.TH AUDIO 1
.SH NAME
mp3dec, mp3enc, oggdec, oggenc, flacdec, flacenc, sundec, wavdec, pcmconv, mixfs \- decode and encode audio files
.SH SYNOPSIS
.B audio/mp3dec
[
.B -s
.I seconds
] [
.B -d
]
.br
.B audio/oggdec
[
.B -s
.I seconds
]
.br
.B audio/flacdec
[
.B -s
.I seconds
]
.br
.B audio/wavdec
[
.B -s
.I seconds
]
.br
.B audio/sundec
.PP
.B audio/oggenc
.br
.B audio/mp3enc
[
.B -hprv
] [
.B -b
.I bitrate
] [
.B -B
.I bitrate
] [
.B -m
.I mode
] [
.B -q
q ] [
.B -s
.I sfreq
] [
.B -V
.I q
] [
.I "long or silly options"
]
.br
.B audio/flacenc
[
.B -b
.I bitspersample
] [
.B -c
.I channels
] [
.B -l
.I compresslevel
] [
.B -s
.I sfreq
] [
.B -P
.I padding
] [
.B -T
.I field=value
]
.PP
.B audio/pcmconv
[
.B -i
.I fmt
] [
.B -o
.I fmt
] [
.B -l
.I length
]
.PP
.B audio/mixfs
[
.B -D
] [
.B -s
.I srvname
] [
.B -m
.I mtpt
]
.PP
.SH DESCRIPTION
These programs decode and encode various audio formats from and to
16-bit stereo PCM (little endian). The decoders read the compressed
audio data from standard input and produce PCM on standard output at
a sampling frequency of 44.1KHz.
.PP
.I Mp3dec
decodes MPEG audio (layer 1, 2 and 3). The
.B -d
option enables debug output to standard error.
.I Oggdec,
.I flacdec,
.I sunwdec
and
.I wavdec
are like
.I mp3dec
but decode OGG Vorbis, FLAC lossless audio, Sun audio and RIFF wave.
.SS Decoding options
.TF "\fB-s seconds \fP"
.TP
.BI -s " seconds"
seek to a specific position in seconds before decoding.
.SS
.PP
The encoders read PCM on standard input and produce compressed audio
on standard output.
.PP
.I Flacenc,
.I oggenc
and
.I mp3enc
produce FLAC, OGG Vorbis and MP3 audio. For
.I mp3enc,
the MP3 file will use `constant bit-rate' (CBR) encoding by default, 
but that can be changed via
.B --abr
(average bitrate desired, ABR)
or
.BR -v
(variable bitrate, VBR).
.PP
.I Flacenc
and
.I oggenc
accept raw PCM in the same byte order as
.B /dev/audio
(little-endian),
while
.I mp3enc -r
expects big-endian.
.SS Encoding options
.TF "\fB-b \fP"
.TP
.B -b
set minimum allowed
.I bitrate
in Kb/s for VBR, default 32Kb/s.
For CBR,
set the exact bitrate in Kb/s, which defaults to 128Kb/s.
.TP
.B -B
set maximum allowed
.I bitrate
in Kb/s for VBR, default 256Kb/s.
.TP
.BI -h
same as
.LR "-q 2" .
.TP
.B -m
.I mode
may be
(s)tereo,
(j)oint,
(f)orce
or
(m)ono
(default j).
.B force
forces mid/side stereo on all frames.
.TP
.B -p
add CRC error protection (adds an additional 16 bits per frame to the stream).
This seems to break playback.
.TP
.B -q
sets output quality to
.I q
(see
.BR -V ).
.TP
.B -r
input is raw pcm
.TP
.B -s
set sampling frequency of input file (in KHz) to
.IR sfreq ,
default is 44.1.
.TP
.B -v
use variable bitrate (VBR) encoding
.TP
.B -V
set quality setting for VBR to
.IR q .
Default
.I q
is 4;
0 produces highest-quality and largest files, and
9 produces lowest-quality and smallest files.
.SS Long options
.TF "\fB--resample sfreq \fP"
.TP
.BI --abr " bitrate"
sets average
.I bitrate
desired in Kb/s, instead of setting quality,
and generates ABR encoding.
.TP
.BI --resample " sfreq"
set sampling frequency of output file (in KHz) to
.IR sfreq ,
default is input sfreq.
.TP
.BI --mp3input " input"
is an MP3 file
.
.SS Silly options
.TF --nohist
.TP
.B -f
same as
.LR "-q 7" .
Such a deal. 
.TP
.B -o
mark as non-original (i.e. do not set the original bit)
.TP
.B -c
mark as copyright
.TP
.B -k
disable sfb=21 cutoff
.TP
.BI -e " emp"
de-emphasis n/5/c
(default n)
.TP
.B -d
allow channels to have different blocktypes
.TP
.B -t
disable Xing VBR informational tag
.TP
.B -a
autoconvert from stereo to mono file for mono encoding
.TP
.B -x
force byte-swapping of input (see
.IR dd (1)
instead)
.TP
.B -S
don't print progress report, VBR histograms
.TP
.B --athonly
only use the ATH for masking
.TP
.B --nohist
disable VBR histogram display
.TP
.B --voice
experimental voice mode
.SS
.I Pcmconv
is a helper program used to convert various PCM sample formats. The
.B -i
and
.B -o
options specify the input and output format
.I fmt
of the conversion.
.I Fmt
is a concatenated string of the following parts:
.TP
.BI s #
sample format is little-endian signed integer where
.I #
specifies the number of bits
.TP
.BI u #
unsigned little-endian integer format
.TP
.BI S #
signed big-endian integer format
.TP
.BI U #
unsigned big-endian integer format
.TP
.BI f #
floating point format where
.I #
has to be 32 or 64 for single- or double-precision
.TP
.B a8
8-bit a-law format
.TP
.B µ8
8-bit µ-law format
.TP
.BI c #
specifies the number of channels
.TP
.BI r #
gives the samplerate in Hz
.PP
The program reads samples from standard
input converting the data and writes the result to standard output
until it reached end of file or, if
.B -l
was given, a number of
.I length
bytes have been consumed from input.
.PP
.I Mixfs
is a fileserver serving a single
.B audio
file which allows simultaneous playback of audio streams. When
run, it binds over
.B /dev/audio
and mixes the audio samples that are written to it.
A service name
.I srvname
can be given with the
.B -s
option which gets posted to
.BR /srv .
By default,
.I mixfs
mounts itself on
.B /mnt/mix
and then binds
.B /mnt/mix/audio
over
.BR /dev .
A alternative mountpoint 
.I mtpt
can be specified with the
.B -m
option.
The
.B -D
option causes
.B 9p
debug messages to be written to file-descriptor 2.
.SH EXAMPLE
Play back an
.L .mp3
.IP
.EX
audio/mp3dec <foo.mp3 >/dev/audio
.EE
.LP
Encode a
.L .wav
file as highest-quality MP3.
.IP
.EX
audio/mp3enc -q 0 -b 320 <foo.wav >foo.mp3
.EE
.LP
Create a fixed 128Kb/s MP3 file from a
.L .wav
file.
.IP
.EX
audio/mp3enc -h <foo.wav >foo.mp3
.EE
.LP
Streaming from stereo 44.1KHz raw PCM data, encoding mono at 16KHz
(you may not need
.IR dd ):
.IP
.EX
dd -conv swab | audio/mp3enc -a -r -m m --resample 16 -b 24
.EE
.SH SOURCE
.B /sys/src/cmd/audio
.SH "SEE ALSO"
.IR play (1),
.IR zuke (1)
.br
.B http://www.underbit.com/products/mad/
.br
.B http://xiph.org/doc/
.br
.B http://flac.sourceforge.net/documentation.html
.SH HISTORY
.I Pcmconv
first appeared in 9front (December, 2012).
.I Mixfs
first appeared in 9front (December, 2013).
.I Flacenc
first appeared in 9front (January, 2021).