out123 - Man Page

send raw PCM audio or a waveform pattern to an output device


cat audio.raw | out123 [ - ] [ options ]

out123 [ options ] filename [ filename ... ]

out123 --wave-freq freq1[,freq2,...] [ options ]

out123 --source geiger [ options ]


out123 reads raw PCM data (in host byte order) from standard input  and plays it on the audio device specified by given options. Alternatively, it can generate periodic or random signals for playback itself.


out123 options may be either the traditional POSIX one letter options, or the GNU style long options.  POSIX style options start with a single '-', while GNU long options start with '--'. Option arguments (if needed) follow separated by whitespace (not '='). Note that some options can be absent from your installation when disabled in the build process.

--name name

Set the name of this instance, possibly used in various places. This sets the client name for JACK output.

-o module, --output module

Select audio output module. You can provide a comma-separated list to use the first one that works. Also see -a.


List the available modules.

-a dev, --audiodevice dev

Specify the audio device to use.  The default as well as the possible values depend on the active output. For the JACK output, a comma-separated list of ports to connect to (for each channel) can be specified.  

-s,  --stdout

The audio samples are written to standard output, instead of playing them through the audio device. The output format is the same as the input ... so in this mode, out123 acts similar the standard tool cat, possibly with some conversions involved. This shortcut is equivalent to '-o raw -a -'.


This variant additionally writes the data to stdout, while still playing it on the output device. So it is more like some flavour of tee than a cat.

-O file, --outfile

Write raw output into a file (instead of simply redirecting standard output to a file with the shell). This shortcut is equivalent to '-o raw -a file'.

-w file, --wav

Write output as WAV file file , or standard output if - is or the empty string used as file name. You can also use --au and --cdr for AU and CDR format, respectively. Note that WAV/AU writing to non-seekable files or redirected stdout needs some thought. The header is written with the first actual data. The result of decoding nothing to WAV/AU is a file consisting just of the header when it is seekable and really nothing when not (not even a header). Correctly writing data with prophetic headers to stdout is no easy business. This shortcut is equivalent to '-o wav -a file'.

--au file

Write to file in SUN audio format.  If - or the empty string is used as the filename, the AU file is written to stdout. See paragraph about WAV writing for header fun with non-seekable streams. This shortcut is equivalent to '-o au -a file'.

--cdr file

Write to file as a CDR (CD-ROM audio, more correctly CDDA for Compact Disc Digital Audio). If - is or the empty string used as the filename, the CDR file is written to stdout. This shortcut is equivalent to '-o cdr -a file'.

-r rate, --rate rate

Set sample rate in Hz (default: 44100). If this does not match the actual input sampling rate, you get changed pitch. Might be intentional;-)

-R rate, --inputrate rate

Set input sample rate to a different value. This triggers resampling if the output rate is indeed different. See --resample.

--speed factor

Speed up/down playback by that factor using resampling. See --resample.

--resample method

This chooses the method for resampling between differing sampling rates or to apply a change in tempo. You can choose between two variants of the syn123 resampler: fine (the default) and dirty. The fine one features 108 dB dynamic range and at worst-case 84% bandwidth. The dirty one uses a bit less CPU time (not that much, though) by reducing the dynamic range to 72 dB with worst-case bandwidth of 85%. The exact properties vary with the sampling rate ratio, as there is interpolation of filter coefficients involved.

-c count, --channels count

Set channel count to given value.

-C count, --inputch count

Set input channel count to a differnt value than for output. This probably means you want some remixing. Also see --mix.

-e enc, --encoding enc

Choose output sample encoding. Possible values look like f32 (32-bit floating point), s32 (32-bit signed integer), u32 (32-bit unsigned integer) and the variants with different numbers of bits (s24, u24, s16, u16, s8, u8) and also special variants like ulaw and alaw 8-bit. See the output of out123's longhelp for actually available encodings. Default is s16.

--endian choice

Select output endianess (byte order). Choice is big, little, or native, which is the default. The processing can only work in native mode, so you need to specify input or output byte order if that does not match your machine. This also sets the input endianess if that is not set separately. See also --inputend and --byteswap.

-E enc, --inputenc enc

Specify input encoding different from output encoding for conversion.

--inputend choice

Select input endianess (byte order). By default it is the same as output byte order. See --endian.


A switch to trigger swapping of byte order just before output, after any other transformations. This works on top of any endianess you specify with

-m,  --mono

Set for single-channel audio (default is two channels, stereo).


Select stereo output (2 channels, default).


List known encoding short and long names to standard output.

--mix matrix

Specify a mixing matrix between input and output channels as linear factors, comma separated list for the input channel factors for output channel 1, then output channel 2, and so forth. The default is a unit matrix if channel counts match, so for 3 channels the equivalent of both channels with halved amplitude, so '--mix 0.5,0.5'. For splitting mono to stereo, it is '--mix 1,1' top keep the symmetry.

--filter coeff

Apply digital filter(s) before pre-amplification (see --preamp) with the coefficient list coeff as


where a_0=1 is mandatory and perhaps helps orientation a bit. Multiple filters are separated by ':'.

-P dbvalue --preamp dbvalue

Enable a pre-amplification stage that amplifies the signal with the given value in dB before output.

--offset value

Apply a PCM offset (floating point value scaled in [-1:1] in the pre-amplification stage. Normally, you would do that to correct a known DC offset in a recording.

--clip mode

Select clipping mode: 'soft' or 'hard' for forced clipping also for floating point output, 'implicit' (default) for implied hard clipping during conversion where necessary.


Enable dithering for conversions to integer. If you insist. This is just some un-spectacular TPDF dither. For some people, that is not fancy enough. Most people cannot be bothered that way or the other.


Check if given format is supported by given driver and device (in command line before encountering this), silently returning 0 as exit value if it is the case.


Print out the short names of encodings supported with the current setup.


If the selected driver and device communicate some default accepted format, print out a command line fragment for out123 setting that format, always in that order: --rate <r> --channels <c> --encoding <e>

-o h, --headphones

Direct audio output to the headphone connector (some hardware only; AIX, HP, SUN).

-o s, --speaker

Direct audio output to the speaker  (some hardware only; AIX, HP, SUN).

-o l, --lineout

Direct audio output to the line-out connector (some hardware only; AIX, HP, SUN).

-b size, --buffer size

Use an audio output buffer of size Kbytes.  This is useful to bypass short periods of heavy system activity, which would normally cause the audio output  to be interrupted.   You should specify a buffer size of at least 1024  (i.e. 1 Mb, which equals about 6 seconds of usual audio data) or more;  less than about 300 does not make much sense.  The default is 0,  which turns buffering off.

--preload fraction

Wait for the buffer to be filled to fraction before starting playback (fraction between 0 and 1). You can tune this prebuffering to either get sound faster to your ears or safer uninterrupted web radio. Default is 0.2 (changed from 1 since version 1.23).

--devbuffer seconds

Set device buffer in seconds; <= 0 means default value. This is the small buffer between the application and the audio backend, possibly directly related to hardware buffers.

--timelimit samples

Set playback time limit in PCM samples if set to a value greater than zero. out123 will stop reading from stdin or playing from the generated wave table after reaching that number of samples.

--seconds seconds

Set time limit in seconds instead.

--source name

Choose the signal source: 'file' (default) for playback of the given file(s) on the command line or standard input if there are none, or one of the generators 'wave' (see --wave-freq), geiger (see --geiger-activity), or just 'white' for some white noise.

--wave-freq frequencies

Set wave generator frequency or list of those with comma separation for enabling a generated test signal instead of standard input. Empty values repeat the previous one.

--wave-pat patterns

Set the waveform patterns of the generated waves as comma-separated list. Choices include sine, square, triangle, sawtooth, gauss, pulse, and shot. Empty values repeat the previous one.

--wave-phase phases

Set waveform phase shift(s) as comma-separated list, negative values inverting the pattern in time and empty value repeating the previous. There is also --wave-direction overriding the negative bit.


Set wave direction explicitly (the sign counts).

--wave-sweep frequency

Sweep a generated wave to the given frequency, from first one specified for --wave-freq, using the first wave pattern and direction, too.

--sweep-time seconds

Set frequency sweep duration in seconds if > 0. This defaults to the configured time limit if set, otherwise one second, as endless sweeps are not sensible.

--sweep-count count

Set timelimit to exactly produce that many (smooth) sweeps

--sweep-type type

Set sweep type: lin(ear) for linear, qua(d) (default) for quadratic, or exp(onential) for an exponential change of frequency with time.


Disable post-sweep smoothing for periodicity.

--genbuffer bytes

Set the buffer size (limit) for signal generators, if > 0 (default), this enforces a periodic buffer also for non-periodic signals, benefit: less runtime CPU overhead, as everything is precomputed as enforced periodic signal.

--wave-limit samples

This is an alias for --genbuffer.

--pink-rows number

Activate pink noise source and choose rows for the algorithm (<1 chooses default). The generator follows code provided by Phil Burk (http://softsynth.com) and uses the Gardner method.

--geiger-activity number

This configures the simulation of a Geiger-Mueller counter as source, with the given numer as average events per second. Play with it. It's fun! -t, --test Test mode.  The audio stream is read, but no output occurs.

-v,  --verbose

Increase the verbosity level.

-q,  --quiet

Quiet.  Suppress diagnostic messages.


Tries to get higher priority

-T,  --realtime

Tries to gain realtime priority.  This option usually requires root privileges to have any effect.

-?,  --help

Shows short usage instructions.


Shows long usage instructions.


Print the version string.



Thomas Orgis <maintainer@mpg123.org>, <thomas@orgis.org>

Creator (ancestry of code inside mpg123):

Michael Hipp

Uses code or ideas from various people, see the AUTHORS file accompanying the source code.


out123 is licensed under the GNU Lesser/Library General Public License, LGPL, version 2.1 .




26 Apr 2020