scream is a Scream audio receiver using Pulseaudio, ALSA or stdout as audio output.
Compilation is done using CMake.
If Pulseaudio or ALSA headers are found, support for those will be compiled in (see below for distro packages to install). If Pulseaudio is compiled in, it will be the default output, otherwise ALSA will be the default. If both Pulseaudio and ALSA are not compiled in, raw/stdout will be the default.
$ mkdir build && cd build
$ cmake ..
$ make$ sudo yum install pulseaudio-libs-devel # Redhat, CentOS, etc.
or
$ sudo apt-get install libpulse-dev # Debian, Ubuntu, etc.$ sudo yum install alsa-lib-devel # Redhat, CentOS, etc.
or
$ sudo apt-get install libasound2-dev # Debian, Ubuntu, etc.You can see the accepted options by using the -h (help) option.
$ scream -h$ screamThis starts the Scream client in multicast mode, using the default audio output. Unicast mode is also supported, and can be used by passing the -u option. If your machine has more than one network interface, you may need to set the interface name which receives scream packets.
$ scream -i eth0This starts the Scream client in libpcap (sniffer) mode. This mode is mostly useful if you are able to the UDP
multicast/unicast transmission in wireshark/tcpdump but unable to have it be delivered to the
user-space Scream client.
$ scream -P -i macvtap0If you have the hard requirement of having to run the receiver as non-root due to pulseaudio/alsa
uid/gid issues, you can do the following:
# setcap cap_net_raw,cap_net_admin=eip ./screamMake sure to have read permission for the shared memory device and execute
$ scream -m /dev/shm/scream-ivshmemIf you experience excessive underruns under normal operating conditions,
lower the process niceness; if it still underruns, raise the default
target latency (50 ms) with -t:
$ scream -o alsa -t 100Note that audio hardware typically has small buffers that result in a latency lower than the target latency.
Run with -v to dump ALSA PCM setup information.
Run with env LIBASOUND_DEBUG=1 to debug ALSA problems.