Java spectrograms: mp3 and microphone

Question

First of all, I am working on a small project to see a spectrum of some sounds.

I worked with a microphone: alt text http://img25.imageshack.us/img25/4271/spectrumanalyzerfourier.png

The image above is just what I say and scream through the microphone for a few seconds. It looks good to me.

But when I try to read an MP3 file and make it, it looks a little different. I tried Aphex Twin - Windowlicker, where you usually should see a face in a spectral image or at least darker colors. But it doesn’t look so good: alt text http://img10.imageshack.us/img10/3475/aphextwinhmm.png

Here is what I did with the microphone:

byte tempBuffer[] = new byte[10000]; ByteArrayOutputStream out = new ByteArrayOutputStream(); counter = 20; // Microphone while (counter != 0) { int count = line.read(tempBuffer, 0, tempBuffer.length); if (count > 0) { out.write(tempBuffer, 0, count); } counter--; } out.close(); // FFT code below ... byte audio[] = out.toByteArray(); // ... 

And here is how I do it with MP3:

I used the same code for the conversion and rendering only for the audio capture part (I only adjusted the height in the drawing method to see if there was a difference, but it wasn’t):

 byte tempBuffer[] = new byte[10000]; ByteArrayOutputStream out = new ByteArrayOutputStream(); FileInputStream input = null; File mp3 = new File("Aphex Twin - Widowlicker.mp3"); input = new FileInputStream(mp3); int len; while((len = input.read(tempBuffer)) > 0) { out.write(tempBuffer, 0, len); } out.close(); input.close(); // FFT code below ... byte audio[] = out.toByteArray(); // ... 

It would be nice if someone could tell me what I'm doing wrong with the MP3 file.

These are my settings:

• Poll Frequency: 44100
• Bit for sample: 8
• Channels: 1 (mono)
• signed: true
• big endian: true (I use AudioFormat in Java)
• tempBuffer for reading audio: 10000 (bytes tempBuffer [] = new byte [10000];)
• and for FFT I share the audio in bytes 4096 (should be 2)

By the way: are these settings OK, or should I use 16 bit / s or stereo or 10000 for a buffer too large or 4096 for small / large?

Thanks in advance

Answer 1

MP3 is a compressed audio format. You must unzip the data first before you can use it as an audio stream comparable to your microphone data. The raw MP3 data has maximum entropy and should look just like the white noise it makes in your spectrogram.