https://github.com/quark-zju/sndcat
Command line tool for mixing or converting audio streams
https://github.com/quark-zju/sndcat
Last synced: 21 days ago
JSON representation
Command line tool for mixing or converting audio streams
- Host: GitHub
- URL: https://github.com/quark-zju/sndcat
- Owner: quark-zju
- License: mit
- Created: 2021-04-11T20:45:04.000Z (about 4 years ago)
- Default Branch: master
- Last Pushed: 2025-05-03T08:14:21.000Z (24 days ago)
- Last Synced: 2025-05-03T09:26:38.448Z (24 days ago)
- Language: Rust
- Size: 195 KB
- Stars: 2
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# sndcat
[](https://github.com/quark-zju/sndcat/actions/workflows/build.yml)
A command line utility that works on audio streams.
Like [`socat`](https://linux.die.net/man/1/socat), `sndcat` has a concept of
inputs and outputs (or, "source" and "sink"). `sndcat` reads from input,
does some necessary processing (ex. resampling), then writes to outputs.
## Usage
### Examples
Encode a large Podcast file from MP3 to [Opus](https://opus-codec.org/) to
make it smaller. `-o stats()` shows progress. It is optional:
sndcat -i mp3(podcast.mp3) -o opus(podcast.opus,16000,1,voip) -o stats()
Record a video conference, including what you said and others said:
# Suppose 5 and 7 are device indexes frm `sndcat list`. One is the
# microphone recording what you said, and the other is the loopback
# device recording what others said.
sndcat -i dev(5) -i dev(7) -o opus(conf.opus)
Keep an output device busy by writing silent samples. This prevents the
device from entering a "paused" state that will make loopback recording
hanging.
sndcat -i silence() -o dev(6)
### Input
Use `-i` to specify input streams.
An input stream could be from an input device, a mp3 file, a sine wav, or
multiple streams mixed together. It uses a mini DSL. For example:
Reading from the 5-th device. Use `sndcat list` to list available devices:
-i dev(5)
Mixing two devices. This can be useful for video conferences, and you want
to record both what you said and what others said:
-i mix(dev(5),dev(7))
Decoding an MP3 file. This can be used to play it, or re-encode into other
formats:
-i mp3('foo.mp3')
Generating a sine wave at the given frequency. This can be used for testing:
-i sin(440)
Specifying `-i` multiple times means reading from multiple streams
simultaneously. It is just a syntax sugar of `mix`. For example, the
following are equivalent:
-i dev(5) -i dev(6) -i sin(440)
-i mix(dev(5),dev(6),sin(440))
Input streams can be endless. Press Ctrl+C to force end it.
### Output
Use `-o` to specify outputs.
An output could be a device, an [Opus](https://opus-codec.org/) file, or
showing statistics in the terminal.
Writing to the 6-th device. Use `sndcat list` to list available devices:
-o dev(6)
Encoding into an Opus file:
-o opus(a.opus,16000,1,voip) # for voice
-o opus(a.opus,24000,2,audio) # for music
Showing simple statistics in terminal. This can be useful to see the
progress of file encoding:
-o stats()
Specifying `-o` multiple times means writing to them simultaneously.
## Build
### Windows
Install [vcpkg](https://github.com/Microsoft/vcpkg) first. Then build
portaudio in the local `vcpkg` directory:
set VCPKG_ROOT=C:\vcpkg
vcpkg\build-portaudio.bat
This will build portaudio patched by
[Audicaty](https://github.com/audacity/audacity) which enables loopback
recording.
Then build `sndcat` using Rust toolchain `cargo`:
cargo install --path .
Currently, Windows is the main platform that `sndcat` is tested.
### Linux
Install dependencies first. For example, For example, Debian/Ubuntu might use
the following command:
sudo apt install build-essential autoconf portaudio19-dev libopus-dev
Then build `sndcat` using Rust toolchain `cargo`:
cargo install --path .