Win32::Sound 16 bit example

[true]

Aldo wrote a great module Win32::Sound. I have been struggling with making an example using 16bit sound though and i would dig some help from any monk willing. Here's the example included with Win32::Sound distro. It generates a 8 bit sound in stereo.

use Win32::Sound;
    # Create the object
    $WAV = new Win32::Sound::WaveOut(44100, 8, 2);
    $data = ""; 
    $counter = 0;
    $increment = 440/44100;
    # Generate 44100 samples ( = 1 second)
    for $i (1..44100) {
        # Calculate the pitch 
        # (range 0..255 for 8 bits)
        $v = sin($counter/2*3.14) * 128 + 128;
        # "pack" it twice for left and right
        $data .= pack("cc", $v, $v);
        $counter += $increment;
    }
    $WAV->Load($data);       # get it
    $WAV->Write();           # hear it
    1 until $WAV->Status();  # wait for completion
    $WAV->Save("sinus.wav"); # write to disk
    $WAV->Unload();          # drop it

Above example works great. My questoin is how to do this with 16bit sound. 16 bit sound has a range of 65335 values. It is stored different from 8bit as a signed integer value -32768 <-> 32767. Zero (no sound is 0). My best guess is to use pack i but i know it is not correct. The below example tries to make a 16 bit mono wav file the same way Aldo's 8bit example works. One 16bit sound is stored in 2 bytes while 8bit sound is stored in one. The pack i line is the line that needs correcting. WAV files are stored in little endian order.

use Win32::Sound;
    # Create the object
    $WAV = new Win32::Sound::WaveOut(44100, 16, 1);
    $data = ""; 
    $counter = 0;
    $increment = 440/44100;
    # Generate 44100 samples ( = 1 second)
    for $i (1..44100) {
        # Calculate the pitch 
        # (range 0..65335 for 16 bits)
        $v = int(sin($counter/2*3.14) * (65335/2));
        #signed integer (v) range -32768 <-> 32767
        $data .= pack("i", $v);
        $counter += $increment;
    }
    $WAV->Load($data);       # get it
    $WAV->Write();           # hear it
    1 until $WAV->Status();  # wait for completion
    $WAV->Save("sinus.wav"); # write to disk
    $WAV->Unload();          # drop it

Thanks for reading.

[BrowserUk]

If your only problem is how to pack a 16-bit unsigned value in le order?, then (from the pack entry in perlfunc):

   v   An unsigned short in "VAX" (little-endian) order.

[true]

By george i think that's the answer. Sound generated with your fix is the exact same tone as the 8bit one!!! Thanks so much for your help BrowserUK! I tried to do this over a year ago and was stuck then too. I have read the pack docs so many times its crazy but their brevity times my ignorance about value types still confuses me. Here is a working example with your fix for a 16bit mono sound wave...

use Win32::Sound;
    # Create the object
    $WAV = new Win32::Sound::WaveOut(44100, 16, 1);
    $data = ""; 
    $counter = 0;
    $increment = 440/44100;
    # Generate 44100 samples ( = 1 second)
    for $i (1..44100) {
        # Calculate the pitch 
        # (range 0..65335 for 16 bits)
        $v = int(sin($counter/2*3.14) * (65335/2));
        #signed integer (v) range -32768 <-> 32767
        $data .= pack("v", $v);
        $counter += $increment;
    }
    $WAV->Load($data);       # get it
    $WAV->Write();           # hear it
    1 until $WAV->Status();  # wait for completion
    $WAV->Save("sinus.wav"); # write to disk
    $WAV->Unload();          # drop it

thanks again, hopefully someone else can use this too. BTW if anyone wants to do this in stereo here's the code for 16bit stereo sound too. Just as above, this generates a 1 second sinusoidal wave at 440Hz but in both channels.

use Win32::Sound;
    # Create the object
    $WAV = new Win32::Sound::WaveOut(44100, 16, 2);
    $data = ""; 
    $counter = 0;
    $increment = 440/44100;
    # Generate 44100 samples ( = 1 second)
    for $i (1..44100) {
        # Calculate the pitch 
        # (range 0..65335 for 16 bits)
        $v = int(sin($counter/2*3.14) * (65335/2));
        #signed integer (v) range -32768 <-> 32767
        $data .= pack("v", $v);
        $data .= pack("v", $v);
        $counter += $increment;
    }
    $WAV->Load($data);       # get it
    $WAV->Write();           # hear it
    1 until $WAV->Status();  # wait for completion
    $WAV->Save("sinus.wav"); # write to disk
    $WAV->Unload();          # drop it

exit;

Thanks again all!

[true]

Yikes, When i try and unpack the 16bit sound value right after i pack it, i get a value which is incorrect if the initial value is less than zero. Basically, the unpack("v",...) is not working correctly. Any advice would be appreciated. I want to unpack the value so i can use it when i read in sound. I use this code only to illustrate. This program outputs two different numbers if the initial value of v is below zero. I need it to output the same numbers so i know i am reading the sound file correctly.

use Win32::Sound;
# Create the object
my $WAV = new Win32::Sound::WaveOut(44100, 16, 1);
my $data = ""; 
my $counter = 0;
my $increment = 440/44100;
# Generate 44100 samples ( = 1 second)
for my $i (1..44100) {
    # Calculate the pitch 
    # (range 0..65335 for 16 bits)
  $v = int(sin($counter/2*3.14) * (65335/2));
  #signed integer (v) range -32768 <-> 32767
print "$v = ".unpack("v",pack("v", $v))."\n";
    $data .= pack("v", $v);
  $counter += $increment;
}
$WAV->Load($data);       # get it
$WAV->Write();           # hear it
1 until $WAV->Status();  # wait for completion
$WAV->Save("16.wav"); # write to disk
$WAV->Unload();          # drop it

exit;

thanks to all for help and for reading.

[BrowserUk]

The 'v' is unsigned 16-bit little-endian. It works fine for packing signed 16-bit values to little-endian, but not for unpacking them.

To recover the signed value use:

print unpack 's', pack 'v', -32767;;
-32767