i think that may work - although i don't know why you're not being able to do the conversion. better yet i hope that ess and sonia will unite into one super-lib one of these days (or at least consolidate some functions)

in the meantime, which features are you missing in sonia?

amit

Ok, got the type conversion working. Ess now plays the grabbed data. However, my fears about performance have been confirmed. There's a definite pause in the sketch when the data is transferred. Is there any way to do this in the background, say using threads? I know of the ability to use threads in Java, but have never done so. Is it even possible in Processing?

~pN~

Krister,

Here's the code. Thanks for taking a look! ~pN~


//THIS PROGRAM USES SONIA TO CAPTURE LIVE SOUND INPUT AND//
//COPIES THE SAMPLE TO AN ESS CHANNEL OBJECT FOR PLAYBACK//

import krister.Ess.*;
import pitaru.sonia_v2_9.*;

Sample mySample;
Channel myChannel;
float[] tempBuffer;

int numBands = 512;
int rate = 60;
int currentChannel = 0;
int sampleLength = 5;
int sampleRate = 44100;
int numFrames = sampleRate*sampleLength;

void setup(){

 size(200,200);
 framerate(rate);
 background(0);
 noLoop();
 //START SONIA AND LIVE INPUT, CREATE EMPTY SAMPLE OBJECT//
 Sonia.start(this);
 LiveInput.start(numBands);
 mySample = new Sample(numFrames);
 //START ESS, CREATE A TEMP BUFFER AND AN EMPTY CHANNEL OBJECT//
 Ess.start(this);
 tempBuffer = new float[numFrames];
 myChannel = new Channel(numFrames);
}

void draw(){

  LiveInput.startRec(mySample); //START RECORDING LIVE INPUT//
  delay(1000*sampleLength);     //WAIT TILL SAMPLE FRAMES ARE FULL//
                                //THERE MUST BE A BETTER WAY!//
  mySample.read(tempBuffer);    //COPIES SAMPLE DATA TO FLOAT ARRAY//
  myChannel.loadSamples();      //ACCESS ESS CHANNEL DATA ARRAY//
 
  //THE FOLLOWING WRITES THE FLOAT ARRAY TO THE ESS DATA ARRAY//
  //OF TYPE INT. THE FLOAT ARRAY HAS VALUES BETWEEN 1 AND -1,//
  //SO SCALE DATA AND CAST TO INT. MUST BE A BETTER WAY TO DO THIS//
 
  for(int i=0; i<tempBuffer.length; i++){
      myChannel.setFrame(myChannel.samples,i,int(tempBuffer[i]*pow(10,5)));
  }
 
  myChannel.updateSamples(); //PUSH DATA INTO THE CHANNEL OBJECT//
  myChannel.play(); //PLAY THE CHANNEL//
}

//CLEAN UP SONIA AND ESS//

public void stop(){
   Sonia.stop();
   Ess.stopAll();
   Ess.stop();
   super.stop();
}

I was wondering the same thing. Started experimenting with Sample.getCurrentFrame() in a conditional loop to check whether or not the buffer was full and then start playing instead of using delay(), but I was having problems. Think it might be the behavior of LiveInput.startRec() when it appears in a loop. For instance,

void draw{
 LiveInput.startRec(mySample);
 mySample.play();
}

This works fine for playback, but my guess is that the sample buffer never gets filled completely, as the current frame pointer is set to zero when startRec is called. At least I think this is the case. Can someone confirm this?

In the meantime, a little test:

void setup(){
 size(100,100);
 background(0);
 framerate(1);
 int numFrames = 44100;
 Sonia.start(this);
 LiveInput.start(numBands);
 mySample = new Sample(numFrames);
}

void draw(){
 LiveInput.startRec(mySample);
 println(mySample.getCurrentFrame());
 mySample.play();
}

When it's run at framerate(1), the output is a series of numbers very close to the number of samples @ 44100. Something like 44033, 42369, etc.  A test at framerate(100) results in much lower numbers, around 500 or less. If I remove the mySample.play() statement from the loop, only zeros are returned. This I did not expect.

Pitaru, perhaps you can shed some light on what's going on behind the scenes of LiveInput.startRec()?

Thanks,

~pN~