CLI to export players' voices from CSGO/CS2 demos into WAV files.
Warning
Valve Matchmaking demos do not contain voice audio data, hence there is nothing to extract from MM demos.
Download the latest release for your OS from GitHub.
The program can export voices in 3 different modes:
- Split compact: extracts and concatenates all of each player's voice segments into separate WAV files. Each player will have their own WAV file containing only their voice data (without silence), and the files will be named after the player's Steam ID. This is the default mode.
- Split full: extracts players' voices into separate WAV files that have the same duration as the demo file. Each player will have their own WAV file with voice segments placed at their original timestamps, and the files will be named after the player's Steam ID.
- Single full: extracts and merges all players' voices into a single WAV file that has the same duration as the demo file, preserving the original timing of all voice communications.
To change the mode, you have to set the -mode
argument. The possible values are:
split-compact
(default)split-full
single-full
csgove.exe demoPaths... [-output]
By default .dll
files are expected to be in the same directory as the executable.
You can change it by setting the LD_LIBRARY_PATH
environment variable. Example:
LD_LIBRARY_PATH="C:\Users\username\Desktop" csgove.exe
Caution
The environment variable DYLD_LIBRARY_PATH
must be set before invoking the program and point to the location of the .dylib
files!
DYLD_LIBRARY_PATH=. csgove demoPaths... [-output]
Caution
The environment variable LD_LIBRARY_PATH
must be set before invoking the program and point to the location of the .so
files!
LD_LIBRARY_PATH=. csgove demoPaths... [-output]
-output <string>
Folder location where audio files will be written. Current working directory by default.
-mode <string>
Output mode that determines how the voices are extracted and saved:
split-compact
(default): separate files per player, without silencesplit-full
: separate files per player, with demo-length silencesingle-full
: single merged file with all players' voices
-steam-ids <string>
Comma-separated list of Steam IDs 64 to extract voices for. If not provided, voices for all players will be extracted.
-exit-on-first-error
Stop the program at the first error encountered. By default, the program will continue to the next demo to process if an error occurs.
Extract voices from the demo myDemo.dem
in the current directory:
csgove myDemo.dem
Extract voices from multiple demos using absolute or relative paths:
csgove myDemo1.dem ../myDemo2.dem "C:\Users\username\Desktop\myDemo3.dem"
Change the output location:
csgove -output "C:\Users\username\Desktop\output" myDemo.dem
Extract all voices into a single merged file:
csgove -mode single-full myDemo.dem
Extract only voices of specific players:
csgove -steam-ids 76561198123456789,76561198123456780 myDemo.dem
- Go
- GCC
- Chocolatey (Windows only)
Debugging is easier on macOS/Linux 64-bit, see warnings below.
Because the CSGO audio library is a 32-bit DLL, you need a 32-bit GCC
and set the Go env variable GOARCH=386
to build the program.
Important
Use a unix like shell such as Git Bash, it will not work with cmd.exe
!
Warning
The $GCC_PATH
variable in the following steps is the path where gcc.exe
is located.
By default, it's C:\TDM-GCC-64\bin
when using TDM-GCC (highly recommended).
- Install
GCC
for Windows, TDM-GCC is recommended because it handles both 32-bit and 64-bit when runninggo build
. If you use MSYS2, it's important to install the 32-bit version (pacman -S mingw-w64-i686-gcc
). - Install
pkg-config
using chocolatey by runningchoco install pkgconfiglite
.
It's highly recommended to usechoco
otherwise you would have to buildpkg-config
and copy/paste thepkg-config.exe
binary in your$GCC_PATH
. - Download the source code of Opus
- Extract the archive, rename the folder to
opus
and place it in the project's root folder mkdir build && cd build
cmake -G "Visual Studio 17 2022" -A Win32 -DCMAKE_BUILD_TYPE:STRING=Release -DBUILD_SHARED_LIBS=ON ..
- Build the
Release
configuration forWin32
(notx64
- it's important to build the 32-bit version!) - Copy/paste the
opus.dll
file in$GCC_PATH
anddist/bin/win32-x64
- Copy/paste the C header files located inside the
include
folder file in$GCC_PATH\include\opus
(create the folders if needed) - Copy/paste the
opus.pc.example
toopus.pc
file and edit theprefix
variable to match yourGCC
installation path if necessary. PKG_CONFIG_PATH=$(realpath .) LD_LIBRARY_PATH=dist/bin/win32-x64 CGO_ENABLED=1 GOARCH=386 go run -tags nolibopusfile .
Warning
Because the Go debugger doesn't support Windows 32-bit and the CSGO lib is a 32-bit DLL, you will not be able to run the Go debugger.
If you want to be able to run the debugger for the Go part only, you could comment on lines that involve C/CGO
calls.
Important
On macOS ARM64
, the x64
version of Homebrew must be installed!
You can install it by adding arch -x86_64
before the official command to install Homebrew (arch -x86_64 /bin/bash -c "$(curl -fsSL https://github.com/raw/Homebrew/install/HEAD/install.sh)"
)
- Install Homebrew x64 version
arch -x86_64 brew install opus
arch -x86_64 brew install pkg-config
cp /usr/local/Cellar/opus/1.4/lib/libopus.0.dylib dist/bin/darwin-x64
(arch -x86_64 brew info opus
to get the path)DYLD_LIBRARY_PATH=dist/bin/darwin-x64 CGO_ENABLED=1 GOARCH=amd64 go run -tags nolibopusfile .
Warning
On macOS ARM64, the Go debugger breakpoints will not work because the executable must target amd64 but your OS is ARM64.
sudo apt install pkg-config libopus-dev
cp /usr/lib/x86_64-linux-gnu/libopus.so.0 dist/bin/linux-x64
(you may need to change the path depending on your distro)LD_LIBRARY_PATH=dist/bin/linux-x64 CGO_ENABLED=1 GOARCH=amd64 go run -tags nolibopusfile .
make build-windows
make build-darwin
make build-linux
Thanks to @saul and @ericek111 for their CSGO investigation.
Thanks to @DandrewsDev for his work on CS2 voice data extraction.