the xiph open source community

Icecast is free server software for streaming multimedia.

IceS 2.0.2 Docs — Input Modules

Several input modules are available, depending on the platform, drivers and libraries available. The general layout is defined as

    <module>module name</module>
    <param name="name1">value</param>
    <param name="name2">value</param>
    <param name="name3">value</param>

For live input you may want to look into various resources on the web for information on sound input. You may find that ALSA for instance supports a particular soundcard better than the Open Sound System.

Open Sound

<param name="rate">44100</param>
<param name="channels">2</param>
<param name="device">/dev/dsp</param>
<param name="metadata">1</param>
<param name="metadatafilename">/home/ices/metadata</param>

This module is for reading live input from the Open Sound System drivers, often found on linux systems but are available on others. This will read audio from the DSP device in a format specified in the parameters provided.

The following can be used to configure the module

The value is in hertz, 44100 is the samplerate used on CD’s, but some drivers may prefer 48000 (DAT) or you may want to use something lower.
The number of channels to record. This is typically 2 for stereo or 1 for mono
The device to read the audio samples from, it’s typically /dev/dsp but there maybe more than one card installed.
Check for metadata arriving, if any are present then the data is marked for an update. The metadata is in the form of tag=value, and while Ogg Vorbis can handle any supplied tags, most players will only do anything with artist and title.
The name of the file to open and read the metadata tags from, with this parameter missing standard input is read. Using a file is often the better approach. When using the file access the procedure is usually to populate the file contents then send a SIGUSR1 to the IceS process.
The format of the file itself is a simple one comment per line format, below is a trivial example of the file, other tags can be used but players tend to only look for artist and title for displaying. The data must be in UTF-8 (this is not checked by ices, however).
title=We Will Rock You


The Advanced Linux Sound Architecture (ALSA) is a completely different sound system on linux but provides OSS compatability so the OSS driver should work with it as well. To use ALSA natively a separate module is used

<param name="rate">44100</param>
<param name="channels">2</param>
<param name="device">hw:0,0</param>
<param name="periods">2</param>
<param name="buffer-time">500</param>
<param name="metadata">1</param>
<param name="metadatafilename">/home/ices/metadata</param>

The parameters to ALSA are mostly the same for OSS, as it performs the same task, ie captures audio from the DSP.

This is the device name as used in ALSA. This can be a physical device as in the case of “hw:0,0” or a virtual device like one with dsnoop.
This specifies how many interrupts will be generated (default: 2)
The size of the buffer measured in mS (default 500)


The Sun Solaris DSP input is similar to OSS. It allows for reading from a soundcard on a Sun Solaris UNIX. OpenBSD also has a sound driver that is similar to solaris and as such should be able to use this module.


The parameters are the same as the OSS and ALSA modules.


<param name="rate">44100</param>
<param name="channels">2</param>
<param name="metadata">1</param>
<param name="metadatafilename">/home/ices/metadata</param>

This module should always be available, and as you can see the parameters are almost the same except for the device. The PCM audio comes from the standard input so it can be generated from some external application feeding into a pipe.

As it’s raw PCM being fed in, it’s impossible to determine the samplerate and channels so make sure the stated parameters match the incoming PCM or the audio will be encoded wrongly.


The playlist module is used to get audio from some pre-encoded Ogg Vorbis files. IceS currently checks to see if the same file gets played in succession and skips it, this means that having a playlist repeat with only one Ogg file listed won’t work. The method of file selection is determined by the playlist type. The current types are basic and script.


 <param name="type">basic</param>
 <param name="file">/path/to/playlist</param>
 <param name="random">0</param>
 <param name="once">0</param>
 <param name="restart-after-reread">1</param>
State a path to a file which will contain a list of Ogg Vorbis filenames to play. One per line with lines beginning with ‘#’ being treated as comments. If a line has a single ‘-‘ then standard input is read, which provides a way of getting some external Ogg Vorbis stream into ices.
When set to 1, the playlist will be randomised when the playlist is loaded. By default random is off
When set to 1, the playlist is gone through once and then ends, this will cause ices to exit. By default once is off.
If the playlist is re-read mid way through, which may occur if the playlist was updated then this will restart at the beginning of the playlist. By default it’s off.


<param name="type">script</param>
<param name="program">/path/to/program</param>
State a path to a program which when run will write to it’s standard output a path to an Ogg Vorbis file. The program can be anything from an executable to a shell script as long as it starts, writes the filename to it’s standard output and then exits.


The RoarAudio module is used to get audio from a RoarAudio Sound Server. This module supports getting both already encoded and raw audio from the sound server. It also allows meta data to be read from the sound server or an file the same way the Open Sound module does.

<param name="rate">44100</param>
<param name="channels">stereo</param>
<param name="codec">ogg_vorbis</param>
<param name="aiprofile">default</param>
<param name="dir">monitor</param>
<param name="device">/tmp/roar</param>
<param name="server">home::</param>
<param name="metadata">file</param>
<param name="metadatafilename">/home/ices/metadata</param>
<param name="plugin">helloworld text="Hello ices2 users!"</param>

The parameters are similar to the of the other modules. For the audio parameters (rate, channels, codec) the values from libroar’s audio info profile “default” are used (normally 44100, 2, pcm_s). The following can be used to configure the module:

The value is in hertz, 44100 is the samplerate used on CD’s, but some drivers may prefer 48000 (DAT) or you may want to use something lower. This can also be a symbolic name as supported by libroar. Examples include “cd” and “dat”.
The number of channels to record. This is typically 2 for stereo or 1 for mono. This can also be a symbolic name as supported by libroar. Examples include “stereo” and “mono”.
The codec to read the audio data in from the sound server. Currently the values “default”, “pcm_s”, “pcm_s_le”, “pcm_s_be”, “ogg_vorbis” and “ogg_general” are supported. “default” is an alias for “pcm_s” which itself is an alias to “pcm_s_le” or “pcm_s_be” depending on the native byte order of the system. “ogg_general” should not be used.
Use the given audio info profile from libroar. By default the profile “default” is used. If mixed with rate, channels or codec options the last one set wins.
This is the stream direction. Currently “monitor” and “record” are supported. If set to “monitor” ices2 will stream a copy of the audio played back by sound server to icecast. If set to “record” it will read from soundcard input (e.g. mic). Defaults to “monitor”.
device, server
This sets the location of the sound server. Both parameters are equal. The location can be in any form libroar understands. Common values are server addresses in form “/path/to/sock” for UNIX sockets, “hostname” for IPv4 and IPv6 hosts and “node::” for DECnet nodes. A full description of this can be found in the RoarAudio documentation. Defaults to the list of default locations (auto detection) libroar has.
This is the meta data source. Currently supported values are “none”, “file” and “stream”. The values “none” and “file” are the same as for the Open Sound module the values “0” and “1”. The value “stream” requests meta data from the sound server. Defaults to “stream”.
This setting is the same as “metadatafilename” for the Open Sound module. It is only active when “metadata” is set to “file”.
This setting loads a plugin. The plugin name my be followed by parameters to be passed to the plugin. Between the plugin name and the parameters needs to be a space. Normal libroar plugin parameter parsing rules apply (keys and values separated by “=”, key-value-pairs separated by “,”).