but I think they have compositional uses if you twist em a bit. If that's a synthetic voice would you share your choir patch?
cheers, Andy
On Fri, 23 Mar 2007 15:06:09 -0600 "David Powers" cyborgk@gmail.com wrote:
PiData, by Cyborg K aka David A. Powers. All synthesis and sequencing was done by a single Pure Data patch, rendered in a single take, then normalized in an external sound editor. More complex synthesis and fx were done with the aid of the Pure Data [vst~] object. The custom [getpi] abstraction was used to output n digits of the first 10,000 digits of Pi at a time: all variable parameters, whether harmonic, rhythmic, or synthetic were driven by instances of the [getpi] abstraction. Overall rhythmic structure was based on the fibonacci sequence: 8 + 13 + 21 + 34 + 21 + 13 + 8 + 34 + 21 + 13 + 8 = 194 seconds = 3:14. The piece took about 15-20 hours to complete, including building abstractions from scratch within Pure Data.
Link: http://www.cyborgk.com/audio/cyborgk-pi_data.mp3
*PS. This is the first time I have composed an entire piece in the Pure Data environment...
~David
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list