PolyGranuSyn 5

[Tim Kleinert]

This is an update on the polyphonic granular synth I made some time ago. It incorporates a completely new elastic audio core circuit (also used in the new DeluxeLooper2 patch). I also added other new stuff. The old PolyGranuSyn patch used a constant grain-clock, which bled into the source material at high frequencies. Because the grain-clock could track the keyboard, this was actually a nice way to create tonally playable sounds from arbitrary sources. The new elastic audio engine however preserves the pitch information of the source material far more faithfully. So there was need for another way of "imprinting a pitch of choice" on its output. The Roland VSynth does this with its so-called "sideband filter", which is Rolands involuted name for the comb filter . I tried this first, but then opted for a Karplus-Strong-type resonator (the same thing with a lowpass filter in the feedback loop) instead since it sounds more organic. I used the StringOsc module for this, but with an external feedback loop to get negative feedback too.

Controls/Features:

INPUT Source, CHANNEL Source: Select audio path for capturing.
SAMPLE Capture: captures 2 seconds of audio. (NOT 2.7 seconds. The resonator also needs delay line memory.)
POSITION <scrub>: scrub around in the sample, independent of pitch. Also assigned to the modwheel.
TIME <Dir>: alter playback direction, independent of pitch, from -64 (backwards) to 0 (freeze) to +64 (forwards). Also assigned to the pitch stick.
TIME KBT: Connects the playback direction to keyboard tracking (like regular sampling).
PITCH Semi: Tune the sampled audio in semitones, independent of speed.
PITCH KBT: lets the sample track the keyboard, independent of speed.
GRAIN Size: Controls the size of grains.
SIZE Jitter: Adds jitter to the grain size, making a more animated sound.
MODE Source ("fixed" and "intelligent"): "Fixed" means that each consecutive grain always contains a fixed amount of audio waveform cycles, regardless of how long those cycles are. As a consequence, low-frequency audio material will produce longer grains than high-frequency material, since the waveform cycles are longer. On the other hand, "Intelligent" mode interprets the measured cycles according to their length, weighing smaller (high-frequency) cycles less than bigger (low-frequency) ones. The result is that the grain size will be more uniform, regardless of the audio material.
RESON8R In Lvl: Attenuator for the audio going into the resonator. Has to be backed off at high feedback settings to prevent clipping.
IN-HPF Freq: A static highpass filter, pre-resonator.
RESON8R Semi, Fine, Damp: Resonator pitch and damping controls.
RESON8R Feedbak: Bipolar resonator feedback control.
FEEDBACK Attack,Release: Envelope for the feedback. Is useful to prevent loud pops which can occur with voice-stealing.

POSITION ENV Attack/Decay, LOOP On/Off, ENV Amount, AMOUNT output type: Loopable envelope for the sample position.
PITCH ENV: ditto for sample pitch.
R-PITCH ENV: ditto for resonator pitch.

VOLUME A, D, S, R: I don't remember what this is for...
CHORUS, DELAY, REVERB: The usual suspects.

I didn't see much point in doing variations since the settings are dependent on the captured source sample material. However, variation 1 is a plain timestretching demonstration, and variation 2 and 3 will turn almost any source material into ethereal ambient soup.


Uses a DSP per voice (sorry...), so 7 voices from an expanded G2. I think it's worth the expense. I get sounds with this patch that are totally unlike anything else I've ever heard from the G2.

Have fun.

tim

---

PolyGranuSyn5 TK.pch2
Description:	Polyphonic granular synth	Download
Filename:	PolyGranuSyn5 TK.pch2
Filesize:	6.23 KB
Downloaded:	2202 Time(s)

---

[glitched]

I just spent 3 hours playing with this. Monomachine IN-->Granulizer. MnM sequences the G2. Amazing!

Dang, I had stuff to do today!