Android:- convert a recorded audio file into a float array

Question

I am making a simple android application where I record something from a smartphone mic, save it in a file and then play that file.

Now I want to apply a high pass filter to that audio file, but to do so, I need to first convert the audio file into a float array. Can someone please help me out with it.

Thanks

package abc.com.please;

import android.content.pm.PackageManager;
import android.media.MediaPlayer;
import android.media.MediaRecorder;
import android.os.Bundle;
import android.os.Environment;
import android.support.v7.app.AppCompatActivity;
import android.view.View;
import android.widget.Button;
import android.widget.Toast;


public class MainActivity extends AppCompatActivity {

private static MediaRecorder mediaRecorder = new MediaRecorder();
private static MediaPlayer mediaPlayer;

private static String audioFilePath;
private static Button stopButton;
private static Button playButton;
private static Button recordButton;

private boolean isRecording = false;


@Override
protected void onCreate(Bundle savedInstanceState) {
    super.onCreate(savedInstanceState);
    setContentView(R.layout.activity_main);
    recordButton = (Button) findViewById(R.id.recordButton);
    playButton = (Button) findViewById(R.id.playButton);
    stopButton = (Button) findViewById(R.id.stopButton);

    if (!hasMicrophone())
    {
        stopButton.setEnabled(false);
        playButton.setEnabled(false);
        recordButton.setEnabled(false);
    } else {
        playButton.setEnabled(false);
        stopButton.setEnabled(false);
    }

    audioFilePath =
            Environment.getExternalStorageDirectory().getAbsolutePath()
                    + "/myaudio.3gp";


    recordButton.setOnClickListener(new View.OnClickListener(){

        public void onClick (View v)
        {

            isRecording = true;
            stopButton.setEnabled(true);
            playButton.setEnabled(false);
            recordButton.setEnabled(false);

            try {
                mediaRecorder.setAudioSource(MediaRecorder.AudioSource.MIC);
                                   mediaRecorder.setOutputFormat(MediaRecorder.OutputFormat.THREE_GPP);
                mediaRecorder.setOutputFile(audioFilePath);
                  mediaRecorder.setAudioEncoder(MediaRecorder.AudioEncoder.AMR_NB);
                mediaRecorder.prepare();
                mediaRecorder.start();
            }catch (Exception e) {
                e.printStackTrace();

            }



        }
    });

    stopButton.setOnClickListener(new View.OnClickListener() {
        public void onClick (View view)
        {

            stopButton.setEnabled(false);
            playButton.setEnabled(true);
            if (isRecording)
            {
                recordButton.setEnabled(false);
                isRecording = false;
                mediaRecorder.stop();
                mediaRecorder.release();
                recordButton.setEnabled(true);
            }
            else
            {
                Toast.makeText(getApplicationContext(),"No recording going      on",Toast.LENGTH_SHORT).show();
            }
        }});

    playButton.setOnClickListener(new View.OnClickListener() {
        public void onClick(View view)

        {

            playButton.setEnabled(false);
            recordButton.setEnabled(false);
            stopButton.setEnabled(true);
            try {
                mediaPlayer = new MediaPlayer();
                mediaPlayer.setDataSource(audioFilePath);
                mediaPlayer.prepare();
                mediaPlayer.start();
            } catch (Exception e) {
                e.printStackTrace();
            }

        }
    });
}

protected boolean hasMicrophone() {
    PackageManager pmanager = this.getPackageManager();
    return pmanager.hasSystemFeature(
            PackageManager.FEATURE_MICROPHONE);
}








}

Answer 1

The details of the bytes-to-float conversion are going to depend on the file format that you use for your audio file. I cannot tell what that format is from your code. If your file is WAV, 44100 fps, 16-bit, little-endian, stereo (this is a standard Java, "CD quality" format) you can try making use of the following code I wrote. The key conversion point is when two bytes are concatenated and converted to a single numeral (where "buffer" contains data being read in from the audio file:

float audioVal = ( buffer[bufferIdx++] & 0xff )
                        | ( buffer[bufferIdx++] << 8 );

If it is big endian, reverse the order of the shifts. If 24 or 32 bit, then you would OR in shifts of << 16 and << 24, respectively. With 16-bit, the result will make use of the range of a short, so division by 32767 is needed to normalize the result to [-1..1].

I've been using the following in the Java context for a while without problems, but I don't know if Android supports javax.sound.sampled.AudioInputStream, etc. Maybe it is still useful to see the conversion in context of a file read? The code assumes that we have the "CD Quality" audio format and that the audio file is not longer than Integer.MAX number of frames.

public float[] loadSoundFileURL(URL url) throws UnsupportedAudioFileException,
        IOException
{
    AudioInputStream ais = AudioSystem.getAudioInputStream(url);

    int framesCount = (int)ais.getFrameLength();

    // assuming stereo format, so two entries per frame
    float[] temp = new float[framesCount * 2];
    long tempCountdown = temp.length;

    int bytesRead = 0;
    int bufferIdx;
    int clipIdx = 0;
    byte[] buffer = new byte[1024];
    while((bytesRead = ais.read(buffer, 0, 1024)) != -1)
    {
        bufferIdx = 0;
        for (int i = 0, n = (bytesRead >> 1); i < n; i ++)
        {
            if ( tempCountdown-- >= 0)
            {
                temp[clipIdx++] = 
                        ( buffer[bufferIdx++] & 0xff )
                        | ( buffer[bufferIdx++] << 8 ) ;
            }
        }
    }

    // QUESTION: better to do following in above loop?
    for (int i = 0; i < temp.length; i++)
    {
        temp[i] = temp[i] / 32767f;
    }

    return temp;
}