From a separate topic:

Quote:


So I think you're saying that I could take two sound clips, convert them both to the frequency domain, multiply their frequencies at each band, and do an inverse transform? Could I then play the resulting audio w/out playing the two original audio clips? What if I pre-processed the two original clips and stored all their frequencies in some data structure?

Taking a step back, what I'm trying to do is take two sound clips, somehow "merge" them together in an "interesting" way, and play the resulting audio. I was thinking the above method would be a likely good approach. Can you think of another method I could maybe try?  On a purely abstract/conceptual level, I'd like to have one sound (like a musical track) be a kind of "mask" for another sound (like a voice track).

Any info would be much appreciated. Thanks!

R

Hi ddf,

Could you maybe help me out a bit more with this? ie, walk me through the pseudo code needed. Basically, I'm trying to load an audio file, modify it somehow in the frequency domain, and play it back. Is that possible with minim? I guess I'd have to continually update the buffer and do forward FFT's on that? And if I could do all this in an Effect, in generate I could do the inverse transform and stuff that back into a buffer for playback.

I tried modifying the ForwardFFT example in the following way, thinking that TestSignal should hopefully be generating a signal that sounds the same or similar to the AudioPlayer being looped ... but the results are noise.

any advice?

thanks so much
R

import ddf.minim.*;
import ddf.minim.analysis.*;
import ddf.minim.signals.*;

AudioPlayer jingle;
FFT fft;
String windowName;
TestSignal test;
AudioOutput out;

void setup()
{
 size(512, 200);
 // always start Minim before you do anything with it
 Minim.start(this);
 jingle = Minim.loadFile("jingle.mp3");
 jingle.loop();
 fft = new FFT(jingle.left.size(), 44100);
 textFont(createFont("Arial", 16));
 windowName = "None";

 out = Minim.getLineOut(Minim.STEREO, 512);
 test = new TestSignal();
 out.addSignal(test);
}

void draw()
{
 background(0);
 stroke(255);
 fft.forward(jingle.left);
 for(int i = 0; i < fft.specSize(); i++)
 {
   line(i, height, i, height - fft.getBand(i)*4);
 }
 fill(255);
 // keep us informed about the window being used
 text("The window being used is: " + windowName, 5, 20);
}

void keyReleased()
{
 if ( key == 'w' )
 {
   fft.window(FFT.HAMMING);
   windowName = "Hamming";
 }

 if ( key == 'e' )
 {
   fft.window(FourierTransform.NONE);
   windowName = "None";
 }
}

void stop()
{
 // always close Minim audio classes when you finish with them
 out.close();
 jingle.close();
 super.stop();
}

class TestSignal implements AudioSignal
{
 void generate(float[] samp)
 {
   fft.inverse(samp);
 }

 void generate(float[] left, float[] right)
 {
   generate(left);
   generate(right);
 }
}