2 summary:: Overview of FFT UGens
4 related:: Classes/FFT, Classes/IFFT
8 SuperCollider implements a number of UGens supporting FFT based processing. The most basic of these are FFT and IFFT which convert data between the time and frequency domains:
10 chain = FFT(buffer, input)
15 link::Classes/FFT:: stores spectral data in a local buffer ( see link::Classes/Buffer:: ) in the following order: DC, nyquist, real 1f, imag 1f, real 2f, imag 2f, ... real (N-1)f, imag (N-1)f, where f is the frequency corresponding to the window size, and N is the window size / 2.
17 The buffer's size must correspond to a power of 2, and must also be a multiple of SC's block size. The window size is equivalent to the buffer size, and the window overlap defaults to 2. Both link::Classes/FFT:: and link::Classes/IFFT:: use a Sine window by default, the combination of which (i.e. raised sine, that is, sine squared) is a Hanning window.
19 section:: Phase Vocoder UGens and Spectral Processing
21 In between an FFT and an IFFT one can chain together a number of Phase Vocoder UGens (i.e. 'PV_...') to manipulate blocks of spectral data before reconversion. The process of buffering the appropriate amount of audio, windowing, conversion, overlap-add, etc. is handled for you automatically.
25 in = WhiteNoise.ar(0.8);
26 chain = FFT(LocalBuf(2048), in);
27 chain = PV_RandComb(chain, 0.95, Impulse.kr(0.4));
33 PV Ugens write their output data emphasis:: in place ::, i.e. back into the same buffer from which they read. PV UGens which require two buffers write their data into the first buffer, usually called 'bufferA'.
35 d = Buffer.read(s, Platform.resourceDir +/+ "sounds/a11wlk01.wav");
38 { var inA, chainA, inB, chainB, chain;
39 inA = LFSaw.ar([100, 150], 0, 0.2);
40 inB = PlayBuf.ar(1, d, BufRateScale.kr(d), loop: 1);
41 chainA = FFT(LocalBuf(2048), inA);
42 chainB = FFT(LocalBuf(2048), inB);
43 chain = PV_MagMul(chainA, chainB); // writes into bufferA
51 Because each PV UGen overwrites the output of the previous one, it is necessary to copy the data to an additional buffer at the desired point in the chain in order to do parallel processing of input without using multiple FFT UGens. link::Classes/PV_Copy:: allows for this.
54 b = Buffer.alloc(s,2048,1); // use global buffers for plotting the data
55 c = Buffer.alloc(s,2048,1);
60 x = { var inA, chainA, chainB;
61 inA = LFClipNoise.ar(100);
63 chainB = PV_Copy(chainA, c);
64 IFFT(chainA) - IFFT(chainB); // cancels to zero so silent!
68 // IFFTed frames contain the same windowed output data
69 b.plot(\b, Rect(200, 430, 700, 300), nil, nil); c.plot(\c, Rect(200, 100, 700, 300), nil, nil);
73 Note that PV UGens convert as needed between cartesian (complex) and polar representations, therefore when using multiple PV UGens it may be impossible to know in which form the values will be at any given time. FFT produces complex output (see above), so while the following produces a reliable magnitude plot:
75 b = Buffer.alloc(s,1024); // use global buffers for plotting the data
76 a = { FFT(b, LFSaw.ar(4000)); 0.0 }.play;
78 b.getn(0, 1024, { arg buf;
80 z = buf.clump(2).flop;
81 z = [Signal.newFrom(z[0]), Signal.newFrom(z[1])];
82 x = Complex(z[0], z[1]);
83 {x.magnitude.plot}.defer
88 any Synth using PV UGens might not.
90 It is possible to manipulate the FFT data directly within a synth graph (if there doesn't already exist a PV UGen which will do what you want), using the methods pvcalc, pvcalc2, pvcollect. Here's an example which uses the link::Classes/SequenceableCollection:: methods clump and flop to rearrange the order of the spectral bins:
92 c = Buffer.read(s, Platform.resourceDir +/+ "sounds/a11wlk01.wav");
96 var in, numFrames=2048, chain, v;
97 in = PlayBuf.ar(1, c, loop: 1);
98 chain = FFT(LocalBuf(numFrames), in);
100 chain = chain.pvcalc(numFrames, {|mags, phases|
101 /* Play with the mags and phases, then return them */
102 [mags, phases].flop.clump(2).flop.flatten
105 Out.ar(0, 0.5 * IFFT(chain).dup);
111 section:: Multichannel Expansion with FFT UGens
113 Care must be taken when using multichannel expansion with FFT UGens, as they require separate buffers. Code such as this can be deceptive:
115 chain = FFT(bufnum, {WhiteNoise.ar(0.2)}.dup);
118 The above may seem to work, but does not. It does result in two FFT UGens, but as they both write to the same buffer, the second simply overwrites the data from the first, thus wasting CPU and accomplishing nothing.
120 When using multichannel expansion with FFT UGens it is necessary to ensure that each one writes to a different buffer. Here's an example of one way to do this:
123 SynthDef("help-multichannel FFT", { arg out=0; // bufnum is an array
125 in = [SinOsc.ar(200, 0, 0.2), WhiteNoise.ar(0.2)];
126 chain = FFT(LocalBuf([2048, 2048]), in); // each FFT has a different buffer
127 // now we can multichannel expand as normal
128 chain = PV_BrickWall(chain, SinOsc.kr(-0.1));
129 Out.ar(out, IFFT(chain));
133 // or using global buffers
135 b = {Buffer.alloc(s,2048,1)}.dup;
138 SynthDef("help-multichannel FFT", { arg out=0, bufnum= #[0, 1]; // bufnum is an array
140 in = [SinOsc.ar(200, 0, 0.2), WhiteNoise.ar(0.2)];
141 chain = FFT(bufnum, in); // each FFT has a different buffer
142 // now we can multichannel expand as normal
143 chain = PV_BrickWall(chain, SinOsc.kr(-0.1));
144 Out.ar(out, IFFT(chain));
145 }).play(s,[\bufnum, b]);
149 Note that dup on a UGen just makes a reference to that UGen, because UGen defines -copy to simply return the receiver. (See link::Classes/UGen:: for more detail.)
156 Code like code::IFFT(chain).dup:: is found throughout the PV help files , and is just a convenient way to copy a mono signal to stereo, without further computation.
158 See also link::Guides/Multichannel-Expansion::.
160 section:: PV and FFT UGens in the Standard Library
162 The following PV UGens are included in the standard SC distribution:
164 ## link::Classes/FFT:: || Fast Fourier Transform
165 ## link::Classes/IFFT:: || Inverse Fast Fourier Transform
166 ## link::Classes/PV_Add:: || complex addition
167 ## link::Classes/PV_BinScramble:: || scramble bins
168 ## link::Classes/PV_BinShift:: || shift and stretch bin position
169 ## link::Classes/PV_BinWipe:: || combine low and high bins from two inputs
170 ## link::Classes/PV_BrickWall:: || zero bins
171 ## link::Classes/PV_ConformalMap:: || complex plane attack
172 ## link::Classes/PV_Copy:: || copy an FFT buffer
173 ## link::Classes/PV_CopyPhase:: || copy magnitudes and phases
174 ## link::Classes/PV_Diffuser:: || random phase shifting
175 ## link::Classes/PV_HainsworthFoote:: || onset detection
176 ## link::Classes/PV_JensenAndersen:: || onset detection
177 ## link::Classes/PV_LocalMax:: || pass bins which are a local maximum
178 ## link::Classes/PV_MagAbove:: || pass bins above a threshold
179 ## link::Classes/PV_MagBelow:: || pass bins below a threshold
180 ## link::Classes/PV_MagClip:: || clip bins to a threshold
181 ## link::Classes/PV_MagFreeze:: || freeze magnitudes
182 ## link::Classes/PV_MagMul:: || multiply magnitudes
183 ## link::Classes/PV_MagDiv:: || division of magnitudes
184 ## link::Classes/PV_MagNoise:: || multiply magnitudes by noise
185 ## link::Classes/PV_MagShift:: || shift and stretch magnitude bin position
186 ## link::Classes/PV_MagSmear:: || average magnitudes across bins
187 ## link::Classes/PV_MagSquared:: || square magnitudes
188 ## link::Classes/PV_Max:: || maximum magnitude
189 ## link::Classes/PV_Min:: || minimum magnitude
190 ## link::Classes/PV_Mul:: || complex multiply
191 ## link::Classes/PV_PhaseShift:: || shift phase of all bins
192 ## link::Classes/PV_PhaseShift270:: || shift phase by 270 degrees
193 ## link::Classes/PV_PhaseShift90:: || shift phase by 90 degrees
194 ## link::Classes/PV_RandComb:: || pass random bins
195 ## link::Classes/PV_RandWipe:: || crossfade in random bin order
196 ## link::Classes/PV_RectComb:: || make gaps in spectrum
197 ## link::Classes/PV_RectComb2:: || make gaps in spectrum
198 ## link::Classes/UnpackFFT::, link::Classes/PackFFT::, link::Classes/Unpack1FFT:: || "unpacking" components used in pvcalc, pvcalc2, pvcollect (can also be used on their own)
200 For a full list of FFT UGens, see strong::UGens>FFT:: in the link::Browse#UGens>FFT:: page.