Playing Dialog Files

[CornbreadCobra]

Say some of us that are running OSX want to listen to the individual voice .wav files of the game, specifically HK-47's dialog. I have access to the StreamVoice folder, just whenever I attempt to play a file, nothing happens or the sound is only static. Changing the .wav to .mp3 does not work either. Some answers that I have read on other forums involved using Miles Sound Studio to play the files as normal audio players would not work.

[Ninko]

You have to un compress the files because they are compressed .wavs

[Mansen]

Originally posted by NinkoKotOR:

You have to un compress the files because they are compressed .wavs

Talk about an oxymoron, heh :D

[CornbreadCobra]

Originally posted by NinkoKotOR:

You have to un compress the files because they are compressed .wavs

And how might I go around doing that? I'm not seeing any options to uncompress the wav file.

[DrMcCoy]

The problem is this: those file aren't actually wave files. They claim they are raw, uncompressed PCM wave files (only the size of the data chunk says it's 0 byte long), but they're actually MP3s. The extra "RIFF" "WAVE" bytes at the start, the fake WAVE header, throws normal audio players off. They then interpret the data as raw, uncompressed PCM data, which this isn't, and so you get static. It's very possible that this was originally done to just prevent people from playing those files.

You need to cut off the fake header, with a hexeditor. It's usually the first 58 bytes (until 4 bytes after the "data" text). The actual MP3 then starts with byte value 0xFF (which is part of the MPEG header bitstream).

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

[CornbreadCobra]

Originally posted by DrMcCoy:

The problem is this: those file aren't actually wave files. They claim they are raw, uncompressed PCM wave files (only the size of the data chunk says it's 0 byte long), but they're actually MP3s. The extra "RIFF" "WAVE" bytes at the start, the fake WAVE header, throws normal audio players off. They then interpret the data as raw, uncompressed PCM data, which this isn't, and so you get static. It's very possible that this was originally done to just prevent people from playing those files.

You need to cut off the fake header, with a hexeditor. It's usually the first 58 bytes (until 4 bytes after the "data" text). The actual MP3 then starts with byte value 0xFF (which is part of the MPEG header bitstream).

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

I followed what you described and it works, thanks.

[just-a-gimmick]

Originally posted by DrMcCoy:

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

You, Sir, are my hero.

Works like a charm!

edit: only setting the first bit to 0x00 already seems to work for VLC

[Gxsoldier77]

Originally posted by DrMcCoy:

The problem is this: those file aren't actually wave files. They claim they are raw, uncompressed PCM wave files (only the size of the data chunk says it's 0 byte long), but they're actually MP3s. The extra "RIFF" "WAVE" bytes at the start, the fake WAVE header, throws normal audio players off. They then interpret the data as raw, uncompressed PCM data, which this isn't, and so you get static. It's very possible that this was originally done to just prevent people from playing those files.

You need to cut off the fake header, with a hexeditor. It's usually the first 58 bytes (until 4 bytes after the "data" text). The actual MP3 then starts with byte value 0xFF (which is part of the MPEG header bitstream).

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

MAN THANK YOU SO MUCH IT WORKED!

[Gxsoldier77]

Originally posted by Gimmick5000:

Originally posted by DrMcCoy:

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

You, Sir, are my hero.

Works like a charm!

edit: only setting the first bit to 0x00 already seems to work for VLC

Yes it works with groove player, doing only this

[Gxsoldier77]

So basically you need to open your wav file in a Hexeditor, select the first two numbers or bytes, and change to 0

[positron01]

I made a python script to remove the beginning bytes up to 0xFF. I only tested this for the StreamMusic folder, but it seems to work.

import os
directory = 'StreamMusic'
file_names = os.listdir(directory)
for file_name in file_names:
    relative_path = directory + '/' + file_name
    with open(relative_path, mode='rb') as file:
        file_bytes = file.read()
        i = 0
        for file_byte in file_bytes:
            i+=1
            if file_byte == 0xFF:
                scrubbed_bytes = file_bytes[i:len(file_bytes)]
                with open(directory + '/' + 'edited_'+file_name, mode='wb') as edited_file:
                    edited_file.write(scrubbed_bytes)
                    break

[positron01]

I made an update that should work with both kotor1 and kotor2. I have really only tested the music and it seems to work. I tested only one dialog file from kotor2 and it worked as well.

import os

def load_files(directory):
    try:
        os.makedirs(directory + '/edited_files')
    except FileExistsError:
        pass
    file_names = os.listdir(directory)
    for file_name in file_names:
        relative_path = directory + '/' + file_name
        if os.path.isfile(relative_path):
            with open(relative_path, mode='rb') as file:
                file_bytes = None
                file_check = file.read(4)
                if file_check == b'RIFF':
                    file.seek(0x3A)
                    file_bytes = file.read()
                else:
                    file.seek(0x1D6)
                    file_check = file.read(4)
                    if file_check == b'RIFF':
                        file.seek(0x1D6)
                        file_bytes = file.read()
                with open(directory + '/edited_files/' + file_name, mode='wb') as edited_file:
                    try:
                        edited_file.write(file_bytes)
                    except TypeError:
                        print('Failed to write ' + file_name)

if __name__ == '__main__':
    kotor1_directory = 'kotor1/streammusic'
    kotor2_directory = 'kotor2/streammusic'
    load_files(kotor1_directory)
    load_files(kotor2_directory)

[positron01]

Originally posted by DrMcCoy:

The problem is this: those file aren't actually wave files. They claim they are raw, uncompressed PCM wave files (only the size of the data chunk says it's 0 byte long), but they're actually MP3s. The extra "RIFF" "WAVE" bytes at the start, the fake WAVE header, throws normal audio players off. They then interpret the data as raw, uncompressed PCM data, which this isn't, and so you get static. It's very possible that this was originally done to just prevent people from playing those files.

You need to cut off the fake header, with a hexeditor. It's usually the first 58 bytes (until 4 bytes after the "data" text). The actual MP3 then starts with byte value 0xFF (which is part of the MPEG header bitstream).

Alternatively, you can overwrite those bytes with 0x00 (i.e. byte value 0, not the text "0x00"), also in a hexeditor. That's a bit dirty, but most players that can play MP3 files should then correctly identify the file, especially when you also rename it to something.mp3 (instead of .wav). VLC works, for example.

EDIT: It's even already enough to only overwrite the first byte with 0x00 for VLC to play the file. Other players might not.

Also thanks to you for getting me on the right track. I also found the following post which was very helpful as well.
https://www.gog.com/forum/star_wars_knights_of_the_old_republic_series/music_from_star_wars_knights_of_the_old_republic/page1