1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
|
CLASS:: FFTSubbandFlatness
summary:: Spectral flatness, divided into subbands
categories:: UGens>Analysis, UGens>FFT
related:: Classes/FFTSubbandPower
DESCRIPTION::
Calculates the spectral flatness measure, in the same manner as FFTFlatness, but divides the spectrum into (adjacent, non-overlapping) subbands, so returns separate measures for the different subbands.
The parameter cutfreqs can only be specified on initialisation - they can't be modulated.
CLASSMETHODS::
METHOD:: kr
argument:: chain
argument:: cutfreqs
must be an array of frequencies. For example, to divide a 44100Hz signal into three subbands we might specify [ 5512, 11025 ] as the cutfreqs, giving subbands of 0-5512Hz, 5512-11025Hz, and 11025-22050Hz. (Frequencies above the Nyquist frequency are not included.)
returns:: #[flatness1, flatness2, ... flatnessN+1]
EXAMPLES::
In this example we take a "flat" sound (white noise) and an "unflat" sound (a sawtooth wave) and mix them together bandwise using [PV_BinWipe]. With the spectrum divided into three bands you should be able to see the results.
code::
s.boot;
b = Buffer.alloc(s,2048,1);
c = Buffer.alloc(s,2048,1);
(
x = {
var inA, inB, chainA, chainB, chain, vals, cutfreqs;
//in = LPF.ar(WhiteNoise.ar, MouseX.kr(10,10000, 1));
inA = WhiteNoise.ar;
inB = Saw.ar;
chainA = FFT(b.bufnum, inA);
chainB = FFT(c.bufnum, inB);
chain = PV_BinWipe(chainA, chainB, MouseX.kr(-1, 1));
vals = FFTSubbandPower.kr(chain, [300, 500, 1500]);
vals[3].poll(10, "hi ");
vals[2].poll(10, "med");
vals[1].poll(10, "low");
Out.ar(0, IFFT(chain).dup * 0.2);
}.play(s);
)
x.free;
::
|
