«003_054» by evilshadowtwin
on 03 Aug'13 03:48 in1
SynthDef('003_054', {|out=0,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y| Out.ar(out, Pan2.ar(Normalizer.ar(( CuspL.ar(e, e, SawDPW.ar(x, SyncSaw.ar( PMOsc.ar(a, g, k, a, x, f ), j, b, p ), v, i ), SyncSaw.ar(u, HenonL.ar(i, m, c, t, l, r, f ), GbmanL.ar(m, x, s, f, y ), n ), o, b ))*( PMOsc.ar( LFSaw.ar(l, b, p, q ), t, x, LFTri.ar(l, LFPulse.ar( Spring.ar(d, x, f ), v, i, k, h ), LFPulse.ar(l, n, h, n, v ), r ), k, q ))*( Formant.ar(d, x, LFDNoise1.ar(l, LFCub.ar(k, l, s, v ), LFClipNoise.ar(f, m, p ) ), SinOsc.ar(g, CML0.ar(g, l, i, c, i, f ), LFPulse.ar(l, f, c, b, s ), LFNoise1.ar(l, x, v ) ), LPF.ar( SyncSaw.ar(m, q, u, r ), BrownNoise.ar(t, h ), f, n ) ))*( QuadL.ar(x, LFNoise1.ar( CombL.ar(e, 1, x, w, d, c ), u, Blip.ar(b, q, m, b ) ), LFPar.ar(h, LFClipNoise.ar(u, v, v ), LFClipNoise.ar(n, d, j ), m ), k, SinOsc.ar(x, l, o, l ), m, HenonL.ar(p, w, h, p, a, j, w ) ) )))) }).play(args: #[ a, 0.45186975651959, b, 4.3496208121178, c, -3.5050481728403, d, 0.21674529882104, e, 3.6302846277664, f, 2.0252093014502, g, -0.50985145568848, h, 4.7362252591183, i, -1.1713491709346, j, 11.646696339178, k, 0.66817343235016, l, 0.11935562425443, m, 1.350574016571, n, -1.7757006468997, o, 2.0951379285004, p, 3.5410180366402, q, 0.39763283729553, r, 0.93207371234894, s, 6.2958073430114, t, 6.1286854287138, u, 4.8556221894114, v, 4.8556221894114, w, 1.7277424641031, x, 4.8556221894114, y, 2.9251942534878 ])
reception
got "class not defined" - CML0
p.s. why single line? )
CML0 is part of Yota Morimoto's nlugens: https://github.com/yotamorimoto/nl4sc
code generated by an evolutionary algorithm, not for human consumption.
thanks for pointing! what did you choose as fitness function?
the fitness is determined by comparing machine listening data of the generated functions to those of an existing sound
interesting! what did you use as a metric of data? ..power spectrum / whole record / something else?
mean + stddev of 20 MFCC, SpecFlatness, SpecCentroid, Amplitude. any synthdef producing even a single NaN or inf must DIE...