Pprob | SuperCollider 3.14.0 Help

Description

Creates an integral table on instantiation (cpu intensive) which is then used by the streams to generate random values efficiently.

Class Methods

Pprob.new(distribution, lo: 0.0, hi: 1.0, length: inf, tableSize)

Arguments:

distribution	desired probability distribution (histogram).
lo	lower bound of the resulting values.
hi	upper bound of the resulting values.
length	number of values to repeat.
tableSize	resample table to this size. If the size of the distribution is smaller than 64, it is (linearly) resampled to this minimum size.
(distribution)	set the distribution, the table is recalculated.
(tableSize)	set the resample size, the table is recalculated.

Examples

// a consistency test
(
var a = Pprob([0,0,0,0,1,1,1,1,3,3,6,6,9].scramble);
var b = a.asStream;
b.nextN(800).sort.plot("sorted distribution");
b.nextN(800).sort.plot("sorted distribution, again");
)

// comparison: emulate a linrand
(
var a, b, x, y;
a = Pprob([1, 0]);
x = Pfunc({ 1.0.linrand });

b = a.asStream;
y = x.asStream;

postf("Pprob mean: % linrand mean: % \n", b.nextN(800).mean, y.nextN(800).mean);

b.nextN(800).sort.plot("this is Pprob");
y.nextN(800).sort.plot("this is linrand");
)

// compare efficiency

bench { Pprob([0, 1]) } // this is fairly expensive
bench { 16.do { Pseq([0, 1] ! 32) } }

x = Pprob([0, 1]).asStream;
y = Pseq([0, 1], inf).asStream;

bench { 100.do { x.next } }; // this very efficient
bench { 100.do { y.next } };

// sound example
(
SynthDef(\help_sinegrain,
    { |out = 0, freq = 440, sustain = 0.05|
        var env;
        env = EnvGen.kr(Env.perc(0.01, sustain, 0.2), doneAction: Done.freeSelf);
        Out.ar(out, SinOsc.ar(freq, 0, env))
    }).add;
)

(
var t;
a = Pprob([0, 0, 1, 0, 1, 1, 0, 0], 60, 80);
t = a.asStream;
Routine({
    loop({
    Synth(\help_sinegrain, [\freq, t.next.midicps]);
    0.01.wait;
    })
}).play;
)

a.distribution = [0, 1];
a.distribution = [1, 0];
a.distribution = [0, 0, 0, 0, 1, 0];
a.distribution = [0, 1, 0, 0, 0, 0];

// higher resolution results in a more accurate distribution:
a.tableSize = 512;
a.tableSize = 2048;

helpfile source: https://github.com/supercollider/supercollider/tree/main/HelpSource/Classes/Pprob.schelp
link::Classes/Pprob::

Pprob : Pattern : AbstractFunction : Object

Description

Class Methods

Pprob.new(distribution, lo: 0.0, hi: 1.0, length: inf, tableSize)

Arguments:

Inherited class methods

Instance Methods

Inherited instance methods

Undocumented instance methods

.distribution

.distribution = list

.embedInStream(inval)

.hi

.hi = value

.length

.length = value

.lo

.lo = value

.table

.tableSize

.tableSize = n

Examples