SOX(1) UNIX System V SOX(1)

NAME
sox -SOund eXchange - universal sound sampletranslator

SYNOPSIS
sox infile outfile
sox infile outfile [ effect [ effect options ... ] ]
sox infile -e effect [ effect options ... ]
sox [ general options ] [ format options ] ifile
[ format options ] ofile [ effect [ effect options ... ] ]

General options: [ -V ] [ -v volume ]

Format options: [ -t filetype ] [ -r rate ] [ -s/-u/-U/-A ]
[ -b/-w/-l/-f/-d/-D ] [ -c channels ] [ -x ]
Effects:
copy
rate
avg
stat
echo delay volume [ delay volume ... ]
vibro speed [ depth ]
lowp center
band [ -n ] center [ width ]

DESCRIPTION
Sox translates sound files from one format to another,
possibly doing a sound effect.

OPTIONS
The option syntax is a littlegrotty,but in essence:
sox file.au file.voc
translates a sound sample in SUN Sparc .AU format into a
SoundBlaster .VOC file, while
sox -v 0.5 file.au -rate12000 file.voc rate
does the sameformat translation but also lowers the
amplitude by 1/2 and changes the sampling rate from 8000
hertz to 12000 hertz via the rate sound effect loop.

File type options:

-t filetype
gives the type of the soundsample file.

-r rate Give sample rate in Hertz of file.

-s/-u/-U/-A
Thesample data is signed linear (2's complement),
unsigned linear, U-law (logarithmic), or A-law
(logarithmic). U-law and A-law arethe U.S. and
international standards forlogarithmic telephone
sound compression.

-b/-w/-l/-f/-d/-D
Thesample data is in bytes, 16-bitwords, 32-bit

Page 1 (printed 2/3/93)

SOX(1) UNIX System V SOX(1)

longwords, 32-bit floats, 64-bit double floats, or
80-bit IEEEfloats. Floatsand double floats are
in native machine format.

-x The sample data is in XINU format; that is, it
comes from a machine with the opposite wordorder
than yours and mustbe swapped according tothe
word-size given above. Only 16-bitand 32-bit
integer data may beswapped. Machine-format
floating-point datais not portable. IEEE floats
area fixed, portable format. ???

-c channels
Thenumber of soundchannels in thedata file.
This may be1, 2, or 4; formono, stereo, or quad
sound data.

General options:

-e after the input file allows you to avoid giving an
output fileand just name an effect. This is only
useful with the stat effect.

-v volume Change amplitude (floating point); less than 1.0
decreases, greater than 1.0increases. Note: we
perceive volume logarithmically, not linearly.
Note: see the stat effect.

-V Print a description of processing phases. Useful
for figuring out exactly how sox is mangling your
sound samples.

The input andoutput files may be standard input and output.
This is specified by '-'. The -t type option must be given
in this case, else sox will not know the format of the given
file. The -t, -r, -s/-u/-U/-A, -b/-w/-l/-f/-d/-D and -x
options referto the input data when given before theinput
file name. After, they referto the output data.

If you don't give an output file name, sox will just read
the input file. Thisis useful for validating structured
file formats; the stat effect may also be used via the -e
option.

FILE TYPES
Sox needs to know the formats of the input and output files.
File formats which have headers are checked, if that header
doesn't seem right, the program exitswith anappropriate
message. Currently, the raw (no header), IRCAM SoundFiles,
SoundBlaster, SPARC .AU (w/header), Mac HCOM, PC/DOS.SOU,
Sndtool, and Sounder,NeXT .SND, Windows 3.1 RIFF/WAV,
Turtle Beach .SMP, and Apple/SGI AIFFand 8SVX formats are

Page 2 (printed 2/3/93)

SOX(1) UNIX System V SOX(1)

supported.

.aiff AIFF files used on Apple IIc/IIgs and SGI. Note:
theAIFF format supports only one SSND chunk. It
does not support multiple sound chunks, or the
8SVX musical instrument descriptionformat. AIFF
files are multimedia archives and and can have
multiple audio and picture chunks.You mayneed a
separate archiver to work with them.

.au SUN Microsystems AU files. There are apparently
many types of .au files; DEC has invented its own
with a different magic number and word order. The
.auhandlercan read these files but will not
write them. Some .au fileshave valid AU headers
andsome donot. The latter are probably original
SUNu-law 8000 hz samples.These can be dealt
with using the .ul format (see below).

.hcom Macintosh HCOM files. These are (apparently) Mac
FSSD files with some variant of Huffman
compression. The Macintoshhas wacky file formats
andthis format handler apparently doesn't handle
allthe ones it should. Mac users will need your
usual arsenal of file converters todeal with an
HCOM file under Unix or DOS.

.raw Raw files (no header).
Thesample rate, size (byte, word, etc), and style
(signed, unsigned, etc.) of the sample file must
be given. The number of channels defaults to 1.

.ub, .sb, .uw, .sw, .ul
These are several suffices which serve as a
shorthand for raw files with a given size and
style. Thus, ub, sb, uw, sw, and ul correspond to
"unsigned byte", "signed byte", "unsigned word",
"signed word", and "ulaw" (byte). The sample rate
defaults to8000 hzif not explicitly set, and the
number of channels (as always) defaults to 1.
There are lots of Sparc samples floating around in
u-law format with no headerand fixed at a sample
rate of 8000 hz. (Certain sound management
software cheerfullyignoresthe headers.)
Similarly, most Macsound files arein unsigned
byte formatwith a sample rate of 11025 or 22050
hz.

.sf IRCAM Sound Files.
SoundFiles are usedby academic music software
such as theCSound package,and theMixViewsound
sample editor.

Page 3 (printed 2/3/93)

SOX(1) UNIX System V SOX(1)

.voc Sound Blaster VOC files.
VOCfiles are multi-part and contain silence
parts, looping, anddifferent sample rates for
different chunks. On input, the silence parts are
filled out,loops are rejected, andsample data
with a new sample rate is rejected. Silence with
a differentsample rate is generated
appropriately. On output, silence is not
detected, nor are impossible samplerates.

.auto This is a ``meta-type'': specifying this type for
an input file triggers somecode that triesto
guess the real typeby looking for magic words in
theheader. If thetype can't be guessed, the
program exits with an errormessage. The input
must be a plain file, not apipe. This type can't
be used foroutput files.

.smp Turtle Beach SampleVision files.
SMPfiles are for use with the PC-DOS package
SampleVision by Turtle Beach Softworks. This
package is for communication to several MIDI
samplers. All sample rates are supported bythe
package, although not all are supported by the
samplers themselves. Currently looppoints are
ignored.

.wav Windows 3.1 .WAV RIFF files.
These appear to be very similar to IFF files, but
notthe same. They are the native sound file
format of Windows 3.1. Obviously, Windows 3.1 is
of such incredible importance to the computer
industry that it just had to have its own sound
file format.

EFFECTS
Only one effect from the palette may be applied to a sound
sample. To do multiple effects you'll need to run sox in a
pipeline.

copyCopy the input file to the
output file. This is the
defaulteffect if both files
have the same sampling rate,
or the rates are "close".

rateTranslate inputsampling rate
to output sampling ratevia
linear interpolation tothe
Least Common Multiple of the
two sampling rates. This is
the default effect if the two

Page 4 (printed 2/3/93)

SOX(1) UNIX System V SOX(1)

files have different sampling
rates.This isfast but
noisy.

avgMix 4- or 2-channel sound file
into 2-or 1-channel file by
averaging the samples for
different speakers.

statDo a statistical check on the
input file, andprint results
on the standarderror file.
stat may copy the file
untouched from input to
output,if you select an
output file. The "Volume
Adjustment:" field in the
statistics gives you the
argument to the -v number
which will makethe sample as
loud aspossible.

echo [ delay volume ... ] Add echoing to a sound sample.
Each delay/volume pair gives
the delay in seconds and the
volume (relative to 1.0) of
that echo. If the volumes add
up to more than1.0, the sound
will melt down instead of
fading away.

vibro speed [ depth ] Add the world-famous Fender
Vibro-Champ sound effect to a
sound sample byusing asine
wave asthe volume knob.
Speed gives the Hertz value of
the wave. Thismust beunder
30. Depth gives the amount
the volume is cut into by the
sine wave, ranging 0.0 to 1.0
and defaulting to 0.5.

lowp center Apply a low-pass filter. The
frequency response drops
logarithmically with center
frequency in the middleof the
drop. The slope of thefilter
is quite gentle.

band [ -n ] center [ width ] Apply a band-pass filter. The
frequency response drops
logarithmicallyaround the

Page 5 (printed 2/3/93)

SOX(1) UNIX System V SOX(1)

center frequency. The width
gives the slopeof the drop.
The frequencies at center +
width and center - width will
be halfof their original
amplitudes. Band defaults to
a mode orientedto pitched
signals, i.e. voice, singing,
or instrumental music. The -n
(for noise) option usesthe
alternate mode for un-pitched
signals. Band introduces
noise in the shape of the
filter,i.e. peaking atthe
center frequency and settling
around it.

Sox enforces certain effects. If the two files have
different sampling rates, therequested effect must be one
of copy, or rate, If the two files have different numbers of
channels, the avg effect must be requested.

BUGS
The syntax ishorrific. It'svery tempting to include a
default system that allows aneffect name as the program
name and justpipes asound sample from standard input to
standard output, but the problem of inputtingthe sample
ratesmakes this unworkable.

FILES
SEE ALSO
NOTICES
The echoplex effect is:
Copyright(C) 1989 by JefPoskanzer.
Permission to use, copy, modify, and distribute this
software and its
documentation forany purpose andwithoutfee is hereby
granted, provided
that the above copyright notice appear inall copies and
that both that
copyrightnotice and thispermission notice appear in
supporting
documentation. This software is provided"as is"
without express or
implied warranty.

Page 6 (printed 2/3/93)