//---------------------------------------------------------------------
// name: mfcc-basic.ck
// desc: mel-frequency cepstral coefficients
// basic MFCC feature extractor
//
// version: need chuck version 1.5.0.0 or higher
// sorting: part of ChAI (ChucK for AI)
//
// uncomment for MFCC API:
// MFCC.help();
//
// author: Ge Wang (https://ccrma.stanford.edu/~ge/)
// date: Spring 2023
//---------------------------------------------------------------------
// synthesis / analysis network
adc => FFT fft =^ MFCC mfcc => blackhole;
// set number of coefficients in MFCC (how many we get out)
// 13 is a commonly used value; using less here for printing
7 => mfcc.numCoeffs;
// set number of mel filters in MFCC
10 => mfcc.numFilters;
// set FFT size
2048 => fft.size;
// set window type and size
Windowing.hann(fft.size()) => fft.window;
// our hop size (how often to perform analysis)
fft.size()::samp => dur HOP;
// let one FFT-size of time pass (to buffer)
fft.size()::samp => now;
// control loop
while( true )
{
//----------------------------------------------------------------
// upchuck() computes our MFCC, automatically computing upstream
// dependencies connected to it using =^ (e.g., FFT)
//----------------------------------------------------------------
mfcc.upchuck();
// print
cherr <= "MFCC: "; // print the MFCC results for( int i; i < mfcc.fvals().size(); i++ ) { cherr <= mfcc.fval(i) <= " "; } // endline cherr <= IO.newline(); // advance time HOP => now;
}