From Linux Man Pages
sox - Sound eXchange : universal sound sample translator
SYNOPSIS
sox infile1 [ infile2 ... ] outfile
sox [ general options ] [ format options ] infile1
[ [ format options ] infile2 ... ] [ format options ] outfile
[ effect [ effect options ] ... ]
soxmix infile1 infile2 [ infile3 ... ] outfile
soxmix [ general options ] [ format options ] infile1
[ format options ] infile2
[ [ format options ] infile3 ... ]
[ format options ] outfile
[ effect [ effect options ] ... ]
General options:
[ -h ] [ -p ] [ -q ] [ -S ] [ -V ]
Format options:
[ -t filetype ] [ -r rate ] [ -s/-u/-U/-A/-a/-i/-g/-f ]
[ -b/-w/-l/-d ] [ -v volume ]
[ -c channels ] [ -x ] [ -e ]
Effects:
avg [ -l | -r | -f | -b | -1 | -2 | -3 | -4 | n,n,...,n ]
band [ -n ] center [ width ]
bandpass frequency bandwidth
bandreject frequency bandwidth
chorus gain-in gain out delay decay speed depth
-s | -t [ delay decay speed depth -s | -t ]
compand attack1,decay1[,attack2,decay2...]
in-dB1,out-dB1[,in-dB2,out-dB2...]
[ gain [ initial-volume [ delay ] ] ]
copy
dcshift shift [ limitergain ]
deemph
earwax
echo gain-in gain-out delay decay [ delay decay ... ]
echos gain-in gain-out delay decay [ delay decay ... ]
fade [ type ] fade-in-length
[ stop-time [ fade-out-length ] ]
filter [ low ]-[ high ] [ window-len [ beta ]]
flanger gain-in gain-out delay decay speed < -s | -t >
highp frequency
highpass frequency
lowp frequency
lowpass frequency
mask
mcompand "attack1,decay1[,attack2,decay2...]
in-dB1,out-dB1[,in-dB2,out-dB2...]
[ gain [ initial-volume [ delay ] ] ]" xover_freq
noiseprof [profile-file]
noisered profile-file [threshold]
pan direction
phaser gain-in gain-out delay decay speed < -s | -t >
pick [ -1 | -2 | -3 | -4 | -l | -r | -f | -b ]
pitch shift [ width interpole fade ]
polyphase [ -w < nut / ham > ]
[ -width < long / short / # > ]
[ -cutoff # ]
rate
repeat count
resample [ -qs | -q | -ql ] [ rolloff [ beta ] ]
reverb gain-out reverb-time delay [ delay ... ]
reverse
silence above_periods [ duration threshold[ d | % ]
[ below_periods duration
threshold[ d | % ]]
speed [ -c ] factor
stat [ -s n ] [ -rms ] [ -v ] [ -d ]
stretch [ factor [ window fade shift fading ]
swap [ 1 2 | 1 2 3 4 ]
synth [ length ] type mix [ freq [ -freq2 ]
[ off ] [ ph ] [ p1 ] [ p2 ] [ p3 ]
trim start [ length ]
vibro speed [ depth ]
vol gain [ type [ limitergain ] ]
DESCRIPTION
SoX is a command line program that can convert most popular audio files to most other popular audio file formats.
It can optionally change the audio sample data type and apply one or more sound effects to the file during this
translation.
If more than one input file is specified then they are concatenated into the output file. In this case, it has a
restriction that all input files must be of the same data type and sample rates.
soxmix is functionally the same as the command line program sox expect that it takes two or more files as input
and mixes the audio together to produce a single file as output. It has a restriction that all input files must
be of the same data type and sample rates.
There are two types of audio file formats that SoX can work with. The first are self-describing file formats.
These contain a header that completely describe the characteristics of the audio data that follows.
The second type are header-less data, or sometimes called raw data. A user must pass enough information to SoX
on the command line so that it knows what type of data it contains.
Audio data can usually be totally described by four characteristics:
rate The sample rate is in samples per second. For example, CD sample rates are at 44100.
data size The precision the data is stored in. Most popular are 8-bit bytes or 16-bit words.
data encoding
What encoding the data type uses. Examples are u-law, ADPCM, or signed linear data.
channels How many channels are contained in the audio data. Mono and Stereo are the two most common.
Please refer to the soxexam(1) manual page for a long description with examples on how to use SoX with various
types of file formats.
OPTIONS
The option syntax is a little grotty, but in essence:
sox file.au file.wav
translates a sound file in SUN Sparc .AU format into a Microsoft .WAV file, while
sox -v 0.5 file.au -r 12000 file.wav mask
does the same format translation but also lowers the amplitude by 1/2, changes the sampling rate to 12000 hertz,
and applies the mask sound effect to the audio data.
The following will mix two sound files together to to produce a single sound file.
soxmix music.wav voice.wav mixed.wav
Filenames:
SoX can be used as a part of pipe operations by using the special filenames of "-". If specified as an input
name, it will read data from stdin. If specified as an output name, it will send data to stdout.
General options:
-h Print version number and usage information.
-p Run in preview mode and run fast. This will somewhat speed up SoX when the output format has a differ-
ent number of channels and a different rate than the input file. Currently, this defaults to using the
rate effect instead of the resample effect for sample rate changes.
-q Run in quite mode when SoX wouldn't otherwise do that. Inverse of -S option.
-S Print status while processing audio data. Tells how much of audio data has been processed in terms of
audio running time instead of samples.
-V Print a description of processing phases. Useful for figuring out exactly how SoX
is mangling your sound samples.
Format options:
Format options effect the input or output file that they immediately precede.
Self describing input files can obtain all the format information directly from the header and so don't generally
need format options. Headerless input files lack this information and so format options must be used to inform
SoX of the file's data type, sample rate, and number of channels.
By default, SoX attempts to write audio data using the same data type, sample rate, and channel count as the
input data. If the user wants the output file to be of a different format then format options can be used to
specify the differences.
If an output file format doesn't support the same data type, sample rate, or channel count as the input file for-
mat, then SoX will auto select the closest values it does support so that the user does not have to specify these
format change options manually.
-t filetype
gives the type of the sound sample file. Useful when file extension is not standard or can not be
determeind by looking at the header of the file.
-r rate Gives the sample rate in Hertz of the file. To cause the output file to have a different sample rate
than the input file, include this option as a part of the output format options.
If the input and output files have different rates then a sample rate change effect must be ran. Since
SoX has multiple rate changing effects, the user can specify which to use as an effect. If no rate
change effect is specified then a default one will be chosen.
-v volume Change amplitude (floating point); less than 1.0 decreases, greater than 1.0 increases. May use a neg-
ative number to invert the phase of the audio data. It is interesting to note that we perceive volume
logarithmically but this adjusts the amplitude linearly.
As with other format options, the volume option effects the file its specified with. This is useful
whe processing mutiple input files as the volume adjustment can be specified for each input file or
just once to adjust the output file. This can be compared to an audio mixer were you can control the
volume of each input as well as a master volume (output side).
soxmix defaults the value of the -v option for each input file to 1/input_file_count. This means if
your mixing two input files together then each input file's volume is adjusted by 0.5. This is done to
prevent clipping of audio data during the mixing operation. Users will most likely not be happy with
this large of a volume adjustment and can specify the -v option to override this default value.
Note: For the non-mixing case, see the stat effect for information on finding the maximum volume
adjustment that can be done with this option without causing audio data to be clipped.
-s/-u/-U/-A/-a/-i/-g/-f
The sample data encoding is signed linear (2's complement), unsigned linear, u-law (logarithmic), A-law
(logarithmic), ADPCM, IMA_ADPCM, GSM, or Floating-point.
U-law (actually shorthand for mu-law) and A-law are the U.S. and international standards for logarith-
mic telephone sound compression. When uncompressed u-law has roughly the precision of 14-bit PCM audio
and A-law has roughly the precision of 13-bit PCM audio.
A-law and u-law data is sometimes encoded using a reversed bit-ordering (ie. MSB becomes LSB). Inter-
nally, SoX understands how to work with this encoding but there is currently no command line option to
specify it. If you need this support then you can use the psuedo file types of ".la" and ".lu" to
inform sox of the encoding. See supported file types for more information.
ADPCM is a form of sound compression that has a good compromise between good sound quality and fast
encoding/decoding time. It is used for telephone sound compression and places were full fidelity is
not as important. When uncompressed it has roughly the precision of 16-bit PCM audio. Popular version
of ADPCM include G.726, MS ADPCM, and IMA ADPCM. The -a flag has different meanings in different file
handlers. In .wav files it represents MS ADPCM files, in all others it means G.726 ADPCM. IMA ADPCM
is a specific form of ADPCM compression, slightly simpler and slightly lower fidelity than Microsoft's
flavor of ADPCM. IMA ADPCM is also called DVI ADPCM.
GSM is a standard used for telephone sound compression in European countries and its gaining popularity
because of its quality. It usually is CPU intensive to work with GSM audio data.
-b/-w/-l/-d
The sample data size is in bytes, 16-bit words, 32-bit long words, or 64-bit double long (long long)
words.
-x The sample data is in XINU format; that is, it comes from a machine with the opposite word order than
yours and must be swapped according to the word-size given above. Only 16-bit and 32-bit integer data
may be swapped. Machine-format floating-point data is not portable.
-c channels
The number of sound channels in the data file. This may be 1, 2, or 4; for mono, stereo, or quad sound
data. To cause the output file to have a different number of channels than the input file, include
this option with the output file options. If the input and output file have a different number of
channels then the avg effect must be used. If the avg effect is not specified on the command line it
will be invoked internally with default parameters.
-e When specified after the last input filename (so that it applies to the output file) it allows you to
avoid giving an output filename and will not produce an output file. It will apply any specified
effects to the input file. This is mainly useful with the stat effect but can be used.
FILE TYPES
SoX attempts to determine the file type of input files automatically by looking at the header of the audio file.
When it is unable to detect the file type or if its an output file then it uses the file extension of the file to
determine what type of file format handler to use. This can be overridden by specifying the "-t" option on the
command line.
The input and output files may be read from standard in and out. This is done by specifying '-' as the filename.
File formats which have headers are checked, if that header doesn't seem right, the program exits with an appro-
priate message.
The following file formats are supported:
.8svx Amiga 8SVX musical instrument description format.
.aiff AIFF files used on Apple IIc/IIgs and SGI. Note: the AIFF format supports only one SSND chunk. It
does not support multiple sound chunks, or the 8SVX musical instrument description format. AIFF files
are multimedia archives and can have multiple audio and picture chunks. You may need a separate
archiver to work with them.
.alsa ALSA /dev/snd/pcmCxDxp device driver
This is a pseudo-file type and can be optionally compiled into SoX. Run sox -h to see if you have sup-
port for this file type. When this driver is used it allows you to open up the ALSA /dev/snd/pcmCxDxp
file and configure it to use the same data format as passed in to SoX. It works for both playing and
recording sound samples. When playing sound files it attempts to set up the ALSA driver to use the
same format as the input file. It is suggested to always override the output values to use the highest
quality samples your sound card can handle. Example: sox infile -t alsa -w -s /dev/snd/pcmC0D0p
.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 .au handler can read these files but will not write them.
Some .au files have valid AU headers and some do not. The latter are probably original SUN u-law 8000
hz samples. These can be dealt with using the .ul format (see below).
.avr Audio Visual Research
The AVR format is produced by a number of commercial packages on the Mac.
.cdr CD-R
CD-R files are used in mastering music on Compact Disks. The audio data on a CD-R disk is a raw audio
file with a format of stereo 16-bit signed samples at a 44khz sample rate. There is a special block-
ing/padding oddity at the end of the audio file and is why it needs its own handler.
.cvs Continuously Variable Slope Delta modulation
Used to compress speech audio for applications such as voice mail.
.dat Text Data files
These files contain a textual representation of the sample data. There is one line at the beginning
that contains the sample rate. Subsequent lines contain two numeric data items: the time since the
beginning of the first sample and the sample value. Values are normalized so that the maximum and min-
imum are 1.00 and -1.00. This file format can be used to create data files for external programs such
as FFT analyzers or graph routines. SoX can also convert a file in this format back into one of the
other file formats.
.gsm GSM 06.10 Lossy Speech Compression
A standard for compressing speech which is used in the Global Standard for Mobil telecommunications
(GSM). Its good for its purpose, shrinking audio data size, but it will introduce lots of noise when a
given sound sample is encoded and decoded multiple times. This format is used by some voice mail
applications. It is rather CPU intensive.
GSM in SoX is optional and requires access to an external GSM library. To see if there is support for
gsm run sox -h and look for it under the list of supported file formats.
.hcom Macintosh HCOM files. These are (apparently) Mac FSSD files with some variant of Huffman compression.
The Macintosh has wacky file formats and this format handler apparently doesn't handle all the ones it
should. Mac users will need your usual arsenal of file converters to deal with an HCOM file under Unix
or DOS.
.maud An Amiga format
An IFF-conform sound file type, registered by MS MacroSystem Computer GmbH, published along with the
"Toccata" sound-card on the Amiga. Allows 8bit linear, 16bit linear, A-Law, u-law in mono and stereo.
.mp3 MP3 Compressed Audio
MP3 audio files come from the MPEG standards for audio and video compression. They are a lossy com-
pression format that achieves good compression rates with a minimum amount of quality loss. Also see
Ogg Vorbis for a similar format. MP3 support in SoX is optional and requires access to either or both
the external libmad and libmp3lame libraries. To see if there is support for Mp3 run sox -h and look
for it under the list of supported file formats as "mp3".
.nul Null file handler. This is a fake file hander that act as if its reading a stream of 0's from a while
or fake writing output to a file. This is not a very useful file handler in most cases. It might be
useful in some scripts were you do not want to read or write from a real file but would like to specify
a filename for consistency.
.ogg Ogg Vorbis Compressed Audio.
Ogg Vorbis is a open, patent-free CODEC designed for compressing music and streaming audio. It is sim-
ilar to MP3, VQF, AAC, and other lossy formats. SoX can decode all types of Ogg Vorbis files, but can
only encode at 128 kbps. Decoding is somewhat CPU intensive and encoding is very CPU intensive.
Ogg Vorbis in SoX is optional and requires access to external Ogg Vorbis libraries. To see if there is
support for Ogg Vorbis run sox -h and look for it under the list of supported file formats as "vorbis".
ossdsp OSS /dev/dsp device driver
This is a pseudo-file type and can be optionally compiled into SoX. Run sox -h to see if you have sup-
port for this file type. When this driver is used it allows you to open up the OSS /dev/dsp file and
configure it to use the same data format as passed in to SoX. It works for both playing and recording
sound samples. When playing sound files it attempts to set up the OSS driver to use the same format as
the input file. It is suggested to always override the output values to use the highest quality sam-
ples your sound card can handle. Example: sox infile -t ossdsp -w -s /dev/dsp
.prc Psion record.app
Used in some Psion devices for System alarms. This format is newer then the .wve format that is used
in some Psion devices.
.sf IRCAM Sound Files.
Sound Files are used by academic music software such as the CSound package, and the MixView sound sam-
ple editor.
.sph
SPHERE (SPeech HEader Resources) is a file format defined by NIST (National Institute of Standards and
Technology) and is used with speech audio. SoX can read these files when they contain u-law and PCM
data. It will ignore any header information that says the data is compressed using shorten compression
and will treat the data as either u-law or PCM. This will allow SoX and the command line shorten pro-
gram to be ran together using pipes to uncompress the data and then pass the result to SoX for process-
ing.
.smp Turtle Beach SampleVision files.
SMP files 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 by the package, although not
all are supported by the samplers themselves. Currently loop points are ignored.
.snd
Under DOS this file format is the same as the .sndt format. Under all other platforms it is the same
as the .au format.
.sndt SoundTool files.
This is an older DOS file format.
sunau Sun /dev/audio device driver
This is a pseudo-file type and can be optionally compiled into SoX. Run sox -h to see if you have sup-
port for this file type. When this driver is used it allows you to open up a Sun /dev/audio file and
configure it to use the same data type as passed in to SoX. It works for both playing and recording
sound samples. When playing sound files it attempts to set up the audio driver to use the same format
as the input file. It is suggested to always override the output values to use the highest quality
samples your hardware can handle. Example: sox infile -t sunau -w -s /dev/audio or sox infile -t sunau
-U -c 1 /dev/audio for older sun equipment.
.txw Yamaha TX-16W sampler.
A file format from a Yamaha sampling keyboard which wrote IBM-PC format 3.5" floppies. Handles reading
of files which do not have the sample rate field set to one of the expected by looking at some other
bytes in the attack/loop length fields, and defaulting to 33kHz if the sample rate is still unknown.
.vms More info to come.
Used to compress speech audio for applications such as voice mail.
.voc Sound Blaster VOC files.
VOC files are multi-part and contain silence parts, looping, and different sample rates for different
chunks. On input, the silence parts are filled out, loops are rejected, and sample data with a new
sample rate is rejected. Silence with a different sample rate is generated appropriately. On output,
silence is not detected, nor are impossible sample rates. Note, this version now supports playing VOC
files with multiple blocks and supports playing files containing u-law and A-law samples.
vorbis See .ogg format.
.vox A headerless file of Dialogic/OKI ADPCM audio data commonly comes with the extension .vox. This ADPCM
data has 12-bit precision packed into only 4-bits.
.wav Microsoft .WAV RIFF files.
These appear to be very similar to IFF files, but not the same. They are the native sound file format
of Windows. (Obviously, Windows was of such incredible importance to the computer industry that it
just had to have its own sound file format.)
Normally .wav files have all formatting information in their headers, and so do not need any format
options specified for an input file. If any are, they will override the file header, and you will be
warned to this effect. You had better know what you are doing! Output format options will cause a for-
mat conversion, and the .wav will written appropriately.
SoX currently can read PCM, ULAW, ALAW, MS ADPCM, and IMA (or DVI) ADPCM. It can write all of these
formats including the ADPCM encoding. Big endian versions of RIFF files, called RIFX, can also be read
and written. To write a RIFX file, use the -x option with the output file options.
.wve Psion 8-bit A-law
These are 8-bit A-law 8khz sound files used on the Psion palmtop portable computer.
.raw Raw files (no header).
The sample rate, size (byte, word, etc), and encoding (signed, unsigned, etc.) of the sample file must
be given. The number of channels defaults to 1.
.ub, .sb, .uw, .sw, .ul, .al, .lu, .la, .sl
These are several suffices which serve as a shorthand for raw files with a given size and encoding.
Thus, ub, sb, uw, sw, ul, al, lu, la and sl correspond to "unsigned byte", "signed byte", "unsigned
word", "signed word", "u-law" (byte), "A-law" (byte), inverse bit order "u-law", inverse bit order "A-
law", and "signed long". The sample rate defaults to 8000 hz if not explicitly set, and the number of
channels defaults to 1. There are lots of Sparc samples floating around in u-law format with no header
and fixed at a sample rate of 8000 hz. (Certain sound management software cheerfully ignores the head-
ers.) Similarly, most Mac sound files are in unsigned byte format with a sample rate of 11025 or 22050
hz.
.auto This is a ``meta-type and is the default file type if the user does not specify one. This file type
attempts to guess the real type by looking for magic words in the header. If the type can't be guessed,
the program exits with an error message. The input must be a plain file, not a pipe. This type can't
be used for output files.
EFFECTS
Multiple effects may be applied to the audio data by specifying them one after another at the end of the command
line.
avg [ -l | -r | -f | -b | -1 | -2 | -3 | -4 | n,n,...,n ]
Reduce the number of channels by averaging the samples, or duplicate channels to increase the number of
channels. This effect is automatically used when the number of input channels differ from the number
of output channels. When reducing the number of channels it is possible to manually specify the avg
effect and use the -l, -r, -f, -b, -1, -2, -3, -4, options to select only the left, right, front, back
channel(s) or specific channel for the output instead of averaging the channels. The -l, and -r
options will do averaging in quad-channel files so select the exact channel to prevent this.
The avg effect can also be invoked with up to 16 double-precision numbers, seperated by commas, which
specify the proportion (0.0 = 0% and 1.0 = 100%) of each input channel that is to be mixed into each
output channel. In two-channel mode, 4 numbers are given: l->l, l->r, r->l, and r->r, respectively.
In four-channel mode, the first 4 numbers give the proportions for the left-front output channel, as
follows: lf->lf, rf->lf, lb->lf, and rb->rf. The next 4 give the right-front output in the same order,
then left-back and right-back.
It is also possible to use the 16 numbers to expand or reduce the channel count; just specify 0 for
unused channels.
Finally, certain reduced combination of numbers can be specified for certain input/output channel com-
binations.
In Ch Out Ch Num Mappings
_____ ______ ___ _____________________________
2 1 2 l->l, r->l
2 2 1 adjust balance
4 1 4 lf->l, rf->l, lb->l, rb-l
4 2 2 lf->l&rf->r, lb->l&rb->r
4 4 1 adjust balance
4 4 2 front balance, back balance
band [ -n ] center [ width ]
Apply a band-pass filter. The frequency response drops logarithmically around the center frequency.
The width gives the slope of the drop. The frequencies at center + width and center - width will be
half of their original amplitudes. Band defaults to a mode oriented to pitched signals, i.e. voice,
singing, or instrumental music. The -n (for noise) option uses the alternate mode for un-pitched sig-
nals. Warning: -n introduces a power-gain of about 11dB in the filter, so beware of output clipping.
Band introduces noise in the shape of the filter, i.e. peaking at the center frequency and settling
around it. See filter for a bandpass effect with steeper shoulders.
bandpass frequency bandwidth
Butterworth bandpass filter. Description coming soon!
bandreject frequency bandwidth
Butterworth bandreject filter. Description coming soon!
chorus gain-in gain-out delay decay speed depth
-s | -t [ delay decay speed depth -s | -t ... ]
Add a chorus to a sound sample. Each quadtuple delay/decay/speed/depth gives the delay in milliseconds
and the decay (relative to gain-in) with a modulation speed in Hz using depth in milliseconds. The
modulation is either sinusoidal (-s) or triangular (-t). Gain-out is the volume of the output.
compand attack1,decay1[,attack2,decay2...]
in-dB1,out-dB1[,in-dB2,out-dB2...]
[gain [initial-volume [delay ] ] ]
Compand (compress or expand) the dynamic range of a sample. The attack and decay time specify the
integration time over which the absolute value of the input signal is integrated to determine its vol-
ume; attacks refer to increases in volume and decays refer to decreases. Where more than one pair of
attack/decay parameters are specified, each channel is treated separately and the number of pairs must
agree with the number of input channels. The second parameter is a list of points on the compander's
transfer function specified in dB relative to the maximum possible signal amplitude. The input values
must be in a strictly increasing order but the transfer function does not have to be monotonically ris-
ing. The special value -inf may be used to indicate that the input volume should be associated output
volume. The points -inf,-inf and 0,0 are assumed; the latter may be overridden, but the former may
not.
The third (optional) parameter is a post-processing gain in dB which is applied after the compression
has taken place; the fourth (optional) parameter is an initial volume to be assumed for each channel
when the effect starts. This permits the user to supply a nominal level initially, so that, for exam-
ple, a very large gain is not applied to initial signal levels before the companding action has begun
to operate: it is quite probable that in such an event, the output would be severely clipped while the
compander gain properly adjusts itself.
The fifth (optional) parameter is a delay in seconds. The input signal is analyzed immediately to con-
trol the compander, but it is delayed before being fed to the volume adjuster. Specifying a delay
approximately equal to the attack/decay times allows the compander to effectively operate in a "predic-
tive" rather than a reactive mode.
copy Copy the input file to the output file. This is the default effect if both files have the same sam-
pling rate.
dcshift shift [ limitergain ]
DC Shift the audio data, with basic linear amplitude formula. This is most useful if your audio data
tends to not be centered around a value of 0. Shifting it back will allow you to get the most volume
adjustments without clipping audio data.
The first option is the dcshift value. It is a floating point number that indicates the amount to
shift.
An option limtergain value can be specified as well. It should have a value much less then 1.0 and is
used only on peaks to prevent clipping.
deemph Apply a treble attenuation shelving filter to samples in audio cd format. The frequency response of
pre-emphasized recordings is rectified. The filtering is defined in the standard document ISO 908.
earwax Makes sound easier to listen to on headphones. Adds audio-cues to samples in audio cd format so that
when listened to on headphones the stereo image is moved from inside your head (standard for head-
phones) to outside and in front of the listener (standard for speakers). See
www.geocities.com/beinges for a full explanation.
echo gain-in gain-out delay decay [ delay decay ... ]
Add echoing to a sound sample. Each delay/decay part gives the delay in milliseconds and the decay
(relative to gain-in) of that echo. Gain-out is the volume of the output.
echos gain-in gain-out delay decay [ delay decay ... ]
Add a sequence of echos to a sound sample. Each delay/decay part gives the delay in milliseconds and
the decay (relative to gain-in) of that echo. Gain-out is the volume of the output.
fade [ type ] fade-in-length
[ stop-time [ fade-out-length ] ]
Add a fade effect to the beginning, end, or both of the audio data.
For fade-ins, this starts from the first sample and ramps the volume of the audio from 0 to full volume
over fade-in-length seconds. Specify 0 seconds if no fade-in is wanted.
For fade-outs, the audio data will be truncated at the stop-time and the volume will be ramped from
full volume down to 0 starting at fade-out-length seconds before the stop-time. If fade-out-length is
not specified, it defaults to the same value as fade-in-length. No fade-out is performed if the stop-
time is not specified.
All times can be specified in either periods of time or sample counts. To specify time periods use the
format hh:mm:ss.frac format. To specify using sample counts, specify the number of samples and append
the letter 's' to the sample count (for example 8000s).
An optional type can be specified to change the type of envelope. Choices are q for quarter of a
sinewave, h for half a sinewave, t for linear slope, l for logarithmic, and p for inverted parabola.
The default is a linear slope.
filter [ low ]-[ high ] [ window-len [ beta ] ]
Apply a Sinc-windowed lowpass, highpass, or bandpass filter of given window length to the signal. low
refers to the frequency of the lower 6dB corner of the filter. high refers to the frequency of the
upper 6dB corner of the filter.
A lowpass filter is obtained by leaving low unspecified, or 0. A highpass filter is obtained by leav-
ing high unspecified, or 0, or greater than or equal to the Nyquist frequency.
The window-len, if unspecified, defaults to 128. Longer windows give a sharper cutoff, smaller windows
a more gradual cutoff.
The beta, if unspecified, defaults to 16. This selects a Kaiser window. You can select a Nuttall win-
dow by specifying anything <= 2.0 here. For more discussion of beta, look under the resample effect.
flanger gain-in gain-out delay decay speed < -s | -t >
Add a flanger to a sound sample. Each triple delay/decay/speed gives the delay in milliseconds and the
decay (relative to gain-in) with a modulation speed in Hz. The modulation is either sinodial (-s) or
triangular (-t). Gain-out is the volume of the output.
highp frequency
Apply a single pole recursive high-pass filter. The frequency response drops logarithmically with I
frequency in the middle of the drop. The slope of the filter is quite gentle. See filter for a high-
pass effect with sharper cutoff.
highpass frequency
Butterworth highpass filter. Description coming soon!
lowp frequency
Apply a single pole recursive low-pass filter. The frequency response drops logarithmically with fre-
quency in the middle of the drop. The slope of the filter is quite gentle. See filter for a lowpass
effect with sharper cutoff.
lowpass frequency
Butterworth lowpass filter. Description coming soon!
mask Add "masking noise" to signal. This effect deliberately adds white noise to a sound in order to mask
quantization effects, created by the process of playing a sound digitally. It tends to mask buzzing
voices, for example. It adds 1/2 bit of noise to the sound file at the output bit depth.
mcompand "attack1,decay1[,attack2,decay2...]
in-dB1,out-dB1[,in-dB2,out-dB2...]
[gain [initial-volume [delay ] ] ]" xover_freq
Multi-band compander is similar to the single band compander but the audio file is first divided up
into bands and then the compander is ran on each band. See the compand effect for definition of its
options. Compand options are specified between double quotes and the crossover frequency for that band
is specefied seperately with xover_fre. This can be repeated multiple times to create multiple bands.
noiseprof [profile-file]
noisered profile-file [threshold]
Noise reduction filter with profiling. This filter is moderately effective at removing consistent back-
ground noise such as hiss or hum. To use it, first run the noiseprof effect on a section of silence
(that is, a section which contains nothing but noise). The noiseprof effect will print a noise profile
to profile-file, or to stdout if no profile-file is specified. If there is sound output on stdout then
the profile will instead be directed to stderr.
To actually remove the noise, run SoX again with the noisered filter. The filter needs one argument,
profile-file, which contains the noise profile from noiseprof. thershold specifies how much noise
should be removed, and may be between 0 and 1 with a default of 0.5. Higher values will remove more
noise but present a greater possibility of distorting the desired audio signal. Experiment with dif-
ferent threshold values to find the optimal one for your sample.
pan direction
Pan the sound of an audio file from one channel to another. This is done by changing the volume of the
input channels so that it fades out on one channel and fades-in on another. If the number of input
channels is different then the number of output channels then this effect tries to intelligently handle
this. For instance, if the input contains 1 channel and the output contains 2 channels, then it will
create the missing channel itself. The direction is a value from -1.0 to 1.0. -1.0 represents far
left and 1.0 represents far right. Numbers in between will start the pan effect without totally muting
the opposite channel.
phaser gain-in gain-out delay decay speed < -s | -t >
Add a phaser to a sound sample. Each triple delay/decay/speed gives the delay in milliseconds and the
decay (relative to gain-in) with a modulation speed in Hz. The modulation is either sinodial (-s) or
triangular (-t). The decay should be less than 0.5 to avoid feedback. Gain-out is the volume of the
output.
pick [ -1 | -2 | -3 | -4 | -l | -r | -f | -b ]
Pick a subset of channels to be copied into the output file. This effect is just an alias of the "avg"
effect but is left here for historical reasons.
pitch shift [ width interpole fade ]
Change the pitch of file without affecting its duration by cross-fading shifted samples. shift is
given in cents. Use a positive value to shift to treble, negative value to shift to bass. Default
shift is 0. width of window is in ms. Default width is 20ms. Try 30ms to lower pitch, and 10ms to
raise pitch. interpole option, can be "cubic" or "linear". Default is "cubic". The fade option, can
be "cos", "hamming", "linear" or "trapezoid". Default is "cos".
polyphase [ -w < nut / ham > ]
[ -width < long / short / # > ]
[ -cutoff # ]
Translate input sampling rate to output sampling rate via polyphase interpolation, a DSP algorithm.
This method is slow and uses lots of RAM, but gives much better results than rate.
-w < nut / ham > : select either a Nuttal (~90 dB stopband) or Hamming (~43 dB stopband) window.
Default is nut.
-width long / short / # : specify the (approximate) width of the filter. long is 1024 samples; short
is 128 samples. Alternatively, an exact number can be used. Default is long. The short option is not
recommended, as it produces poor quality results.
-cutoff # : specify the filter cutoff frequency in terms of fraction of frequency bandwidth, also know
as the Nyquist frequency. Please see the resample effect for further information on Nyquist frequency.
If upsampling, then this is the fraction of the original signal that should go through. If downsam-
pling, this is the fraction of the signal left after downsampling. Default is 0.95. Remember that
this is a float.
rate Translate input sampling rate to output sampling rate via linear interpolation to the Least Common Mul-
tiple of the two sampling rates. This is the default effect if the two files have different sampling
rates and the preview options was specified. This is fast but noisy: the spectrum of the original
sound will be shifted upwards and duplicated faintly when up-translating by a multiple.
Lerp-ing is acceptable for cheap 8-bit sound hardware, but for CD-quality sound you should instead use
either resample or polyphase. If you are wondering which rate changing effects to use, you will want
to read a detailed analysis of all of them at http://leute.server.de/wilde/resample.html
repeat count
Repeats the audio data count times. Requires disk space to store the data to be repeated.
resample [ -qs | -q | -ql ] [ rolloff [ beta ] ]
Translate input sampling rate to output sampling rate via simulated analog filtration. This method is
slower than rate, but gives much better results.
By default, linear interpolation is used, with a window width about 45 samples at the lower of the two
rate. This gives an accuracy of about 16 bits, but insufficient stopband rejection in the case that
you want to have rolloff greater than about 0.80 of the Nyquist frequency.
The -q* options will change the default values for rolloff and beta as well as use quadratic interpola-
tion of filter coefficients, resulting in about 24 bits precision. The -qs, -q, or -ql options specify
increased accuracy at the cost of lower execution speed. It is optional to specify rolloff and beta
parameters when using the -q* options.
Following is a table of the reasonable defaults which are built-in to SoX:
Option Window rolloff beta interpolation
------ ------ ------- ---- -------------
(none) 45 0.80 16 linear
-qs 45 0.80 16 quadratic
-q 75 0.875 16 quadratic
-ql 149 0.94 16 quadratic
------ ------ ------- ---- -------------
-qs, -q, or -ql use window lengths of 45, 75, or 149 samples, respectively, at the lower sample-rate of
the two files. This means progressively sharper stop-band rejection, at proportionally slower execu-
tion times.
rolloff refers to the cut-off frequency of the low pass filter and is given in terms of the Nyquist
frequency for the lower sample rate. rolloff therefore should be something between 0.0 and 1.0, in
practice 0.8-0.95. The defaults are indicated above.
The Nyquist frequency is equal to (sample rate / 2). Logically, this is because the A/D converter
needs at least 2 samples to detect 1 cycle at the Nyquist frequency. Frequencies higher then the
Nyquist will actually appear as lower frequencies to the A/D converter and is called aliasing. Nor-
mally, A/D converts run the signal through a highpass filter first to avoid these problems.
Similar problems will happen in software when reducing the sample rate of an audio file (frequencies
above the new Nyquist frequency can be aliased to lower frequencies). Therefore, a good resample
effect will remove all frequency information above the new Nyquist frequency.
The rolloff refers to how close to the Nyquist frequency this cutoff is, with closer being better.
When increasing the sample rate of an audio file you would not expect to have any frequencies exist
that are past the original Nyquist frequency. Because of resampling properties, it is common to have
aliasing data created that is above the old Nyquist frequency. In that case the rolloff refers to how
close to the original Nyquist frequency to use a highpass filter to remove this false data, with closer
also being better.
The beta parameter determines the type of filter window used. Any value greater than 2.0 is the beta
for a Kaiser window. Beta <= 2.0 selects a Nuttall window. If unspecified, the default is a Kaiser
window with beta 16.
In the case of Kaiser window (beta > 2.0), lower betas produce a somewhat faster transition from pass-
band to stopband, at the cost of noticeable artifacts. A beta of 16 is the default, beta less than 10
is not recommended. If you want a sharper cutoff, don't use low beta's, use a longer sample window. A
Nuttall window is selected by specifying any 'beta' <= 2, and the Nuttall window has somewhat steeper
cutoff than the default Kaiser window. You will probably not need to use the beta parameter at all,
unless you are just curious about comparing the effects of Nuttall vs. Kaiser windows.
This is the default effect if the two files have different sampling rates. Default parameters are, as
indicated above, Kaiser window of length 45, rolloff 0.80, beta 16, linear interpolation.
NOTE: -qs is only slightly slower, but more accurate for 16-bit or higher precision.
NOTE: In many cases of up-sampling, no interpolation is needed, as exact filter coefficients can be
computed in a reasonable amount of space. To be precise, this is done when
input_rate < output_rate
&&
output_rate/gcd(input_rate,output_rate) <= 511
reverb gain-out reverbe-time delay [ delay ... ]
Add reverberation to a sound sample. Each delay is given in milliseconds and its feedback is depending
on the reverb-time in milliseconds. Each delay should be in the range of half to quarter of reverb-
time to get a realistic reverberation. Gain-out is the volume of the output.
reverse Reverse the sound sample completely. Included for finding Satanic subliminals.
silence above_periods [ duration threshold[ d | % ]
[ below_periods duration
threshold[ d | % ]]
Removes silence from the beginning, middle, or end of a sound file. Silence is anything below a speci-
fied threshold.
The above_periods value is used to indicate if sound should be trimmed at the beginning of the audio
file. A value of zero indicates no silence should be trimmed from the beginning. When specifing an
non-zero above_periods, it trims audio up until it finds non-silence. Normally, when trimming silence
from beginning of audio the above_periods will be 1 but it can be increased to higher values to trim
all data up to a specific count of non-silence periods. For example, if you had an audio file with two
songs that each contained 2 seconds of silence before the song, you could specify an above_period of 2
to strip out both silence periods and the first song.
When above_periods is non-zero, you must also specify a duration and threshold. Duration indications
the amount of time that non-silence must be detected before it stops trimming data. By increasing the
duration, burst of noise can be treated as silence and trimmed off.
Threshold is used to indicate what sample value you should treat as silence. For digital audio, a
value of 0 may be fine but for audio recorded from analog, you may wish to increase ths value to
account for background noise.
When optionally trimming silence from the end of a sound file, you specify a below_periods count. In
this case, below_period means to remove all audio data after silence is detected. Normally, this will
be a value 1 of but it can be increased to skip over periods of silence that are wanted. For example,
if you have a song with 2 seconds of silence in the middle and 2 second at the end, you could set
below_period to a value of 2 to skip over the silence in the middle of the audio file.
For below_periods, duration specifies a period of silence that must exist before data is not copied any
more. By specifying a higher duration, silence that is wanted can be left in the audio. For example,
if you have a song with an expected 1 second of silence in the middle and 2 seconds of silence at the
end, a duration of 2 seconds could be used to skip over the middle silence.
Unfortunetly, you must know the length of the silence at the end of your audio file to trim off silence
reliably. A work around is to use the silence effect in combination with the reverse effect. By first
reversing the audio, you can use the above_periods to reliably trim all audio from what looks like the
front of the file. Then reverse the file again to get back to normal.
To remove silence from the middle of a file, specify a below_periods that is negative. This value is
then treated as a positive value and is also used to indicate the effect should restart processing as
specified by the above_periods, making it suitable for removing periods of silence in the middle of the
sound file.
The period counts are in units of samples. Duration counts may be in the format of hh:mm:ss.frac, or
the exact count of samples. Threshold numbers may be suffixed iwth d, or % to indicate the value is in
decibels or a percentage of maximum value of the sample value (0% specifies pure digital silence).
speed [ -c ] factor
Speed up or down the sound, as a magnetic tape with a speed control. It affects both pitch and time. A
factor of 1.0 means no change, and is the default. 2.0 doubles speed, thus time length is cut by a
half and pitch is one octave higher. 0.5 halves speed thus time length doubles and pitch is one octave
lower. If the optional -c parameter is used then the factor is specified in "cents".
stat [ -s n ] [-rms ] [ -v ] [ -d ]
Do a statistical check on the input file, and print results on the standard error file. Audio data is
passed unmodified from input to output file unless used along with the -e option.
The "Volume Adjustment:" field in the statistics gives you the argument to the -v number which will
make the sample as loud as possible without clipping.
The option -v will print out the "Volume Adjustment:" field's value only and return. This could be of
use in scripts to auto convert the volume.
The -s n option is used to scale the input data by a given factor. The default value of n is the max
value of a signed long variable (0x7fffffff). Internal effects always work with signed long PCM data
and so the value should relate to this fact.
The -rms option will convert all output average values to root mean square format.
There is also an optional parameter -d that will print out a hex dump of the sound file from the inter-
nal buffer that is in 32-bit signed PCM data. This is mainly only of use in tracking down endian prob-
lems that creep in to SoX on cross-platform versions.
stretch factor [window fade shift fading]
Time stretch file by a given factor. Change duration without affecting the pitch. factor of stretch-
ing: >1.0 lengthen, <1.0 shorten duration. window size is in ms. Default is 20ms. The fade option, can
be "lin". shift ratio, in [0.0 1.0]. Default depends on stretch factor. 1.0 to shorten, 0.8 to
lengthen. The fading ratio, in [0.0 0.5]. The amount of a fade's default depends on factor and shift.
swap [ 1 2 | 1 2 3 4 ]
Swap channels in multi-channel sound files. Optionally, you may specify the channel order you would
like the output in. This defaults to output channel 2 and then 1 for stereo and 2, 1, 4, 3 for quad-
channels. An interesting feature is that you may duplicate a given channel by overwriting another.
This is done by repeating an output channel on the command line. For example, swap 2 2 will overwrite
channel 1 with channel 2's data; creating a stereo file with both channels containing the same audio
data.
synth [ length ] type mix [ freq [ -freq2 ]
[ off ] [ ph ] [ p1 ] [ p2 ] [ p3 ]
The synth effect will generate various types of audio data. Although this effect is used to generate
audio data, an input file must be specified. The length of the input audio file determines the length
of the output audio file.
<length> length in sec or hh:mm:ss.frac, 0=inputlength, default=0
<type> is sine, square, triangle, sawtooth, trapetz, exp, whitenoise, pinknoise, brownnoise,
default=sine
<mix> is create, mix, amod, default=create
<freq> frequency at beginning in Hz, not used for noise..
<freq2> frequency at end in Hz, not used for noise.. <freq/2> can be given as %%n, where 'n' is the
number of half notes in respect to A (440Hz)
<off> Bias (DC-offset) of signal in percent, default=0
<ph> phase shift 0..100 shift phase 0..2*Pi, not used for noise..
<p1> square: Ton/Toff, triangle+trapetz: rising slope time (0..100)
<p2> trapetz: ON time (0..100)
<p3> trapetz: falling slope position (0..100)
trim start [ length ]
Trim can trim off unwanted audio data from the beginning and end of the audio file. Audio samples are
not sent to the output stream until the start location is reached.
The optional length parameter tells the number of samples to output after the start sample and is used
to trim off the back side of the audio data. Using a value of 0 for the start parameter will allow
trimming off the back side only.
Both options can be specified using either an amount of time and an exact count of samples. The format
for specifying lengths in time is hh:mm:ss.frac. A start value of 1:30.5 will not start until 1
minute, thirty and 1/2 seconds into the audio data. The format for specifying sample counts is the
number of samples with the letter 's' appended to it. A value of 8000s will wait until 8000 samples
are read before starting to process audio data.
vibro speed [ depth ]
Add the world-famous Fender Vibro-Champ sound effect to a sound sample by using a sine wave as the vol-
ume knob. Speed gives the Hertz value of the wave. This must be under 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.
vol gain [ type [ limitergain ] ]
The vol effect is much like the command line option -v. It allows you to adjust the volume of an input
file and allows you to specify the adjustment in relation to amplitude, power, or dB. If type is not
specified then it defaults to amplitude.
When type is amplitude then a linear change of the amplitude is performed based on the gain. There-
fore, a value of 1.0 will keep the volume the same, 0.0 to < 1.0 will cause the volume to decrease and
values of > 1.0 will cause the volume to increase. Beware of clipping audio data when the gain is
greater then 1.0. A negative value performs the same adjustment while also changing the phase.
When type is power then a value of 1.0 also means no change in volume.
When type is dB the amplitude is changed logarithmically. 0.0 is constant while +6 doubles the ampli-
tude.
An optional limitergain value can be specified and should be a value much less then 1.0 (ie 0.05 or
0.02) and is used only on peaks to prevent clipping. Not specifying this parameter will cause no lim-
iter to be used. In verbose mode, this effect will display the percentage of audio data that needed to
be limited.
BUGS
The syntax is horrific. Thats the breaks when trying to handle all things from the command line.
Please report any bugs found in this version of SoX to Chris Bagwell (cbagwell@users.sourceforge.net)
FILES
RELATED
play(1), rec(1), soxexam(1)
NOTICES
The version of SoX that accompanies this manual page is support by Chris Bagwell (cbagwell@users.source-
forge.net). Please refer any questions regarding it to this address. You may obtain the latest version at the
the web site http://sox.sourceforge.net/
CATEGORY