LANSink is simple unreliable half-duplex audio transport over UDP implemented as an ALSA plugin on sender side and as a stand-alone daemon on receiver side. It currently performs no audio compression or latency control. It is designed as a means to connect HTPC with other computers on the same network.

• protobuf
• alsa

• g++
• cmake >= 2.8.5
• pkg-config
• alsa-lib
• git (for getting the source code)

Under Fedora you can install required packages using:

$ sudo yum install cmake gcc-c++ alsa-lib-devel protobuf-devel git

Clone git repository:

$ git clone https://github.com/NikitaKarnauhov/lansink.git

Build:

$ cd lansink
$ mkdir Release
$ cd Release
$ cmake CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr ..
$ make
$ sudo make install

This installs both ALSA plugin and receiver daemon. You will likely need to repeat intallation both on sender and on receiver side.

Audio sender is implemented as an ALSA plugin lansink, see asoundrc.sample in the source tree for annotated sample configuration (copy it's contents into your $HOME/.asoundrc). Only type parameter is strictly required but you probably should also specify host to point to the receiver machine.

You can test configuration with speaker-test program from alsa-utils package:

$ speaker-test -c 2 -b 20000 -D lansink

It shouldn't produce any errors though no sound will actually appear until receiver daemon is set up.

If you followed this instructions receiver daemon can be launched with 'lansinkd' command. To test the installation run:

$ src/lansinkd -L3 -n

This command launches receiver attached to the terminal with log level set to info. The output should be approximately like the following:

$ lansinkd -L3 -n
2013-08-27T21:20:58+0700: INFO: Logging to /tmp/lansinkd.log
2013-08-27T21:20:58+0700: INFO: PID file written to /tmp/lansinkd.pid
2013-08-27T21:20:58+0700: INFO: Server listening on 0.0.0.0:26751

On incoming connections ALSA device default will be opened. To override this use -D command line argument, e.g. run lansinkd -D hw:0,0 to use hardware output. When using plain sound card (without PulseAudio/dmix layer) that doesn't support mixing use -x parameter to prevent 'Device or resource busy' errors on multiple simultaneous connections. See lansinkd.conf.sample file for annotated configuration. You can specify custom config with -c argument.

You may repeat running speaker-test on the sender machine. You will hear noise if everything is set up properly.

For ease of switching outputs you can use LANSink from PulseAudio with alsa-sink module. Add following line to the PulseAudio configuration ($HOME/.config/pulse/default.pa or /etc/pulse/default.pa):

load-module module-alsa-sink device=lansink sink_name=lansink tsched=0 fragment_size=6144 fragments=1

fragment_size and fragments parameters are not required but can help when playing audo from network, e.g. with Flash Player from your browser.

PulseAudio can also be used on receiver in order to mix streams and convert audio formats to ones supported by hardware. You will need to install 'pulse' plugin for ALSA. On Fedora run:

$ sudo yum install alsa-plugins-pulseaudio

It will add pulse alsa device that can then be passed to lansinkd with -D argument.