I've never done this myself before but as far as I know the situation is as follows.
If you're working with Flash Player 10 your in luck, otherwise it would seem that you have to compute the spectrum as the sound is playing. Like in Chris's link you can get the sound data out of the sound object into a ByteArray with Sound.extract() (FP10 only). You then go through the ByteArray read the values and draw something on screen according to the values.
On FP9 you'll have to play the sound on a SoundChannel and read the left and right peaks from the channel as it plays. I found this example code for that (link)
var snd:Sound = new Sound();
var req:URLRequest = new URLRequest("your track url");
var channel:SoundChannel;
var bytes:ByteArray = new ByteArray();
snd.load(req);
this.addEventListener(Event.ENTER_FRAME, enterFrameEvent);
channel = snd.play(0,3);
function enterFrameEvent(event:Event):void {
SoundMixer.computeSpectrum(bytes, true,0);
// bytes has 512 values 0-255 leftchannel 256-512 right channel
for (var i:Number = 0; i < 256; i++) {
val = bytes.readFloat();
}
}
So in short try to stick to FP10 and if you're able to do that (no client restrictions) you should be able to take the Waveform class straight out of Chris's link and use that. It looks like the class also draws the actual waveform (on top of extracting the audio data) but you can either move the drawing to it's own class or modify the code to get your own kind of look for the waveform.
