Between Pitch and Noise
Preamble
An important corollary when talking about pitch is to talk about unpitched sound or noise. In periods of music dominated by pitched sounds, disintegration or erosion of pitch into noise can be an important technique to drive a set forward, or just provide sonic interest. I find a lot of this kind of thing in the transformations of instruments within Holden's Music for example. Here are some techniques to achieve this.
SinOscFB
A Ugen I use a lot (read: far too much) is SinOscFB, a 'sine oscillator that has phase modulation feedback'. I've always been a big fan of bare sine waves, and SinOscFB's feedback
argument allows a sine wave to be modulated into noise and back very easily, with extreme modulations creating a strange-sounding degraded sine wave.
//SinOscFB - A sine wave that can move between pitch and noise and noisy pitch
(
//polling the modulation of the 'feedback' argument, to show the way in which SinOscFB degrades sine waves
~sinfbstatic = {SinOscFB.ar([330,440],XLine.kr(0.1,500,60).poll(10),0.6)};
~sinfbstatic.play;
)
A stalwart of my SynthDef arsenal is sinfb
, a SinOscFB
Ugen inside of an Env.perc
which is used to control its amplitude curve. This SynthDef is very flexible - great for basses, melodies and chords, but also great for flexibly turning melodic riffs into textural noise, as well as blending the two. Notice that from values 0.0
to 20.0
there is a full spectrum from clean sine wave to noise feedback, any values above 30.0
will blend the two and are what I would consider 'extreme modulation'. In general usage during sets I tend to use the range 0.0 to 3.0, as anything above tends to be too noisy and interferes with the percussion i'm using.
//a pattern I use regularly with its feedback being modulated from 0 to 20. Notice the difference in sound across the spectrum
(
~sinfb = Pbind(\instrument,\sinfb,\scale,Scale.minor,\octave,[3,4,5],\degree,Pseq([0,0,4,5],inf),\dur,Pbjorklund2(3,8)/4,\amp,0.3,\fb,0.1,\rel,0.3);
~feedback = {SinOsc.kr(0.1,-1,1).range(0,20.0).poll(30)};
~sinfb.set(\fb,~feedback);
~sinfb.play;
)
Harmonic series and extreme pitch values
In 4.2 I talked about the Harmonic Series. An interesting quality of using a fundamental frequency to determine the pitch of various NodeProxies by multiplying that fundamental frequency to create a scale structure.
Some interesting techniques for distorting this harmonic series technique into the territory of noise are extreme modulation, which pushes the frequency into supersonics (and sometimes back again):
//Extreme modulation of fundamental frequency
//taking the up-down scale given in the 'riffs' section
(
//up
p.clock.tempo = 2.4;
~r = {75};
~sinfb1 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10),inf)*Pkr(~r),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,1.4),\rel,0.1);
~sinfb2 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10).reverse,inf)*Pkr(~r),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,1.4),\rel,0.1);
~sinfb3 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10).scramble,inf)*Pkr(~r),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,2.0),\rel,0.1);
~sinfb1.play;~sinfb2.play;~sinfb3.play;
)
//moving the frequency up and beyond sensible into supersonics - after reading around 5000Hz some interesting aliasing starts to happen
(
~r = {XLine.kr(75,8000,60).poll(10)}
)
//and even further, lower frequencies start reappearing
(
~r = {XLine.kr(8000,30000,60).poll(10)};
)
//using very extreme modulation also gives some interesing results
(
~r = {LFNoise1.kr(0.2).range(30000,90000).poll(10)};
)
And extreme pitch values - which appear to rise continually into supersonic frequencies and aliasing, and then looping back to the bottom of the pitch scale:
//extreme multiplaction of fundamental frequency
//using the previous example, a NodeProxy holding a second multiplier is added onto the \freq argument of each Pbind
(
~r = {75};
~mult = {1};
~sinfb1 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10),inf)*(Pkr(~r)*Pkr(~mult)),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,1.4),\rel,0.1);
~sinfb2 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10).reverse,inf)*(Pkr(~r)*Pkr(~mult)),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,1.4),\rel,0.1);
~sinfb3 = Pbind(\instrument,\sinfb,\freq,Pseq((1..10).scramble,inf)*(Pkr(~r)*Pkr(~mult)),\dur,0.25,\amp,0.3,\fb,Pwhite(0.1,2.0),\rel,0.1);
~sinfb1.play;~sinfb2.play;~sinfb3.play;
)
//increase the multiplcation over time using a .round on a Line.kr UGen. Listen to how the scale is distorted as the multiplcation increases, eventually ending as a series of pulses
(
~mult = {Line.kr(1,60,60).round(1).poll(5)}
)
Chaos UGens
SuperCollider has support for UGens that use Chaos Theory for synthesis - the Chaos UGens (note that there are also a number of additional Chaos UGens in sc3-plugins which are worth having).
While (at the time of writing) I don't know a whole lot about the particularities of chaos theory works, but the Chaos UGens are great for creating musical structures that move freely between pitched sound and noise, and these are usually handled both in the equation variables of the UGens as well as the initial conditions.
I'll use HenonN as an example of the use of chaos theory to move between melody, noise and percussion:
//HenonN - Chaos synths and moving between pitch and noise
(
//henon using a minor pentatonic scale at a high octave.
//The chaos Ugens will need some experimentations if you want subtle variance in sound
//For Henon I found that an a value of 1.3 and a b value of 0.3 renders a pitch in a pattern pretty reliably
//note that the pitches aren't quite the same as 'concert pitch'
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,8,\dur,Pbjorklund2(3,8)/4,\a,Pexprand(1.3,1.3),\b,Pexprand(0.3,0.3),\atk,0,\sus,0,\rel,Pexprand(0.1,0.1),\amp,1);
~henon.play;
)
//increase the variation in the a and b arguments to add more noise to the mix
(
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,8,\dur,Pbjorklund2(3,8)/4,\a,Pexprand(1.3,1.31),\b,Pexprand(0.3,0.31),\atk,0,\sus,0,\rel,Pexprand(0.1,0.1),\amp,1);
)
//notice that this gets very noisy VERY fast.
//adding a little more possiblity to the Pexprands in a and b turns it into pure noise very very fast, while still retaining a little of its pitched character
(
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,8,\dur,Pbjorklund2(3,8)/4,\a,Pexprand(1.3,1.35),\b,Pexprand(0.3,0.35),\atk,0,\sus,0,\rel,Pexprand(0.1,0.1),\amp,1);
)
//even more and noises become cut off and non-sounding.
//the cut off sounds would sound as DC bias, but the SynthDef \henon has a LeakDC on its output to prevent this as it can damage sound systems and is generally quite an unpleasant thing to deal with.
(
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,8,\dur,Pbjorklund2(3,8)/4,\a,Pexprand(1.3,1.45),\b,Pexprand(0.3,0.55),\atk,0,\sus,0,\rel,Pexprand(0.1,0.1),\amp,1);
)
//at this point decreasing the \dur and \rel value turns it into rhythmic percussion
(
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,8,\dur,0.25,\a,Pexprand(1.3,1.45),\b,Pexprand(0.3,0.55),\atk,0,\sus,0,\rel,Pexprand(0.01,0.1),\amp,1);
)
//more extreme possible values - \dur varied, octaves doubled up, more variation in a and b values, more octaves
(
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,[8,12,9,10],\dur,Pwrand([0.25,Pbjorklund2(Pwhite(3,5),8,1)/4,Pseq([0.125],4)],[7,4,1].normalizeSum,inf),\a,Pexprand(1.2,1.55),\b,Pexprand(0.21,0.55),\atk,0,\sus,0,\rel,Pexprand(0.01,0.6),\amp,1);
)
//against a kick drum it takes on a really strange character
(
~k = Pbind(\instrument,\bplay,\buf,d["k"][0],\dur,1,\amp,1);
~k.play;
)
A thing to note about the Chaos synths is the type of interpolation used - taking Henon as an example; HenonC, HenonL and HenonN stand for Cubic, Linear and None respectively. The sonic effect of the type of interpolation used is in the 'smoothness' of the sound, with Cubic being the most smooth and None being the least.
//sound of different types of interpolation
//the default in my SynthDefs.scd file is currently to use none:
(
SynthDef(\henon,
{arg freq=440,mfreq=440,a=1.3,b=0.3,x0=0.30501993062401,y0=0.20938865431933,atk=0.01,sus=1,rel=1,ts=1,out=0,pan=0,amp=0.3;
var sig,env;
sig = Henon2DN.ar(freq,freq+mfreq,a,b,x0,y0,amp);
env = EnvGen.ar(Env.linen(atk,sus,rel),1,1,0,ts,2);
sig = LeakDC.ar(sig);
sig = sig*env;
Out.ar(out,Pan2.ar(sig,pan));
}).add;
);
//the example earlier, with no interpolation (default)
(
p.clock.tempo = 2.2;
~henon = Pbind(\instrument,\henon,\scale,Scale.minorPentatonic,\degree,Pseq([0,2,4,6,7],inf),\octave,[8,12,9,10],\dur,Pwrand([0.25,Pbjorklund2(Pwhite(3,5),8,1)/4,Pseq([0.125],4)],[7,4,1].normalizeSum,inf),\a,Pexprand(1.2,1.55),\b,Pexprand(0.21,0.55),\atk,0,\sus,0,\rel,Pexprand(0.01,0.6),\amp,1);
~henon.play;
)
//now with Linear interpolation
(
SynthDef(\henon,
{arg freq=440,mfreq=440,a=1.3,b=0.3,x0=0.30501993062401,y0=0.20938865431933,atk=0.01,sus=1,rel=1,ts=1,out=0,pan=0,amp=0.3;
var sig,env;
sig = Henon2DL.ar(freq,freq+mfreq,a,b,x0,y0,amp);
env = EnvGen.ar(Env.linen(atk,sus,rel),1,1,0,ts,2);
sig = LeakDC.ar(sig);
sig = sig*env;
Out.ar(out,Pan2.ar(sig,pan));
}).add;
);
//now with Cubic interpolation
(
SynthDef(\henon,
{arg freq=440,mfreq=440,a=1.3,b=0.3,x0=0.30501993062401,y0=0.20938865431933,atk=0.01,sus=1,rel=1,ts=1,out=0,pan=0,amp=0.3;
var sig,env;
sig = Henon2DC.ar(freq,freq+mfreq,a,b,x0,y0,amp);
env = EnvGen.ar(Env.linen(atk,sus,rel),1,1,0,ts,2);
sig = LeakDC.ar(sig);
sig = sig*env;
Out.ar(out,Pan2.ar(sig,pan));
}).add;
);