1SoX(1) Sound eXchange SoX(1)
2
6 SoX - Sound eXchange, the Swiss Army knife of audio manipulation
7
9 sox [global-options] [format-options] infile1
10 [[format-options] infile2] ... [format-options] outfile
11 [effect [effect-options]] ...
12 play [global-options] [format-options] infile1
14 [[format-options] infile2] ... [format-options]
15 [effect [effect-options]] ...
16 rec [global-options] [format-options] outfile
18 [effect [effect-options]] ...
19
21 Introduction
22 SoX reads and writes audio files in most popular formats and can
23 optionally apply effects to them. It can combine multiple input
24 sources, synthesise audio, and, on many systems, act as a general pur‐
25 pose audio player or a multi-track audio recorder. It also has limited
26 ability to split the input into multiple output files.
27 All SoX functionality is available using just the sox command. To sim‐
29 plify playing and recording audio, if SoX is invoked as play, the out‐
30 put file is automatically set to be the default sound device, and if
31 invoked as rec, the default sound device is used as an input source.
32 Additionally, the soxi(1) command provides a convenient way to just
33 query audio file header information.
34 The heart of SoX is a library called libSoX. Those interested in
36 extending SoX or using it in other programs should refer to the libSoX
37 manual page: libsox(3).
38 SoX is a command-line audio processing tool, particularly suited to
40 making quick, simple edits and to batch processing. If you need an
41 interactive, graphical audio editor, use audacity(1).
42 * * *
44 The overall SoX processing chain can be summarised as follows:
46 Input(s) → Combiner → Effects → Output(s)
48 Note however, that on the SoX command line, the positions of the Out‐
50 put(s) and the Effects are swapped w.r.t. the logical flow just shown.
51 Note also that whilst options pertaining to files are placed before
52 their respective file name, the opposite is true for effects. To show
53 how this works in practice, here is a selection of examples of how SoX
54 might be used. The simple
55 sox recital.au recital.wav
56 translates an audio file in Sun AU format to a Microsoft WAV file,
57 whilst
58 sox recital.au -b 16 recital.wav channels 1 rate 16k fade 3 norm
59 performs the same format translation, but also applies four effects
60 (down-mix to one channel, sample rate change, fade-in, nomalize), and
61 stores the result at a bit-depth of 16.
62 sox -r 16k -e signed -b 8 -c 1 voice-memo.raw voice-memo.wav
63 converts `raw' (a.k.a. `headerless') audio to a self-describing file
64 format,
65 sox slow.aiff fixed.aiff speed 1.027
66 adjusts audio speed,
67 sox short.wav long.wav longer.wav
68 concatenates two audio files, and
69 sox -m music.mp3 voice.wav mixed.flac
70 mixes together two audio files.
71 play "The Moonbeams/Greatest/*.ogg" bass +3
72 plays a collection of audio files whilst applying a bass boosting
73 effect,
74 play -n -c1 synth sin %-12 sin %-9 sin %-5 sin %-2 fade h 0.1 1 0.1
75 plays a synthesised `A minor seventh' chord with a pipe-organ sound,
76 rec -c 2 radio.aiff trim 0 30:00
77 records half an hour of stereo audio, and
78 play -q take1.aiff & rec -M take1.aiff take1-dub.aiff
79 (with POSIX shell and where supported by hardware) records a new track
80 in a multi-track recording. Finally,
81 rec -r 44100 -b 16 -e signed-integer -p \
82 silence 1 0.50 0.1% 1 10:00 0.1% | \
83 sox -p song.ogg silence 1 0.50 0.1% 1 2.0 0.1% : \
84 newfile : restart
85 records a stream of audio such as LP/cassette and splits in to multiple
86 audio files at points with 2 seconds of silence. Also, it does not
87 start recording until it detects audio is playing and stops after it
88 sees 10 minutes of silence.
89 N.B. The above is just an overview of SoX's capabilities; detailed
91 explanations of how to use all SoX parameters, file formats, and
92 effects can be found below in this manual, in soxformat(7), and in
93 soxi(1).
94 File Format Types
96 SoX can work with `self-describing' and `raw' audio files. `self-
97 describing' formats (e.g. WAV, FLAC, MP3) have a header that completely
98 describes the signal and encoding attributes of the audio data that
99 follows. `raw' or `headerless' formats do not contain this information,
100 so the audio characteristics of these must be described on the SoX com‐
101 mand line or inferred from those of the input file.
102 The following four characteristics are used to describe the format of
104 audio data such that it can be processed with SoX:
105 sample rate
107 The sample rate in samples per second (`Hertz' or `Hz'). Digi‐
108 tal telephony traditionally uses a sample rate of 8000 Hz
109 (8 kHz), though these days, 16 and even 32 kHz are becoming more
110 common. Audio Compact Discs use 44100 Hz (44.1 kHz). Digital
111 Audio Tape and many computer systems use 48 kHz. Professional
112 audio systems often use 96 kHz.
113 sample size
115 The number of bits used to store each sample. Today, 16-bit is
116 commonly used. 8-bit was popular in the early days of computer
117 audio. 24-bit is used in the professional audio arena. Other
118 sizes are also used.
119 data encoding
121 The way in which each audio sample is represented (or
122 `encoded'). Some encodings have variants with different byte-
123 orderings or bit-orderings. Some compress the audio data so
124 that the stored audio data takes up less space (i.e. disk space
125 or transmission bandwidth) than the other format parameters and
126 the number of samples would imply. Commonly-used encoding types
127 include floating-point, μ-law, ADPCM, signed-integer PCM, MP3,
128 and FLAC.
129 channels
131 The number of audio channels contained in the file. One
132 (`mono') and two (`stereo') are widely used. `Surround sound'
133 audio typically contains six or more channels.
134 The term `bit-rate' is a measure of the amount of storage occupied by
136 an encoded audio signal over a unit of time. It can depend on all of
137 the above and is typically denoted as a number of kilo-bits per second
138 (kbps). An A-law telephony signal has a bit-rate of 64 kbps.
139 MP3-encoded stereo music typically has a bit-rate of 128-196 kbps.
140 FLAC-encoded stereo music typically has a bit-rate of 550-760 kbps.
141 Most self-describing formats also allow textual `comments' to be embed‐
143 ded in the file that can be used to describe the audio in some way,
144 e.g. for music, the title, the author, etc.
145 One important use of audio file comments is to convey `Replay Gain'
147 information. SoX supports applying Replay Gain information (for cer‐
148 tain input file formats only; currently, at least FLAC and Ogg Vorbis),
149 but not generating it. Note that by default, SoX copies input file
150 comments to output files that support comments, so output files may
151 contain Replay Gain information if some was present in the input file.
152 In this case, if anything other than a simple format conversion was
153 performed then the output file Replay Gain information is likely to be
154 incorrect and so should be recalculated using a tool that supports this
155 (not SoX).
156 The soxi(1) command can be used to display information from audio file
158 headers.
159 Determining & Setting The File Format
161 There are several mechanisms available for SoX to use to determine or
162 set the format characteristics of an audio file. Depending on the cir‐
163 cumstances, individual characteristics may be determined or set using
164 different mechanisms.
165 To determine the format of an input file, SoX will use, in order of
167 precedence and as given or available:
168 1. Command-line format options.
170 2. The contents of the file header.
172 3. The filename extension.
174 To set the output file format, SoX will use, in order of precedence and
176 as given or available:
177 1. Command-line format options.
179 2. The filename extension.
181 3. The input file format characteristics, or the closest that is sup‐
183 ported by the output file type.
184 For all files, SoX will exit with an error if the file type cannot be
186 determined. Command-line format options may need to be added or changed
187 to resolve the problem.
188 Playing & Recording Audio
190 The play and rec commands are provided so that basic playing and
191 recording is as simple as
192 play existing-file.wav
193 and
194 rec new-file.wav
195 These two commands are functionally equivalent to
196 sox existing-file.wav -d
197 and
198 sox -d new-file.wav
199 Of course, further options and effects (as described below) can be
200 added to the commands in either form.
201 * * *
203 Some systems provide more than one type of (SoX-compatible) audio
205 driver, e.g. ALSA & OSS, or SUNAU & AO. Systems can also have more
206 than one audio device (a.k.a. `sound card'). If more than one audio
207 driver has been built-in to SoX, and the default selected by SoX when
208 recording or playing is not the one that is wanted, then the AUDIO‐
209 DRIVER environment variable can be used to override the default. For
210 example (on many systems):
211 set AUDIODRIVER=oss
212 play ...
213 The AUDIODEV environment variable can be used to override the default
214 audio device, e.g.
215 set AUDIODEV=/dev/dsp2
216 play ...
217 sox ... -t oss
218 or
219 set AUDIODEV=hw:soundwave,1,2
220 play ...
221 sox ... -t alsa
222 Note that the way of setting environment variables varies from system
223 to system - for some specific examples, see `SOX_OPTS' below.
224 When playing a file with a sample rate that is not supported by the
226 audio output device, SoX will automatically invoke the rate effect to
227 perform the necessary sample rate conversion. For compatibility with
228 old hardware, the default rate quality level is set to `low'. This can
229 be changed by explicitly specifying the rate effect with a different
230 quality level, e.g.
231 play ... rate -m
232 or by using the --play-rate-arg option (see below).
233 * * *
235 On some systems, SoX allows audio playback volume to be adjusted whilst
237 using play. Where supported, this is achieved by tapping the `v' & `V'
238 keys during playback.
239 To help with setting a suitable recording level, SoX includes a peak-
241 level meter which can be invoked (before making the actual recording)
242 as follows:
243 rec -n
244 The recording level should be adjusted (using the system-provided mixer
245 program, not SoX) so that the meter is at most occasionally full scale,
246 and never `in the red' (an exclamation mark is shown). See also -S
247 below.
248 Accuracy
250 Many file formats that compress audio discard some of the audio signal
251 information whilst doing so. Converting to such a format and then con‐
252 verting back again will not produce an exact copy of the original
253 audio. This is the case for many formats used in telephony (e.g. A-
254 law, GSM) where low signal bandwidth is more important than high audio
255 fidelity, and for many formats used in portable music players (e.g.
256 MP3, Vorbis) where adequate fidelity can be retained even with the
257 large compression ratios that are needed to make portable players prac‐
258 tical.
259 Formats that discard audio signal information are called `lossy'. For‐
261 mats that do not are called `lossless'. The term `quality' is used as
262 a measure of how closely the original audio signal can be reproduced
263 when using a lossy format.
264 Audio file conversion with SoX is lossless when it can be, i.e. when
266 not using lossy compression, when not reducing the sampling rate or
267 number of channels, and when the number of bits used in the destination
268 format is not less than in the source format. E.g. converting from an
269 8-bit PCM format to a 16-bit PCM format is lossless but converting from
270 an 8-bit PCM format to (8-bit) A-law isn't.
271 N.B. SoX converts all audio files to an internal uncompressed format
273 before performing any audio processing. This means that manipulating a
274 file that is stored in a lossy format can cause further losses in audio
275 fidelity. E.g. with
276 sox long.mp3 short.mp3 trim 10
277 SoX first decompresses the input MP3 file, then applies the trim
278 effect, and finally creates the output MP3 file by re-compressing the
279 audio - with a possible reduction in fidelity above that which occurred
280 when the input file was created. Hence, if what is ultimately desired
281 is lossily compressed audio, it is highly recommended to perform all
282 audio processing using lossless file formats and then convert to the
283 lossy format only at the final stage.
284 N.B. Applying multiple effects with a single SoX invocation will, in
286 general, produce more accurate results than those produced using multi‐
287 ple SoX invocations.
288 Dithering
290 Dithering is a technique used to maximise the dynamic range of audio
291 stored at a particular bit-depth. Any distortion introduced by quanti‐
292 sation is decorrelated by adding a small amount of white noise to the
293 signal. In most cases, SoX can determine whether the selected process‐
294 ing requires dither and will add it during output formatting if appro‐
295 priate.
296 Specifically, by default, SoX automatically adds TPDF dither when the
298 output bit-depth is less than 24 and any of the following are true:
299 · bit-depth reduction has been specified explicitly using a command-
301 line option
302 · the output file format supports only bit-depths lower than that of
304 the input file format
305 · an effect has increased effective bit-depth within the internal
307 processing chain
308 For example, adjusting volume with vol 0.25 requires two additional
310 bits in which to losslessly store its results (since 0.25 decimal
311 equals 0.01 binary). So if the input file bit-depth is 16, then SoX's
312 internal representation will utilise 18 bits after processing this vol‐
313 ume change. In order to store the output at the same depth as the
314 input, dithering is used to remove the additional bits.
315 Use the -V option to see what processing SoX has automatically added.
317 The -D option may be given to override automatic dithering. To invoke
318 dithering manually (e.g. to select a noise-shaping curve), see the
319 dither effect.
320 Clipping
322 Clipping is distortion that occurs when an audio signal level (or `vol‐
323 ume') exceeds the range of the chosen representation. In most cases,
324 clipping is undesirable and so should be corrected by adjusting the
325 level prior to the point (in the processing chain) at which it occurs.
326 In SoX, clipping could occur, as you might expect, when using the vol
328 or gain effects to increase the audio volume. Clipping could also occur
329 with many other effects, when converting one format to another, and
330 even when simply playing the audio.
331 Playing an audio file often involves resampling, and processing by ana‐
333 logue components can introduce a small DC offset and/or amplification,
334 all of which can produce distortion if the audio signal level was ini‐
335 tially too close to the clipping point.
336 For these reasons, it is usual to make sure that an audio file's signal
338 level has some `headroom', i.e. it does not exceed a particular level
339 below the maximum possible level for the given representation. Some
340 standards bodies recommend as much as 9dB headroom, but in most cases,
341 3dB (≈ 70% linear) is enough. Note that this wisdom seems to have been
342 lost in modern music production; in fact, many CDs, MP3s, etc. are now
343 mastered at levels above 0dBFS i.e. the audio is clipped as delivered.
344 SoX's stat and stats effects can assist in determining the signal level
346 in an audio file. The gain or vol effect can be used to prevent clip‐
347 ping, e.g.
348 sox dull.wav bright.wav gain -6 treble +6
349 guarantees that the treble boost will not clip.
350 If clipping occurs at any point during processing, SoX will display a
352 warning message to that effect.
353 See also -G and the gain and norm effects.
355 Input File Combining
357 SoX's input combiner can be configured (see OPTIONS below) to combine
358 multiple files using any of the following methods: `concatenate',
359 `sequence', `mix', `mix-power', `merge', or `multiply'. The default
360 method is `sequence' for play, and `concatenate' for rec and sox.
361 For all methods other than `sequence', multiple input files must have
363 the same sampling rate. If necessary, separate SoX invocations can be
364 used to make sampling rate adjustments prior to combining.
365 If the `concatenate' combining method is selected (usually, this will
367 be by default) then the input files must also have the same number of
368 channels. The audio from each input will be concatenated in the order
369 given to form the output file.
370 The `sequence' combining method is selected automatically for play. It
372 is similar to `concatenate' in that the audio from each input file is
373 sent serially to the output file. However, here the output file may be
374 closed and reopened at the corresponding transition between input
375 files. This may be just what is needed when sending different types of
376 audio to an output device, but is not generally useful when the output
377 is a normal file.
378 If either the `mix' or `mix-power' combining method is selected then
380 two or more input files must be given and will be mixed together to
381 form the output file. The number of channels in each input file need
382 not be the same, but SoX will issue a warning if they are not and some
383 channels in the output file will not contain audio from every input
384 file. A mixed audio file cannot be un-mixed without reference to the
385 original input files.
386 If the `merge' combining method is selected then two or more input
388 files must be given and will be merged together to form the output
389 file. The number of channels in each input file need not be the same.
390 A merged audio file comprises all of the channels from all of the input
391 files. Un-merging is possible using multiple invocations of SoX with
392 the remix effect. For example, two mono files could be merged to form
393 one stereo file. The first and second mono files would become the left
394 and right channels of the stereo file.
395 The `multiply' combining method multiplies the sample values of corre‐
397 sponding channels (treated as numbers in the interval -1 to +1). If
398 the number of channels in the input files is not the same, the missing
399 channels are considered to contain all zero.
400 When combining input files, SoX applies any specified effects (includ‐
402 ing, for example, the vol volume adjustment effect) after the audio has
403 been combined. However, it is often useful to be able to set the volume
404 of (i.e. `balance') the inputs individually, before combining takes
405 place.
406 For all combining methods, input file volume adjustments can be made
408 manually using the -v option (below) which can be given for one or more
409 input files. If it is given for only some of the input files then the
410 others receive no volume adjustment. In some circumstances, automatic
411 volume adjustments may be applied (see below).
412 The -V option (below) can be used to show the input file volume adjust‐
414 ments that have been selected (either manually or automatically).
415 There are some special considerations that need to made when mixing
417 input files:
418 Unlike the other methods, `mix' combining has the potential to cause
420 clipping in the combiner if no balancing is performed. In this case,
421 if manual volume adjustments are not given, SoX will try to ensure that
422 clipping does not occur by automatically adjusting the volume (ampli‐
423 tude) of each input signal by a factor of ¹/n, where n is the number of
424 input files. If this results in audio that is too quiet or otherwise
425 unbalanced then the input file volumes can be set manually as described
426 above. Using the norm effect on the mix is another alternative.
427 If mixed audio seems loud enough at some points but too quiet in others
429 then dynamic range compression should be applied to correct this - see
430 the compand effect.
431 With the `mix-power' combine method, the mixed volume is approximately
433 equal to that of one of the input signals. This is achieved by balanc‐
434 ing using a factor of ¹/√n instead of ¹/n. Note that this balancing
435 factor does not guarantee that clipping will not occur, but the number
436 of clips will usually be low and the resultant distortion is generally
437 imperceptible.
438 Output Files
440 SoX's default behaviour is to take one or more input files and write
441 them to a single output file.
442 This behaviour can be changed by specifying the pseudo-effect `newfile'
444 within the effects list. SoX will then enter multiple output mode.
445 In multiple output mode, a new file is created when the effects prior
447 to the `newfile' indicate they are done. The effects chain listed
448 after `newfile' is then started up and its output is saved to the new
449 file.
450 In multiple output mode, a unique number will automatically be appended
452 to the end of all filenames. If the filename has an extension then the
453 number is inserted before the extension. This behaviour can be custom‐
454 ized by placing a %n anywhere in the filename where the number should
455 be substituted. An optional number can be placed after the % to indi‐
456 cate a minimum fixed width for the number.
457 Multiple output mode is not very useful unless an effect that will stop
459 the effects chain early is specified before the `newfile'. If end of
460 file is reached before the effects chain stops itself then no new file
461 will be created as it would be empty.
462 The following is an example of splitting the first 60 seconds of an
464 input file into two 30 second files and ignoring the rest.
465 sox song.wav ringtone%1n.wav trim 0 30 : newfile : trim 0 30
466 Stopping SoX
468 Usually SoX will complete its processing and exit automatically once it
469 has read all available audio data from the input files.
470 If desired, it can be terminated earlier by sending an interrupt signal
472 to the process (usually by pressing the keyboard interrupt key which is
473 normally Ctrl-C). This is a natural requirement in some circumstances,
474 e.g. when using SoX to make a recording. Note that when using SoX to
475 play multiple files, Ctrl-C behaves slightly differently: pressing it
476 once causes SoX to skip to the next file; pressing it twice in quick
477 succession causes SoX to exit.
478 Another option to stop processing early is to use an effect that has a
480 time period or sample count to determine the stopping point. The trim
481 effect is an example of this. Once all effects chains have stopped
482 then SoX will also stop.
483
485 Filenames can be simple file names, absolute or relative path names, or
486 URLs (input files only). Note that URL support requires that wget(1)
487 is available.
488 Note: Giving SoX an input or output filename that is the same as a SoX
490 effect-name will not work since SoX will treat it as an effect
491 specification. The only work-around to this is to avoid such
492 filenames. This is generally not difficult since most audio filenames
493 have a filename `extension', whilst effect-names do not.
494 Special Filenames
496 The following special filenames may be used in certain circumstances in
497 place of a normal filename on the command line:
498 - SoX can be used in simple pipeline operations by using the
500 special filename `-' which, if used as an input filename, will
501 cause SoX will read audio data from `standard input' (stdin),
502 and which, if used as the output filename, will cause SoX will
503 send audio data to `standard output' (stdout). Note that when
504 using this option for the output file, and sometimes when using
505 it for an input file, the file-type (see -t below) must also be
506 given.
507 "|program [options] ..."
509 This can be used in place of an input filename to specify the
510 the given program's standard output (stdout) be used as an input
511 file. Unlike - (above), this can be used for several inputs to
512 one SoX command. For example, if `genw' generates mono WAV
513 formatted signals to its standard output, then the following
514 command makes a stereo file from two generated signals:
515 sox -M "|genw --imd -" "|genw --thd -" out.wav
516 For headerless (raw) audio, -t (and perhaps other format
517 options) will need to be given, preceding the input command.
518 "wildcard-filename"
520 Specifies that filename `globbing' (wild-card matching) should
521 be performed by SoX instead of by the shell. This allows a sin‐
522 gle set of file options to be applied to a group of files. For
523 example, if the current directory contains three `vox' files,
524 file1.vox, file2.vox, and file3.vox, then
525 play --rate 6k *.vox
526 will be expanded by the `shell' (in most environments) to
527 play --rate 6k file1.vox file2.vox file3.vox
528 which will treat only the first vox file as having a sample rate
529 of 6k. With
530 play --rate 6k "*.vox"
531 the given sample rate option will be applied to all three vox
532 files.
533 -p, --sox-pipe
535 This can be used in place of an output filename to specify that
536 the SoX command should be used as in input pipe to another SoX
537 command. For example, the command:
538 play "|sox -n -p synth 2" "|sox -n -p synth 2 tremolo 10" stat
539 plays two `files' in succession, each with different effects.
540 -p is in fact an alias for `-t sox -'.
542 -d, --default-device
544 This can be used in place of an input or output filename to
545 specify that the default audio device (if one has been built
546 into SoX) is to be used. This is akin to invoking rec or play
547 (as described above).
548 -n, --null
550 This can be used in place of an input or output filename to
551 specify that a `null file' is to be used. Note that here, `null
552 file' refers to a SoX-specific mechanism and is not related to
553 any operating-system mechanism with a similar name.
554 Using a null file to input audio is equivalent to using a normal
556 audio file that contains an infinite amount of silence, and as
557 such is not generally useful unless used with an effect that
558 specifies a finite time length (such as trim or synth).
559 Using a null file to output audio amounts to discarding the
561 audio and is useful mainly with effects that produce information
562 about the audio instead of affecting it (such as noiseprof or
563 stat).
564 The sampling rate associated with a null file is by default
566 48 kHz, but, as with a normal file, this can be overridden if
567 desired using command-line format options (see below).
568 Supported File & Audio Device Types
570 See soxformat(7) for a list and description of the supported file for‐
571 mats and audio device drivers.
572
574 Global Options
575 These options can be specified on the command line at any point before
576 the first effect name.
577 The SOX_OPTS environment variable can be used to provide alternative
579 default values for SoX's global options. For example:
580 SOX_OPTS="--buffer 20000 --play-rate-arg -hs --temp /mnt/temp"
581 Note that setting SOX_OPTS can potentially create unwanted changes in
582 the behaviour of scripts or other programs that invoke SoX. SOX_OPTS
583 might best be used for things (such as in the given example) that
584 reflect the environment in which SoX is being run. Enabling options
585 such as --no-clobber as default might be handled better using a shell
586 alias since a shell alias will not affect operation in scripts etc.
587 One way to ensure that a script cannot be affected by SOX_OPTS is to
589 clear SOX_OPTS at the start of the script, but this of course loses the
590 benefit of SOX_OPTS carrying some system-wide default options. An
591 alternative approach is to explicitly invoke SoX with default option
592 values, e.g.
593 SOX_OPTS="-V --no-clobber"
594 ...
595 sox -V2 --clobber $input $output ...
596 Note that the way to set environment variables varies from system to
597 system. Here are some examples:
598 Unix bash:
600 export SOX_OPTS="-V --no-clobber"
601 Unix csh:
602 setenv SOX_OPTS "-V --no-clobber"
603 MS-DOS/MS-Windows:
604 set SOX_OPTS=-V --no-clobber
605 MS-Windows GUI: via Control Panel : System : Advanced : Environment
606 Variables
607 Mac OS X GUI: Refer to Apple's Technical Q&A QA1067 document.
609 --buffer BYTES, --input-buffer BYTES
611 Set the size in bytes of the buffers used for processing audio
612 (default 8192). --buffer applies to input, effects, and output
613 processing; --input-buffer applies only to input processing (for
614 which it overrides --buffer if both are given).
615 Be aware that large values for --buffer will cause SoX to be
617 become slow to respond to requests to terminate or to skip the
618 current input file.
619 --clobber
621 Don't prompt before overwriting an existing file with the same
622 name as that given for the output file. This is the default be‐
623 haviour.
624 --combine concatenate|merge|mix|mix-power|multiply|sequence
626 Select the input file combining method; for some of these, short
627 options are available: -m selects `mix', -M selects `merge', and
628 -T selects `multiply'.
629 See Input File Combining above for a description of the differ‐
631 ent combining methods.
632 -D, --no-dither
634 Disable automatic dither - see `Dithering' above. An example of
635 why this might occasionally be useful is if a file has been con‐
636 verted from 16 to 24 bit with the intention of doing some pro‐
637 cessing on it, but in fact no processing is needed after all and
638 the original 16 bit file has been lost, then, strictly speaking,
639 no dither is needed if converting the file back to 16 bit. See
640 also the stats effect for how to determine the actual bit depth
641 of the audio within a file.
642 --effects-file FILENAME
644 Use FILENAME to obtain all effects and their arguments. The
645 file is parsed as if the values were specified on the command
646 line. A new line can be used in place of the special : marker
647 to separate effect chains. For convenience, such markers at the
648 end of the file are normally ignored; if you want to specify an
649 empty last effects chain, use an explicit : by itself on the
650 last line of the file. This option causes any effects specified
651 on the command line to be discarded.
652 -G, --guard
654 Automatically invoke the gain effect to guard against clipping.
655 E.g.
656 sox -G infile -b 16 outfile rate 44100 dither -s
657 is shorthand for
658 sox infile -b 16 outfile gain -h rate 44100 gain -rh dither -s
659 See also -V, --norm, and the gain effect.
660 -h, --help
662 Show version number and usage information.
663 --help-effect NAME
665 Show usage information on the specified effect. The name all
666 can be used to show usage on all effects.
667 --help-format NAME
669 Show information about the specified file format. The name all
670 can be used to show information on all formats.
671 --i, --info
673 Only if given as the first parameter to sox, behave as soxi(1).
674 -m|-M Equivalent to --combine mix and --combine merge, respectively.
676 --magic
678 If SoX has been built with the optional `libmagic' library then
679 this option can be given to enable its use in helping to detect
680 audio file types.
681 --multi-threaded | --single-threaded
683 By default, SoX is `single threaded'. If the --multi-threaded
684 option is given however then SoX will process audio channels for
685 most multi-channel effects in parallel on hyper-threading/multi-
686 core architectures. This may reduce processing time, though
687 sometimes it may be necessary to use this option in conjunction
688 with a larger buffer size than is the default to gain any bene‐
689 fit from multi-threaded processing (e.g. 131072; see --buffer
690 above).
691 --no-clobber
693 Prompt before overwriting an existing file with the same name as
694 that given for the output file.
695 N.B. Unintentionally overwriting a file is easier than you
697 might think, for example, if you accidentally enter
698 sox file1 file2 effect1 effect2 ...
699 when what you really meant was
700 play file1 file2 effect1 effect2 ...
701 then, without this option, file2 will be overwritten. Hence,
702 using this option is recommended. SOX_OPTS (above), a `shell'
703 alias, script, or batch file may be an appropriate way of perma‐
704 nently enabling it.
705 --norm[=dB-level]
707 Automatically invoke the gain effect to guard against clipping
708 and to normalise the audio. E.g.
709 sox --norm infile -b 16 outfile rate 44100 dither -s
710 is shorthand for
711 sox infile -b 16 outfile gain -h rate 44100 gain -nh dither -s
712 Optionally, the audio can be normalized to a given level (usu‐
713 ally) below 0 dBFS:
714 sox --norm=-3 infile outfile
715 See also -V, -G, and the gain effect.
717 --play-rate-arg ARG
719 Selects a quality option to be used when the `rate' effect is
720 automatically invoked whilst playing audio. This option is typ‐
721 ically set via the SOX_OPTS environment variable (see above).
722 --plot gnuplot|octave|off
724 If not set to off (the default if --plot is not given), run in a
725 mode that can be used, in conjunction with the gnuplot program
726 or the GNU Octave program, to assist with the selection and con‐
727 figuration of many of the transfer-function based effects. For
728 the first given effect that supports the selected plotting pro‐
729 gram, SoX will output commands to plot the effect's transfer
730 function, and then exit without actually processing any audio.
731 E.g.
732 sox --plot octave input-file -n highpass 1320 > highpass.plt
733 octave highpass.plt
734 -q, --no-show-progress
736 Run in quiet mode when SoX wouldn't otherwise do so. This is
737 the opposite of the -S option.
738 -R Run in `repeatable' mode. When this option is given, where
740 applicable, SoX will embed a fixed time-stamp in the output file
741 (e.g. AIFF) and will `seed' pseudo random number generators
742 (e.g. dither) with a fixed number, thus ensuring that succes‐
743 sive SoX invocations with the same inputs and the same parame‐
744 ters yield the same output.
745 --replay-gain track|album|off
747 Select whether or not to apply replay-gain adjustment to input
748 files. The default is off for sox and rec, album for play where
749 (at least) the first two input files are tagged with the same
750 Artist and Album names, and track for play otherwise.
751 -S, --show-progress
753 Display input file format/header information, and processing
754 progress as input file(s) percentage complete, elapsed time, and
755 remaining time (if known; shown in brackets), and the number of
756 samples written to the output file. Also shown is a peak-level
757 meter, and an indication if clipping has occurred. The peak-
758 level meter shows up to two channels and is calibrated for digi‐
759 tal audio as follows (right channel shown):
760 dB FSD Display dB FSD Display
762 -25 - -11 ====
763 -23 = -9 ====-
764 -21 =- -7 =====
765 -19 == -5 =====-
766 -17 ==- -3 ======
767 -15 === -1 =====!
768 -13 ===-
769 A three-second peak-held value of headroom in dBs will be shown
771 to the right of the meter if this is below 6dB.
772 This option is enabled by default when using SoX to play or
774 record audio.
775 -T Equivalent to --combine multiply.
777 --temp DIRECTORY
779 Specify that any temporary files should be created in the given
780 DIRECTORY. This can be useful if there are permission or free-
781 space problems with the default location. In this case, using
782 `--temp .' (to use the current directory) is often a good solu‐
783 tion.
784 --version
786 Show SoX's version number and exit.
787 -V[level]
789 Set verbosity. This is particularly useful for seeing how any
790 automatic effects have been invoked by SoX.
791 SoX displays messages on the console (stderr) according to the
793 following verbosity levels:
794 0 No messages are shown at all; use the exit status to
796 determine if an error has occurred.
797 1 Only error messages are shown. These are generated if
799 SoX cannot complete the requested commands.
800 2 Warning messages are also shown. These are generated if
802 SoX can complete the requested commands, but not exactly
803 according to the requested command parameters, or if
804 clipping occurs.
805 3 Descriptions of SoX's processing phases are also shown.
807 Useful for seeing exactly how SoX is processing your
808 audio.
809 4 and above
811 Messages to help with debugging SoX are also shown.
812 By default, the verbosity level is set to 2 (shows errors and
814 warnings). Each occurrence of the -V option increases the ver‐
815 bosity level by 1. Alternatively, the verbosity level can be
816 set to an absolute number by specifying it immediately after the
817 -V, e.g. -V0 sets it to 0.
818 Input File Options
820 These options apply only to input files and may precede only input
821 filenames on the command line.
822 --ignore-length
824 Override an (incorrect) audio length given in an audio file's
825 header. If this option is given then SoX will keep reading audio
826 until it reaches the end of the input file.
827 -v, --volume FACTOR
829 Intended for use when combining multiple input files, this
830 option adjusts the volume of the file that follows it on the
831 command line by a factor of FACTOR. This allows it to be `bal‐
832 anced' w.r.t. the other input files. This is a linear (ampli‐
833 tude) adjustment, so a number less than 1 decreases the volume
834 and a number greater than 1 increases it. If a negative number
835 is given then in addition to the volume adjustment, the audio
836 signal will be inverted.
837 See also the norm, vol, and gain effects, and see Input File
839 Balancing above.
840 Input & Output File Format Options
842 These options apply to the input or output file whose name they immedi‐
843 ately precede on the command line and are used mainly when working with
844 headerless file formats or when specifying a format for the output file
845 that is different to that of the input file.
846 -b BITS, --bits BITS
848 The number of bits (a.k.a. bit-depth or sometimes word-length)
849 in each encoded sample. Not applicable to complex encodings
850 such as MP3 or GSM. Not necessary with encodings that have a
851 fixed number of bits, e.g. A/μ-law, ADPCM.
852 For an input file, the most common use for this option is to
854 inform SoX of the number of bits per sample in a `raw' (`header‐
855 less') audio file. For example
856 sox -r 16k -e signed -b 8 input.raw output.wav
857 converts a particular `raw' file to a self-describing `WAV'
858 file.
859 For an output file, this option can be used (perhaps along with
861 -e) to set the output encoding size. By default (i.e. if this
862 option is not given), the output encoding size will (providing
863 it is supported by the output file type) be set to the input
864 encoding size. For example
865 sox input.cdda -b 24 output.wav
866 converts raw CD digital audio (16-bit, signed-integer) to a
867 24-bit (signed-integer) `WAV' file.
868 -c CHANNELS, --channels CHANNELS
870 The number of audio channels in the audio file. This can be any
871 number greater than zero.
872 For an input file, the most common use for this option is to
874 inform SoX of the number of channels in a `raw' (`headerless')
875 audio file. Occasionally, it may be useful to use this option
876 with a `headered' file, in order to override the (presumably
877 incorrect) value in the header - note that this is only sup‐
878 ported with certain file types. Examples:
879 sox -r 48k -e float -b 32 -c 2 input.raw output.wav
880 converts a particular `raw' file to a self-describing `WAV'
881 file.
882 play -c 1 music.wav
883 interprets the file data as belonging to a single channel
884 regardless of what is indicated in the file header. Note that
885 if the file does in fact have two channels, this will result in
886 the file playing at half speed.
887 For an output file, this option provides a shorthand for speci‐
889 fying that the channels effect should be invoked in order to
890 change (if necessary) the number of channels in the audio signal
891 to the number given. For example, the following two commands
892 are equivalent:
893 sox input.wav -c 1 output.wav bass -b 24
894 sox input.wav output.wav bass -b 24 channels 1
895 though the second form is more flexible as it allows the effects
896 to be ordered arbitrarily.
897 -e ENCODING, --encoding ENCODING
899 The audio encoding type. Sometimes needed with file-types that
900 support more than one encoding type. For example, with raw, WAV,
901 or AU (but not, for example, with MP3 or FLAC). The available
902 encoding types are as follows:
903 signed-integer
905 PCM data stored as signed (`two's complement') integers.
906 Commonly used with a 16 or 24 -bit encoding size. A
907 value of 0 represents minimum signal power.
908 unsigned-integer
910 PCM data stored as unsigned integers. Commonly used with
911 an 8-bit encoding size. A value of 0 represents maximum
912 signal power.
913 floating-point
915 PCM data stored as IEEE 753 single precision (32-bit) or
916 double precision (64-bit) floating-point (`real') num‐
917 bers. A value of 0 represents minimum signal power.
918 a-law International telephony standard for logarithmic encoding
920 to 8 bits per sample. It has a precision equivalent to
921 roughly 13-bit PCM and is sometimes encoded with reversed
922 bit-ordering (see the -X option).
923 u-law, mu-law
925 North American telephony standard for logarithmic encod‐
926 ing to 8 bits per sample. A.k.a. μ-law. It has a preci‐
927 sion equivalent to roughly 14-bit PCM and is sometimes
928 encoded with reversed bit-ordering (see the -X option).
929 oki-adpcm
931 OKI (a.k.a. VOX, Dialogic, or Intel) 4-bit ADPCM; it has
932 a precision equivalent to roughly 12-bit PCM. ADPCM is a
933 form of audio compression that has a good compromise
934 between audio quality and encoding/decoding speed.
935 ima-adpcm
937 IMA (a.k.a. DVI) 4-bit ADPCM; it has a precision equiva‐
938 lent to roughly 13-bit PCM.
939 ms-adpcm
941 Microsoft 4-bit ADPCM; it has a precision equivalent to
942 roughly 14-bit PCM.
943 gsm-full-rate
945 GSM is currently used for the vast majority of the
946 world's digital wireless telephone calls. It utilises
947 several audio formats with different bit-rates and asso‐
948 ciated speech quality. SoX has support for GSM's origi‐
949 nal 13kbps `Full Rate' audio format. It is usually CPU-
950 intensive to work with GSM audio.
951 Encoding names can be abbreviated where this would not be
953 ambiguous; e.g. `unsigned-integer' can be given as `un', but not
954 `u' (ambiguous with `u-law').
955 For an input file, the most common use for this option is to
957 inform SoX of the encoding of a `raw' (`headerless') audio file
958 (see the examples in -b and -c above).
959 For an output file, this option can be used (perhaps along with
961 -b) to set the output encoding type For example
962 sox input.cdda -e float output1.wav
963 sox input.cdda -b 64 -e float output2.wav
965 convert raw CD digital audio (16-bit, signed-integer) to float‐
966 ing-point `WAV' files (single & double precision respectively).
967 By default (i.e. if this option is not given), the output encod‐
969 ing type will (providing it is supported by the output file
970 type) be set to the input encoding type.
971 --no-glob
973 Specifies that filename `globbing' (wild-card matching) should
974 not be performed by SoX on the following filename. For example,
975 if the current directory contains the two files `five-sec‐
976 onds.wav' and `five*.wav', then
977 play --no-glob "five*.wav"
978 can be used to play just the single file `five*.wav'.
979 -r, --rate RATE[k]
981 Gives the sample rate in Hz (or kHz if appended with `k') of the
982 file.
983 For an input file, the most common use for this option is to
985 inform SoX of the sample rate of a `raw' (`headerless') audio
986 file (see the examples in -b and -c above). Occasionally it may
987 be useful to use this option with a `headered' file, in order to
988 override the (presumably incorrect) value in the header - note
989 that this is only supported with certain file types. For exam‐
990 ple, if audio was recorded with a sample-rate of say 48k from a
991 source that played back a little, say 1.5%, too slowly, then
992 sox -r 48720 input.wav output.wav
993 effectively corrects the speed by changing only the file header
994 (but see also the speed effect for the more usual solution to
995 this problem).
996 For an output file, this option provides a shorthand for speci‐
998 fying that the rate effect should be invoked in order to change
999 (if necessary) the sample rate of the audio signal to the given
1000 value. For example, the following two commands are equivalent:
1001 sox input.wav -r 48k output.wav bass -b 24
1002 sox input.wav output.wav bass -b 24 rate 48k
1003 though the second form is more flexible as it allows rate
1004 options to be given, and allows the effects to be ordered arbi‐
1005 trarily.
1006 -t, --type FILE-TYPE
1008 Gives the type of the audio file. For both input and output
1009 files, this option is commonly used to inform SoX of the type a
1010 `headerless' audio file (e.g. raw, mp3) where the actual/desired
1011 type cannot be determined from a given filename extension. For
1012 example:
1013 another-command | sox -t mp3 - output.wav
1014 sox input.wav -t raw output.bin
1016 It can also be used to override the type implied by an input
1017 filename extension, but if overriding with a type that has a
1018 header, SoX will exit with an appropriate error message if such
1019 a header is not actually present.
1020 See soxformat(7) for a list of supported file types.
1022 -L, --endian little
1024 -B, --endian big
1025 -x, --endian swap
1026 These options specify whether the byte-order of the audio data
1027 is, respectively, `little endian', `big endian', or the opposite
1028 to that of the system on which SoX is being used. Endianness
1029 applies only to data encoded as floating-point, or as signed or
1030 unsigned integers of 16 or more bits. It is often necessary to
1031 specify one of these options for headerless files, and sometimes
1032 necessary for (otherwise) self-describing files. A given
1033 endian-setting option may be ignored for an input file whose
1034 header contains a specific endianness identifier, or for an out‐
1035 put file that is actually an audio device.
1036 N.B. Unlike other format characteristics, the endianness (byte,
1038 nibble, & bit ordering) of the input file is not automatically
1039 used for the output file; so, for example, when the following is
1040 run on a little-endian system:
1041 sox -B audio.s16 trimmed.s16 trim 2
1042 trimmed.s16 will be created as little-endian;
1043 sox -B audio.s16 -B trimmed.s16 trim 2
1044 must be used to preserve big-endianness in the output file.
1045 The -V option can be used to check the selected orderings.
1047 -N, --reverse-nibbles
1049 Specifies that the nibble ordering (i.e. the 2 halves of a byte)
1050 of the samples should be reversed; sometimes useful with ADPCM-
1051 based formats.
1052 N.B. See also N.B. in section on -x above.
1054 -X, --reverse-bits
1056 Specifies that the bit ordering of the samples should be
1057 reversed; sometimes useful with a few (mostly headerless) for‐
1058 mats.
1059 N.B. See also N.B. in section on -x above.
1061 Output File Format Options
1063 These options apply only to the output file and may precede only the
1064 output filename on the command line.
1065 --add-comment TEXT
1067 Append a comment in the output file header (where applicable).
1068 --comment TEXT
1070 Specify the comment text to store in the output file header
1071 (where applicable).
1072 SoX will provide a default comment if this option (or --com‐
1074 ment-file) is not given. To specify that no comment should be
1075 stored in the output file, use --comment "" .
1076 --comment-file FILENAME
1078 Specify a file containing the comment text to store in the out‐
1079 put file header (where applicable).
1080 -C, --compression FACTOR
1082 The compression factor for variably compressing output file for‐
1083 mats. If this option is not given then a default compression
1084 factor will apply. The compression factor is interpreted dif‐
1085 ferently for different compressing file formats. See the
1086 description of the file formats that use this option in soxfor‐
1087 mat(7) for more information.
1088
1090 In addition to converting, playing and recording audio files, SoX can
1091 be used to invoke a number of audio `effects'. Multiple effects may be
1092 applied by specifying them one after another at the end of the SoX com‐
1093 mand line, forming an `effects chain'. Note that applying multiple
1094 effects in real-time (i.e. when playing audio) is likely to require a
1095 high performance computer. Stopping other applications may alleviate
1096 performance issues should they occur.
1097 Some of the SoX effects are primarily intended to be applied to a sin‐
1099 gle instrument or `voice'. To facilitate this, the remix effect and
1100 the global SoX option -M can be used to isolate then recombine tracks
1101 from a multi-track recording.
1102 Multiple Effects Chains
1104 A single effects chain is made up of one or more effects. Audio from
1105 the input runs through the chain until either the end of the input file
1106 is reached or an effect in the chain requests to terminate the chain.
1107 SoX supports running multiple effects chains over the input audio. In
1109 this case, when one chain indicates it is done processing audio, the
1110 audio data is then sent through the next effects chain. This continues
1111 until either no more effects chains exist or the input has reached the
1112 end of the file.
1113 An effects chain is terminated by placing a : (colon) after an effect.
1115 Any following effects are a part of a new effects chain.
1116 It is important to place the effect that will stop the chain as the
1118 first effect in the chain. This is because any samples that are
1119 buffered by effects to the left of the terminating effect will be dis‐
1120 carded. The amount of samples discarded is related to the --buffer
1121 option and it should be kept small, relative to the sample rate, if the
1122 terminating effect cannot be first. Further information on stopping
1123 effects can be found in the Stopping SoX section.
1124 There are a few pseudo-effects that aid using multiple effects chains.
1126 These include newfile which will start writing to a new output file
1127 before moving to the next effects chain and restart which will move
1128 back to the first effects chain. Pseudo-effects must be specified as
1129 the first effect in a chain and as the only effect in a chain (they
1130 must have a : before and after they are specified).
1131 The following is an example of multiple effects chains. It will split
1133 the input file into multiple files of 30 seconds in length. Each out‐
1134 put filename will have unique number in its name as documented in the
1135 Output Files section.
1136 sox infile.wav output.wav trim 0 30 : newfile : restart
1137 Common Notation And Parameters
1139 In the descriptions that follow, brackets [ ] are used to denote param‐
1140 eters that are optional, braces { } to denote those that are both
1141 optional and repeatable, and angle brackets < > to denote those that
1142 are repeatable but not optional. Where applicable, default values for
1143 optional parameters are shown in parenthesis ( ).
1144 The following parameters are used with, and have the same meaning for,
1146 several effects:
1147 center[k]
1149 See frequency.
1150 frequency[k]
1152 A frequency in Hz, or, if appended with `k', kHz.
1153 gain A power gain in dB. Zero gives no gain; less than zero gives an
1155 attenuation.
1156 position
1158 A position within the audio stream; the syntax is [=|+|-]time‐
1159 spec, where timespec is a time specification (see below). The
1160 optional first character indicates whether the timespec is to be
1161 interpreted relative to the start (=) or end (-) of audio, or to
1162 the previous position if the effect accepts multiple position
1163 arguments (+). The audio length must be known for end-relative
1164 locations to work; some effects do accept -0 for end-of-audio,
1165 though, even if the length is unknown. Which of =, +, - is the
1166 default depends on the effect and is shown in its syntax as,
1167 e.g., position(+).
1168 Examples: =2:00 (two minutes into the audio stream), -100s (one
1170 hundred samples before the end of audio), +0:12+10s (twelve sec‐
1171 onds and ten samples after the previous position), -0.5+1s (one
1172 sample less than half a second before the end of audio).
1173 width[h|k|o|q]
1175 Used to specify the band-width of a filter. A number of differ‐
1176 ent methods to specify the width are available (though not all
1177 for every effect). One of the characters shown may be appended
1178 to select the desired method as follows:
1179 Method Notes
1181 h Hz
1182 k kHz
1183 o Octaves
1184 q Q-factor See [2]
1185 For each effect that uses this parameter, the default method
1187 (i.e. if no character is appended) is the one that it listed
1188 first in the first line of the effect's description.
1189 Most effects that expect an audio position or duration in a parameter,
1191 i.e. a time specification, accept either of the following two forms:
1192 [[hours:]minutes:]seconds[.frac][t]
1194 A specification of `1:30.5' corresponds to one minute, thirty
1195 and ½ seconds. The t suffix is entirely optional (however, see
1196 the silence effect for an exception). Note that the component
1197 values do not have to be normalized; e.g., `1:23:45', `83:45',
1198 `79:0285', `1:0:1425', `1::1425' and `5025' all are legal and
1199 equivalent to each other.
1200 sampless
1202 Specifies the number of samples directly, as in `8000s'. For
1203 large sample counts, e notation is supported: `1.7e6s' is the
1204 same as `1700000s'.
1205 Time specifications can also be chained with + or - into a new time
1207 specification where the right part is added to or subtracted from the
1208 left, respectively: `3:00-200s' means two hundred samples less than
1209 three minutes.
1210 To see if SoX has support for an optional effect, enter sox -h and look
1212 for its name under the list: `EFFECTS'.
1213 Supported Effects
1215 Note: a categorised list of the effects can be found in the accompany‐
1216 ing `README' file.
1217 allpass frequency[k] width[h|k|o|q]
1219 Apply a two-pole all-pass filter with central frequency (in Hz)
1220 frequency, and filter-width width. An all-pass filter changes
1221 the audio's frequency to phase relationship without changing its
1222 frequency to amplitude relationship. The filter is described in
1223 detail in [1].
1224 This effect supports the --plot global option.
1226 band [-n] center[k] [width[h|k|o|q]]
1228 Apply a band-pass filter. The frequency response drops loga‐
1229 rithmically around the center frequency. The width parameter
1230 gives the slope of the drop. The frequencies at center + width
1231 and center - width will be half of their original amplitudes.
1232 band defaults to a mode oriented to pitched audio, i.e. voice,
1233 singing, or instrumental music. The -n (for noise) option uses
1234 the alternate mode for un-pitched audio (e.g. percussion).
1235 Warning: -n introduces a power-gain of about 11dB in the filter,
1236 so beware of output clipping. band introduces noise in the
1237 shape of the filter, i.e. peaking at the center frequency and
1238 settling around it.
1239 This effect supports the --plot global option.
1241 See also sinc for a bandpass filter with steeper shoulders.
1243 bandpass|bandreject [-c] frequency[k] width[h|k|o|q]
1245 Apply a two-pole Butterworth band-pass or band-reject filter
1246 with central frequency frequency, and (3dB-point) band-width
1247 width. The -c option applies only to bandpass and selects a
1248 constant skirt gain (peak gain = Q) instead of the default: con‐
1249 stant 0dB peak gain. The filters roll off at 6dB per octave
1250 (20dB per decade) and are described in detail in [1].
1251 These effects support the --plot global option.
1253 See also sinc for a bandpass filter with steeper shoulders.
1255 bandreject frequency[k] width[h|k|o|q]
1257 Apply a band-reject filter. See the description of the bandpass
1258 effect for details.
1259 bass|treble gain [frequency[k] [width[s|h|k|o|q]]]
1261 Boost or cut the bass (lower) or treble (upper) frequencies of
1262 the audio using a two-pole shelving filter with a response simi‐
1263 lar to that of a standard hi-fi's tone-controls. This is also
1264 known as shelving equalisation (EQ).
1265 gain gives the gain at 0 Hz (for bass), or whichever is the
1267 lower of ∼22 kHz and the Nyquist frequency (for treble). Its
1268 useful range is about -20 (for a large cut) to +20 (for a large
1269 boost). Beware of Clipping when using a positive gain.
1270 If desired, the filter can be fine-tuned using the following
1272 optional parameters:
1273 frequency sets the filter's central frequency and so can be used
1275 to extend or reduce the frequency range to be boosted or cut.
1276 The default value is 100 Hz (for bass) or 3 kHz (for treble).
1277 width determines how steep is the filter's shelf transition. In
1279 addition to the common width specification methods described
1280 above, `slope' (the default, or if appended with `s') may be
1281 used. The useful range of `slope' is about 0.3, for a gentle
1282 slope, to 1 (the maximum), for a steep slope; the default value
1283 is 0.5.
1284 The filters are described in detail in [1].
1286 These effects support the --plot global option.
1288 See also equalizer for a peaking equalisation effect.
1290 bend [-f [22mframe-rate(25)] [-o [22mover-sample(16)] { start-posi‐
1292 tion(+),cents,end-position(+) }
1293 Changes pitch by specified amounts at specified times. Each
1294 given triple: start-position,cents,end-position specifies one
1295 bend. cents is the number of cents (100 cents = 1 semitone) by
1296 which to bend the pitch. The other values specify the points in
1297 time at which to start and end bending the pitch, respectively.
1298 The pitch-bending algorithm utilises the Discrete Fourier Trans‐
1300 form (DFT) at a particular frame rate and over-sampling rate.
1301 The -f and -o parameters may be used to adjust these parameters
1302 and thus control the smoothness of the changes in pitch.
1303 For example, an initial tone is generated, then bent three
1305 times, yielding four different notes in total:
1306 play -n synth 2.5 sin 667 gain 1 \
1307 bend .35,180,.25 .15,740,.53 0,-520,.3
1308 Here, the first bend runs from 0.35 to 0.6, and the second one
1309 from 0.75 to 1.28 seconds. Note that the clipping that is pro‐
1310 duced in this example is deliberate; to remove it, use gain -5
1311 in place of gain 1.
1312 See also pitch.
1314 biquad b0 b1 b2 a0 a1 a2
1316 Apply a biquad IIR filter with the given coefficients. Where b*
1317 and a* are the numerator and denominator coefficients respec‐
1318 tively.
1319 See http://en.wikipedia.org/wiki/Digital_biquad_filter (where a0
1321 = 1).
1322 This effect supports the --plot global option.
1324 channels CHANNELS
1326 Invoke a simple algorithm to change the number of channels in
1327 the audio signal to the given number CHANNELS: mixing if
1328 decreasing the number of channels or duplicating if increasing
1329 the number of channels.
1330 The channels effect is invoked automatically if SoX's -c option
1332 specifies a number of channels that is different to that of the
1333 input file(s). Alternatively, if this effect is given explic‐
1334 itly, then SoX's -c option need not be given. For example, the
1335 following two commands are equivalent:
1336 sox input.wav -c 1 output.wav bass -b 24
1337 sox input.wav output.wav bass -b 24 channels 1
1338 though the second form is more flexible as it allows the effects
1339 to be ordered arbitrarily.
1340 See also remix for an effect that allows channels to be
1342 mixed/selected arbitrarily.
1343 chorus gain-in gain-out <delay decay speed depth -s|-t>
1345 Add a chorus effect to the audio. This can make a single vocal
1346 sound like a chorus, but can also be applied to instrumentation.
1347 Chorus resembles an echo effect with a short delay, but whereas
1349 with echo the delay is constant, with chorus, it is varied using
1350 sinusoidal or triangular modulation. The modulation depth
1351 defines the range the modulated delay is played before or after
1352 the delay. Hence the delayed sound will sound slower or faster,
1353 that is the delayed sound tuned around the original one, like in
1354 a chorus where some vocals are slightly off key. See [3] for
1355 more discussion of the chorus effect.
1356 Each four-tuple parameter delay/decay/speed/depth gives the
1358 delay in milliseconds and the decay (relative to gain-in) with a
1359 modulation speed in Hz using depth in milliseconds. The modula‐
1360 tion is either sinusoidal (-s) or triangular (-t). Gain-out is
1361 the volume of the output.
1362 A typical delay is around 40ms to 60ms; the modulation speed is
1364 best near 0.25Hz and the modulation depth around 2ms. For exam‐
1365 ple, a single delay:
1366 play guitar1.wav chorus 0.7 0.9 55 0.4 0.25 2 -t
1367 Two delays of the original samples:
1368 play guitar1.wav chorus 0.6 0.9 50 0.4 0.25 2 -t \
1369 60 0.32 0.4 1.3 -s
1370 A fuller sounding chorus (with three additional delays):
1371 play guitar1.wav chorus 0.5 0.9 50 0.4 0.25 2 -t \
1372 60 0.32 0.4 2.3 -t 40 0.3 0.3 1.3 -s
1373 compand attack1,decay1{,attack2,decay2}
1375 [soft-knee-dB:]in-dB1[,out-dB1]{,in-dB2,out-dB2}
1376 [gain [initial-volume-dB [delay]]]
1377 Compand (compress or expand) the dynamic range of the audio.
1379 The attack and decay parameters (in seconds) determine the time
1381 over which the instantaneous level of the input signal is aver‐
1382 aged to determine its volume; attacks refer to increases in vol‐
1383 ume and decays refer to decreases. For most situations, the
1384 attack time (response to the music getting louder) should be
1385 shorter than the decay time because the human ear is more sensi‐
1386 tive to sudden loud music than sudden soft music. Where more
1387 than one pair of attack/decay parameters are specified, each
1388 input channel is companded separately and the number of pairs
1389 must agree with the number of input channels. Typical values
1390 are 0.3,0.8 seconds.
1391 The second parameter is a list of points on the compander's
1393 transfer function specified in dB relative to the maximum possi‐
1394 ble signal amplitude. The input values must be in a strictly
1395 increasing order but the transfer function does not have to be
1396 monotonically rising. If omitted, the value of out-dB1 defaults
1397 to the same value as in-dB1; levels below in-dB1 are not com‐
1398 panded (but may have gain applied to them). The point 0,0 is
1399 assumed but may be overridden (by 0,out-dBn). If the list is
1400 preceded by a soft-knee-dB value, then the points at where adja‐
1401 cent line segments on the transfer function meet will be rounded
1402 by the amount given. Typical values for the transfer function
1403 are 6:-70,-60,-20.
1404 The third (optional) parameter is an additional gain in dB to be
1406 applied at all points on the transfer function and allows easy
1407 adjustment of the overall gain.
1408 The fourth (optional) parameter is an initial level to be
1410 assumed for each channel when companding starts. This permits
1411 the user to supply a nominal level initially, so that, for exam‐
1412 ple, a very large gain is not applied to initial signal levels
1413 before the companding action has begun to operate: it is quite
1414 probable that in such an event, the output would be severely
1415 clipped while the compander gain properly adjusts itself. A
1416 typical value (for audio which is initially quiet) is -90 dB.
1417 The fifth (optional) parameter is a delay in seconds. The input
1419 signal is analysed immediately to control the compander, but it
1420 is delayed before being fed to the volume adjuster. Specifying
1421 a delay approximately equal to the attack/decay times allows the
1422 compander to effectively operate in a `predictive' rather than a
1423 reactive mode. A typical value is 0.2 seconds.
1424 * * *
1426 The following example might be used to make a piece of music
1428 with both quiet and loud passages suitable for listening to in a
1429 noisy environment such as a moving vehicle:
1430 sox asz.wav asz-car.wav compand 0.3,1 6:-70,-60,-20 -5 -90 0.2
1431 The transfer function (`6:-70,...') says that very soft sounds
1432 (below -70dB) will remain unchanged. This will stop the compan‐
1433 der from boosting the volume on `silent' passages such as
1434 between movements. However, sounds in the range -60dB to 0dB
1435 (maximum volume) will be boosted so that the 60dB dynamic range
1436 of the original music will be compressed 3-to-1 into a 20dB
1437 range, which is wide enough to enjoy the music but narrow enough
1438 to get around the road noise. The `6:' selects 6dB soft-knee
1439 companding. The -5 (dB) output gain is needed to avoid clipping
1440 (the number is inexact, and was derived by experimentation).
1441 The -90 (dB) for the initial volume will work fine for a clip
1442 that starts with near silence, and the delay of 0.2 (seconds)
1443 has the effect of causing the compander to react a bit more
1444 quickly to sudden volume changes.
1445 In the next example, compand is being used as a noise-gate for
1447 when the noise is at a lower level than the signal:
1448 play infile compand .1,.2 -inf,-50.1,-inf,-50,-50 0 -90 .1
1449 Here is another noise-gate, this time for when the noise is at a
1450 higher level than the signal (making it, in some ways, similar
1451 to squelch):
1452 play infile compand .1,.1 -45.1,-45,-inf,0,-inf 45 -90 .1
1453 This effect supports the --plot global option (for the transfer
1454 function).
1455 See also mcompand for a multiple-band companding effect.
1457 contrast [enhancement-amount(75)]
1459 Comparable with compression, this effect modifies an audio sig‐
1460 nal to make it sound louder. enhancement-amount controls the
1461 amount of the enhancement and is a number in the range 0-100.
1462 Note that enhancement-amount = 0 still gives a significant con‐
1463 trast enhancement.
1464 See also the compand and mcompand effects.
1466 dcshift shift [limitergain]
1468 Apply a DC shift to the audio. This can be useful to remove a
1469 DC offset (caused perhaps by a hardware problem in the recording
1470 chain) from the audio. The effect of a DC offset is reduced
1471 headroom and hence volume. The stat or stats effect can be used
1472 to determine if a signal has a DC offset.
1473 The given dcshift value is a floating point number in the range
1475 of ±2 that indicates the amount to shift the audio (which is in
1476 the range of ±1).
1477 An optional limitergain can be specified as well. It should
1479 have a value much less than 1 (e.g. 0.05 or 0.02) and is used
1480 only on peaks to prevent clipping.
1481 * * *
1483 An alternative approach to removing a DC offset (albeit with a
1485 short delay) is to use the highpass filter effect at a frequency
1486 of say 10Hz, as illustrated in the following example:
1487 sox -n dc.wav synth 5 sin %0 50
1488 sox dc.wav fixed.wav highpass 10
1489 deemph Apply Compact Disc (IEC 60908) de-emphasis (a treble attenuation
1491 shelving filter).
1492 Pre-emphasis was applied in the mastering of some CDs issued in
1494 the early 1980s. These included many classical music albums, as
1495 well as now sought-after issues of albums by The Beatles, Pink
1496 Floyd and others. Pre-emphasis should be removed at playback
1497 time by a de-emphasis filter in the playback device. However,
1498 not all modern CD players have this filter, and very few PC CD
1499 drives have it; playing pre-emphasised audio without the correct
1500 de-emphasis filter results in audio that sounds harsh and is far
1501 from what its creators intended.
1502 With the deemph effect, it is possible to apply the necessary
1504 de-emphasis to audio that has been extracted from a pre-empha‐
1505 sised CD, and then either burn the de-emphasised audio to a new
1506 CD (which will then play correctly on any CD player), or simply
1507 play the correctly de-emphasised audio files on the PC. For
1508 example:
1509 sox track1.wav track1-deemph.wav deemph
1510 and then burn track1-deemph.wav to CD, or
1511 play track1-deemph.wav
1512 or simply
1513 play track1.wav deemph
1514 The de-emphasis filter is implemented as a biquad and requires
1515 the input audio sample rate to be either 44.1kHz or 48kHz. Max‐
1516 imum deviation from the ideal response is only 0.06dB (up to
1517 20kHz).
1518 This effect supports the --plot global option.
1520 See also the bass and treble shelving equalisation effects.
1522 delay {position(=)}
1524 Delay one or more audio channels such that they start at the
1525 given position. For example, delay 1.5 +1 3000s delays the
1526 first channel by 1.5 seconds, the second channel by 2.5 seconds
1527 (one second more than the previous channel), the third channel
1528 by 3000 samples, and leaves any other channels that may be
1529 present un-delayed. The following (one long) command plays a
1530 chime sound:
1531 play -n synth -j 3 sin %3 sin %-2 sin %-5 sin %-9 \
1532 sin %-14 sin %-21 fade h .01 2 1.5 delay \
1533 1.3 1 .76 .54 .27 remix - fade h 0 2.7 2.5 norm -1
1534 and this plays a guitar chord:
1535 play -n synth pl G2 pl B2 pl D3 pl G3 pl D4 pl G4 \
1536 delay 0 .05 .1 .15 .2 .25 remix - fade 0 4 .1 norm -1
1537 dither [-S|-s|-f filter] [-a] [-p precision]
1539 Apply dithering to the audio. Dithering deliberately adds a
1540 small amount of noise to the signal in order to mask audible
1541 quantization effects that can occur if the output sample size is
1542 less than 24 bits. With no options, this effect will add trian‐
1543 gular (TPDF) white noise. Noise-shaping (only for certain sam‐
1544 ple rates) can be selected with -s. With the -f option, it is
1545 possible to select a particular noise-shaping filter from the
1546 following list: lipshitz, f-weighted, modified-e-weighted,
1547 improved-e-weighted, gesemann, shibata, low-shibata, high-shi‐
1548 bata. Note that most filter types are available only with
1549 44100Hz sample rate. The filter types are distinguished by the
1550 following properties: audibility of noise, level of (inaudible,
1551 but in some circumstances, otherwise problematic) shaped high
1552 frequency noise, and processing speed.
1553 See http://sox.sourceforge.net/SoX/NoiseShaping for graphs of
1554 the different noise-shaping curves.
1555 The -S option selects a slightly `sloped' TPDF, biased towards
1557 higher frequencies. It can be used at any sampling rate but
1558 below ≈22k, plain TPDF is probably better, and above ≈ 37k,
1559 noise-shaping (if available) is probably better.
1560 The -a option enables a mode where dithering (and noise-shaping
1562 if applicable) are automatically enabled only when needed. The
1563 most likely use for this is when applying fade in or out to an
1564 already dithered file, so that the redithering applies only to
1565 the faded portions. However, auto dithering is not fool-proof,
1566 so the fades should be carefully checked for any noise modula‐
1567 tion; if this occurs, then either re-dither the whole file, or
1568 use trim, fade, and concatencate.
1569 The -p option allows overriding the target precision.
1571 If the SoX global option -R option is not given, then the
1573 pseudo-random number generator used to generate the white noise
1574 will be `reseeded', i.e. the generated noise will be different
1575 between invocations.
1576 This effect should not be followed by any other effect that
1578 affects the audio.
1579 See also the `Dithering' section above.
1581 downsample [factor(2)]
1583 Downsample the signal by an integer factor: Only the first out
1584 of each factor samples is retained, the others are discarded.
1585 No decimation filter is applied. If the input is not a properly
1587 bandlimited baseband signal, aliasing will occur. This may be
1588 desirable, e.g., for frequency translation.
1589 For a general resampling effect with anti-aliasing, see rate.
1591 See also upsample.
1592 earwax Makes audio easier to listen to on headphones. Adds `cues' to
1594 44.1kHz stereo (i.e. audio CD format) audio so that when lis‐
1595 tened to on headphones the stereo image is moved from inside
1596 your head (standard for headphones) to outside and in front of
1597 the listener (standard for speakers).
1598 echo gain-in gain-out <delay decay>
1600 Add echoing to the audio. Echoes are reflected sound and can
1601 occur naturally amongst mountains (and sometimes large build‐
1602 ings) when talking or shouting; digital echo effects emulate
1603 this behaviour and are often used to help fill out the sound of
1604 a single instrument or vocal. The time difference between the
1605 original signal and the reflection is the `delay' (time), and
1606 the loudness of the reflected signal is the `decay'. Multiple
1607 echoes can have different delays and decays.
1608 Each given delay decay pair gives the delay in milliseconds and
1610 the decay (relative to gain-in) of that echo. Gain-out is the
1611 volume of the output. For example: This will make it sound as
1612 if there are twice as many instruments as are actually playing:
1613 play lead.aiff echo 0.8 0.88 60 0.4
1614 If the delay is very short, then it sound like a (metallic) ro‐
1615 bot playing music:
1616 play lead.aiff echo 0.8 0.88 6 0.4
1617 A longer delay will sound like an open air concert in the moun‐
1618 tains:
1619 play lead.aiff echo 0.8 0.9 1000 0.3
1620 One mountain more, and:
1621 play lead.aiff echo 0.8 0.9 1000 0.3 1800 0.25
1622 echos gain-in gain-out <delay decay>
1624 Add a sequence of echoes to the audio. Each delay decay pair
1625 gives the delay in milliseconds and the decay (relative to gain-
1626 in) of that echo. Gain-out is the volume of the output.
1627 Like the echo effect, echos stand for `ECHO in Sequel', that is
1629 the first echos takes the input, the second the input and the
1630 first echos, the third the input and the first and the second
1631 echos, ... and so on. Care should be taken using many echos; a
1632 single echos has the same effect as a single echo.
1633 The sample will be bounced twice in symmetric echos:
1635 play lead.aiff echos 0.8 0.7 700 0.25 700 0.3
1636 The sample will be bounced twice in asymmetric echos:
1637 play lead.aiff echos 0.8 0.7 700 0.25 900 0.3
1638 The sample will sound as if played in a garage:
1639 play lead.aiff echos 0.8 0.7 40 0.25 63 0.3
1640 equalizer frequency[k] width[q|o|h|k] gain
1642 Apply a two-pole peaking equalisation (EQ) filter. With this
1643 filter, the signal-level at and around a selected frequency can
1644 be increased or decreased, whilst (unlike band-pass and band-
1645 reject filters) that at all other frequencies is unchanged.
1646 frequency gives the filter's central frequency in Hz, width, the
1648 band-width, and gain the required gain or attenuation in dB.
1649 Beware of Clipping when using a positive gain.
1650 In order to produce complex equalisation curves, this effect can
1652 be given several times, each with a different central frequency.
1653 The filter is described in detail in [1].
1655 This effect supports the --plot global option.
1657 See also bass and treble for shelving equalisation effects.
1659 fade [type] fade-in-length [stop-position(=) [fade-out-length]]
1661 Apply a fade effect to the beginning, end, or both of the audio.
1662 An optional type can be specified to select the shape of the
1664 fade curve: q for quarter of a sine wave, h for half a sine
1665 wave, t for linear (`triangular') slope, l for logarithmic, and
1666 p for inverted parabola. The default is logarithmic.
1667 A fade-in starts from the first sample and ramps the signal
1669 level from 0 to full volume over the time given as fade-in-
1670 length. Specify 0 if no fade-in is wanted.
1671 For fade-outs, the audio will be truncated at stop-position and
1673 the signal level will be ramped from full volume down to 0 over
1674 an interval of fade-out-length before the stop-position. If
1675 fade-out-length is not specified, it defaults to the same value
1676 as fade-in-length. No fade-out is performed if stop-position is
1677 not specified. If the audio length can be determined from the
1678 input file header and any previous effects, then -0 (or, for
1679 historical reasons, 0) may be specified for stop-position to
1680 indicate the usual case of a fade-out that ends at the end of
1681 the input audio stream.
1682 Any time specification may be used for fade-in-length and fade-
1684 out-length.
1685 See also the splice effect.
1687 fir [coefs-file|coefs]
1689 Use SoX's FFT convolution engine with given FIR filter coeffi‐
1690 cients. If a single argument is given then this is treated as
1691 the name of a file containing the filter coefficients (white-
1692 space separated; may contain `#' comments). If the given file‐
1693 name is `-', or if no argument is given, then the coefficients
1694 are read from the `standard input' (stdin); otherwise, coeffi‐
1695 cients may be given on the command line. Examples:
1696 sox infile outfile fir 0.0195 -0.082 0.234 0.891 -0.145 0.043
1697 sox infile outfile fir coefs.txt
1698 with coefs.txt containing
1699 # HP filter
1700 # freq=10000
1701 1.2311233052619888e-01
1702 -4.4777096106211783e-01
1703 5.1031563346705155e-01
1704 -6.6502926320995331e-02
1705 ...
1706 This effect supports the --plot global option.
1708 flanger [delay depth regen width speed shape phase interp]
1710 Apply a flanging effect to the audio. See [3] for a detailed
1711 description of flanging.
1712 All parameters are optional (right to left).
1714 Range Default Description
1716 delay 0 - 30 0 Base delay in milliseconds.
1717 depth 0 - 10 2 Added swept delay in milliseconds.
1718 regen -95 - 95 0 Percentage regeneration (delayed
1719 signal feedback).
1720 width 0 - 100 71 Percentage of delayed signal mixed
1721 with original.
1722 speed 0.1 - 10 0.5 Sweeps per second (Hz).
1723 shape sin Swept wave shape: sine|triangle.
1724 phase 0 - 100 25 Swept wave percentage phase-shift
1725 for multi-channel (e.g. stereo)
1726 flange; 0 = 100 = same phase on
1727 each channel.
1728 interp lin Digital delay-line interpolation:
1729 linear|quadratic.
1730 gain [-e|-B|-b|-r] [-n] [-l|-h] [gain-dB]
1732 Apply amplification or attenuation to the audio signal, or, in
1733 some cases, to some of its channels. Note that use of any of
1734 -e, -B, -b, -r, or -n requires temporary file space to store the
1735 audio to be processed, so may be unsuitable for use with
1736 `streamed' audio.
1737 Without other options, gain-dB is used to adjust the signal
1739 power level by the given number of dB: positive amplifies
1740 (beware of Clipping), negative attenuates. With other options,
1741 the gain-dB amplification or attenuation is (logically) applied
1742 after the processing due to those options.
1743 Given the -e option, the levels of the audio channels of a
1745 multi-channel file are `equalised', i.e. gain is applied to all
1746 channels other than that with the highest peak level, such that
1747 all channels attain the same peak level (but, without also giv‐
1748 ing -n, the audio is not `normalised').
1749 The -B (balance) option is similar to -e, but with -B, the RMS
1751 level is used instead of the peak level. -B might be used to
1752 correct stereo imbalance caused by an imperfect record turntable
1753 cartridge. Note that unlike -e, -B might cause some clipping.
1754 -b is similar to -B but has clipping protection, i.e. if neces‐
1756 sary to prevent clipping whilst balancing, attenuation is
1757 applied to all channels. Note, however, that in conjunction
1758 with -n, -B and -b are synonymous.
1759 The -r option is used in conjunction with a prior invocation of
1761 gain with the -h option - see below for details.
1762 The -n option normalises the audio to 0dB FSD; it is often used
1764 in conjunction with a negative gain-dB to the effect that the
1765 audio is normalised to a given level below 0dB. For example,
1766 sox infile outfile gain -n
1767 normalises to 0dB, and
1768 sox infile outfile gain -n -3
1769 normalises to -3dB.
1770 The -l option invokes a simple limiter, e.g.
1772 sox infile outfile gain -l 6
1773 will apply 6dB of gain but never clip. Note that limiting more
1774 than a few dBs more than occasionally (in a piece of audio) is
1775 not recommended as it can cause audible distortion. See the
1776 compand effect for a more capable limiter.
1777 The -h option is used to apply gain to provide head-room for
1779 subsequent processing. For example, with
1780 sox infile outfile gain -h bass +6
1781 6dB of attenuation will be applied prior to the bass boosting
1782 effect thus ensuring that it will not clip. Of course, with
1783 bass, it is obvious how much headroom will be needed, but with
1784 other effects (e.g. rate, dither) it is not always as clear.
1785 Another advantage of using gain -h rather than an explicit
1786 attenuation, is that if the headroom is not used by subsequent
1787 effects, it can be reclaimed with gain -r, for example:
1788 sox infile outfile gain -h bass +6 rate 44100 gain -r
1789 The above effects chain guarantees never to clip nor amplify; it
1790 attenuates if necessary to prevent clipping, but by only as much
1791 as is needed to do so.
1792 Output formatting (dithering and bit-depth reduction) also
1794 requires headroom (which cannot be `reclaimed'), e.g.
1795 sox infile outfile gain -h bass +6 rate 44100 gain -rh dither
1796 Here, the second gain invocation, reclaims as much of the head‐
1797 room as it can from the preceding effects, but retains as much
1798 headroom as is needed for subsequent processing. The SoX global
1799 option -G can be given to automatically invoke gain -h and gain
1800 -r.
1801 See also the norm and vol effects.
1803 highpass|lowpass [-1|-2] frequency[k] [width[q|o|h|k]]
1805 Apply a high-pass or low-pass filter with 3dB point frequency.
1806 The filter can be either single-pole (with -1), or double-pole
1807 (the default, or with -2). width applies only to double-pole
1808 filters; the default is Q = 0.707 and gives a Butterworth
1809 response. The filters roll off at 6dB per pole per octave (20dB
1810 per pole per decade). The double-pole filters are described in
1811 detail in [1].
1812 These effects support the --plot global option.
1814 See also sinc for filters with a steeper roll-off.
1816 hilbert [-n taps]
1818 Apply an odd-tap Hilbert transform filter, phase-shifting the
1819 signal by 90 degrees.
1820 This is used in many matrix coding schemes and for analytic sig‐
1822 nal generation. The process is often written as a multiplica‐
1823 tion by i (or j), the imaginary unit.
1824 An odd-tap Hilbert transform filter has a bandpass characteris‐
1826 tic, attenuating the lowest and highest frequencies. Its band‐
1827 width can be controlled by the number of filter taps, which can
1828 be specified with -n. By default, the number of taps is chosen
1829 for a cutoff frequency of about 75 Hz.
1830 This effect supports the --plot global option.
1832 ladspa [-l|-r] module [plugin] [argument ...]
1834 Apply a LADSPA [5] (Linux Audio Developer's Simple Plugin API)
1835 plugin. Despite the name, LADSPA is not Linux-specific, and a
1836 wide range of effects is available as LADSPA plugins, such as
1837 cmt [6] (the Computer Music Toolkit) and Steve Harris's plugin
1838 collection [7]. The first argument is the plugin module, the
1839 second the name of the plugin (a module can contain more than
1840 one plugin), and any other arguments are for the control ports
1841 of the plugin. Missing arguments are supplied by default values
1842 if possible.
1843 Normally, the number of input ports of the plugin must match the
1845 number of input channels, and the number of output ports deter‐
1846 mines the output channel count. However, the -r (replicate)
1847 option allows cloning a mono plugin to handle multi-channel
1848 input.
1849 Some plugins introduce latency which SoX may optionally compen‐
1851 sate for. The -l (latency compensation) option automatically
1852 compensates for latency as reported by the plugin via an output
1853 control port named "latency".
1854 If found, the environment variable LADSPA_PATH will be used as
1856 search path for plugins.
1857 loudness [gain [reference]]
1859 Loudness control - similar to the gain effect, but provides
1860 equalisation for the human auditory system. See
1861 http://en.wikipedia.org/wiki/Loudness for a detailed description
1862 of loudness. The gain is adjusted by the given gain parameter
1863 (usually negative) and the signal equalised according to ISO 226
1864 w.r.t. a reference level of 65dB, though an alternative refer‐
1865 ence level may be given if the original audio has been equalised
1866 for some other optimal level. A default gain of -10dB is used
1867 if a gain value is not given.
1868 See also the gain effect.
1870 lowpass [-1|-2] frequency[k] [width[q|o|h|k]]
1872 Apply a low-pass filter. See the description of the highpass
1873 effect for details.
1874 mcompand "attack1,decay1{,attack2,decay2}
1876 [soft-knee-dB:]in-dB1[,out-dB1]{,in-dB2,out-dB2}
1877 [gain [initial-volume-dB [delay]]]" {crossover-freq[k]
1878 "attack1,..."}
1879 The multi-band compander is similar to the single-band compander
1881 but the audio is first divided into bands using Linkwitz-Riley
1882 cross-over filters and a separately specifiable compander run on
1883 each band. See the compand effect for the definition of its
1884 parameters. Compand parameters are specified between double
1885 quotes and the crossover frequency for that band is given by
1886 crossover-freq; these can be repeated to create multiple bands.
1887 For example, the following (one long) command shows how multi-
1889 band companding is typically used in FM radio:
1890 play track1.wav gain -3 sinc 8000- 29 100 mcompand \
1891 "0.005,0.1 -47,-40,-34,-34,-17,-33" 100 \
1892 "0.003,0.05 -47,-40,-34,-34,-17,-33" 400 \
1893 "0.000625,0.0125 -47,-40,-34,-34,-15,-33" 1600 \
1894 "0.0001,0.025 -47,-40,-34,-34,-31,-31,-0,-30" 6400 \
1895 "0,0.025 -38,-31,-28,-28,-0,-25" \
1896 gain 15 highpass 22 highpass 22 sinc -n 255 -b 16 -17500 \
1897 gain 9 lowpass -1 17801
1898 The audio file is played with a simulated FM radio sound (or
1899 broadcast signal condition if the lowpass filter at the end is
1900 skipped). Note that the pipeline is set up with US-style 75us
1901 pre-emphasis.
1902 See also compand for a single-band companding effect.
1904 noiseprof [profile-file]
1906 Calculate a profile of the audio for use in noise reduction.
1907 See the description of the noisered effect for details.
1908 noisered [profile-file [amount]]
1910 Reduce noise in the audio signal by profiling and filtering.
1911 This effect is moderately effective at removing consistent back‐
1912 ground noise such as hiss or hum. To use it, first run SoX with
1913 the noiseprof effect on a section of audio that ideally would
1914 contain silence but in fact contains noise - such sections are
1915 typically found at the beginning or the end of a recording.
1916 noiseprof will write out a noise profile to profile-file, or to
1917 stdout if no profile-file or if `-' is given. E.g.
1918 sox speech.wav -n trim 0 1.5 noiseprof speech.noise-profile
1919 To actually remove the noise, run SoX again, this time with the
1920 noisered effect; noisered will reduce noise according to a noise
1921 profile (which was generated by noiseprof), from profile-file,
1922 or from stdin if no profile-file or if `-' is given. E.g.
1923 sox speech.wav cleaned.wav noisered speech.noise-profile 0.3
1924 How much noise should be removed is specified by amount-a number
1925 between 0 and 1 with a default of 0.5. Higher numbers will
1926 remove more noise but present a greater likelihood of removing
1927 wanted components of the audio signal. Before replacing an
1928 original recording with a noise-reduced version, experiment with
1929 different amount values to find the optimal one for your audio;
1930 use headphones to check that you are happy with the results,
1931 paying particular attention to quieter sections of the audio.
1932 On most systems, the two stages - profiling and reduction - can
1934 be combined using a pipe, e.g.
1935 sox noisy.wav -n trim 0 1 noiseprof | play noisy.wav noisered
1936 norm [dB-level]
1938 Normalise the audio. norm is just an alias for gain -n; see the
1939 gain effect for details.
1940 oops Out Of Phase Stereo effect. Mixes stereo to twin-mono where
1942 each mono channel contains the difference between the left and
1943 right stereo channels. This is sometimes known as the `karaoke'
1944 effect as it often has the effect of removing most or all of the
1945 vocals from a recording. It is equivalent to remix 1,2i 1,2i.
1946 overdrive [gain(20) [colour(20)]]
1948 Non linear distortion. The colour parameter controls the amount
1949 of even harmonic content in the over-driven output.
1950 pad { length[@position(=)] }
1952 Pad the audio with silence, at the beginning, the end, or any
1953 specified points through the audio. length is the amount of
1954 silence to insert and position the position in the input audio
1955 stream at which to insert it. Any number of lengths and posi‐
1956 tions may be specified, provided that a specified position is
1957 not less that the previous one, and any time specification may
1958 be used for them. position is optional for the first and last
1959 lengths specified and if omitted correspond to the beginning and
1960 the end of the audio respectively. For example, pad 1.5 1.5
1961 adds 1.5 seconds of silence padding at each end of the audio,
1962 whilst pad 4000s@3:00 inserts 4000 samples of silence 3 minutes
1963 into the audio. If silence is wanted only at the end of the
1964 audio, specify either the end position or specify a zero-length
1965 pad at the start.
1966 See also delay for an effect that can add silence at the begin‐
1968 ning of the audio on a channel-by-channel basis.
1969 phaser gain-in gain-out delay decay speed [-s|-t]
1971 Add a phasing effect to the audio. See [3] for a detailed
1972 description of phasing.
1973 delay/decay/speed gives the delay in milliseconds and the decay
1975 (relative to gain-in) with a modulation speed in Hz. The modu‐
1976 lation is either sinusoidal (-s) - preferable for multiple
1977 instruments, or triangular (-t) - gives single instruments a
1978 sharper phasing effect. The decay should be less than 0.5 to
1979 avoid feedback, and usually no less than 0.1. Gain-out is the
1980 volume of the output.
1981 For example:
1983 play snare.flac phaser 0.8 0.74 3 0.4 0.5 -t
1984 Gentler:
1985 play snare.flac phaser 0.9 0.85 4 0.23 1.3 -s
1986 A popular sound:
1987 play snare.flac phaser 0.89 0.85 1 0.24 2 -t
1988 More severe:
1989 play snare.flac phaser 0.6 0.66 3 0.6 2 -t
1990 pitch [-q] shift [segment [search [overlap]]]
1992 Change the audio pitch (but not tempo).
1993 shift gives the pitch shift as positive or negative `cents'
1995 (i.e. 100ths of a semitone). See the tempo effect for a
1996 description of the other parameters.
1997 See also the bend, speed, and tempo effects.
1999 rate [-q|-l|-m|-h|-v] [override-options] RATE[k]
2001 Change the audio sampling rate (i.e. resample the audio) to any
2002 given RATE (even non-integer if this is supported by the output
2003 file format) using a quality level defined as follows:
2004 Quality Band- Rej dB Typical Use
2006 width
2007 -q quick n/a ≈30 @ playback on
2008 Fs/4 ancient hardware
2009 -l low 80% 100 playback on old
2010 hardware
2011 -m medium 95% 100 audio playback
2012 -h high 95% 125 16-bit mastering
2015 (use with dither)
2016 -v very high 95% 175 24-bit mastering
2017 where Band-width is the percentage of the audio frequency band
2019 that is preserved and Rej dB is the level of noise rejection.
2020 Increasing levels of resampling quality come at the expense of
2021 increasing amounts of time to process the audio. If no quality
2022 option is given, the quality level used is `high' (but see
2023 `Playing & Recording Audio' above regarding playback).
2024 The `quick' algorithm uses cubic interpolation; all others use
2026 band-limited interpolation. By default, all algorithms have a
2027 `linear' phase response; for `medium', `high' and `very high',
2028 the phase response is configurable (see below).
2029 The rate effect is invoked automatically if SoX's -r option
2031 specifies a rate that is different to that of the input file(s).
2032 Alternatively, if this effect is given explicitly, then SoX's -r
2033 option need not be given. For example, the following two com‐
2034 mands are equivalent:
2035 sox input.wav -r 48k output.wav bass -b 24
2036 sox input.wav output.wav bass -b 24 rate 48k
2037 though the second command is more flexible as it allows rate
2038 options to be given, and allows the effects to be ordered arbi‐
2039 trarily.
2040 * * *
2042 Warning: technically detailed discussion follows.
2044 The simple quality selection described above provides settings
2046 that satisfy the needs of the vast majority of resampling tasks.
2047 Occasionally, however, it may be desirable to fine-tune the
2048 resampler's filter response; this can be achieved using over‐
2049 ride options, as detailed in the following table:
2050 -M/-I/-L Phase response = minimum/intermediate/linear
2052 -s Steep filter (band-width = 99%)
2053 -a Allow aliasing/imaging above the pass-band
2054 -b 74-99.7 Any band-width %
2055 -p 0-100 Any phase response (0 = minimum, 25 = intermediate,
2056 50 = linear, 100 = maximum)
2057 N.B. Override options cannot be used with the `quick' or `low'
2059 quality algorithms.
2060 All resamplers use filters that can sometimes create `echo'
2062 (a.k.a. `ringing') artefacts with transient signals such as
2063 those that occur with `finger snaps' or other highly percussive
2064 sounds. Such artefacts are much more noticeable to the human
2065 ear if they occur before the transient (`pre-echo') than if they
2066 occur after it (`post-echo'). Note that frequency of any such
2067 artefacts is related to the smaller of the original and new sam‐
2068 pling rates but that if this is at least 44.1kHz, then the arte‐
2069 facts will lie outside the range of human hearing.
2070 A phase response setting may be used to control the distribution
2072 of any transient echo between `pre' and `post': with minimum
2073 phase, there is no pre-echo but the longest post-echo; with lin‐
2074 ear phase, pre and post echo are in equal amounts (in signal
2075 terms, but not audibility terms); the intermediate phase setting
2076 attempts to find the best compromise by selecting a small length
2077 (and level) of pre-echo and a medium lengthed post-echo.
2078 Minimum, intermediate, or linear phase response is selected
2080 using the -M, -I, or -L option; a custom phase response can be
2081 created with the -p option. Note that phase responses between
2082 `linear' and `maximum' (greater than 50) are rarely useful.
2083 A resampler's band-width setting determines how much of the fre‐
2085 quency content of the original signal (w.r.t. the original sam‐
2086 ple rate when up-sampling, or the new sample rate when down-sam‐
2087 pling) is preserved during conversion. The term `pass-band' is
2088 used to refer to all frequencies up to the band-width point
2089 (e.g. for 44.1kHz sampling rate, and a resampling band-width of
2090 95%, the pass-band represents frequencies from 0Hz (D.C.) to
2091 circa 21kHz). Increasing the resampler's band-width results in
2092 a slower conversion and can increase transient echo artefacts
2093 (and vice versa).
2094 The -s `steep filter' option changes resampling band-width from
2096 the default 95% (based on the 3dB point), to 99%. The -b option
2097 allows the band-width to be set to any value in the range
2098 74-99.7 %, but note that band-width values greater than 99% are
2099 not recommended for normal use as they can cause excessive tran‐
2100 sient echo.
2101 If the -a option is given, then aliasing/imaging above the pass-
2103 band is allowed. For example, with 44.1kHz sampling rate, and a
2104 resampling band-width of 95%, this means that frequency content
2105 above 21kHz can be distorted; however, since this is above the
2106 pass-band (i.e. above the highest frequency of interest/audi‐
2107 bility), this may not be a problem. The benefits of allowing
2108 aliasing/imaging are reduced processing time, and reduced (by
2109 almost half) transient echo artefacts. Note that if this option
2110 is given, then the minimum band-width allowable with -b
2111 increases to 85%.
2112 Examples:
2114 sox input.wav -b 16 output.wav rate -s -a 44100 dither -s
2115 default (high) quality resampling; overrides: steep filter,
2116 allow aliasing; to 44.1kHz sample rate; noise-shaped dither to
2117 16-bit WAV file.
2118 sox input.wav -b 24 output.aiff rate -v -I -b 90 48k
2119 very high quality resampling; overrides: intermediate phase,
2120 band-width 90%; to 48k sample rate; store output to 24-bit AIFF
2121 file.
2122 * * *
2124 The pitch and speed effects use the rate effect at their core.
2126 remix [-a|-m|-p] <out-spec>
2128 out-spec = in-spec{,in-spec} | 0
2129 in-spec = [in-chan][-[in-chan2]][vol-spec]
2130 vol-spec = p|i|v[volume]
2131 Select and mix input audio channels into output audio channels.
2133 Each output channel is specified, in turn, by a given out-spec:
2134 a list of contributing input channels and volume specifications.
2135 Note that this effect operates on the audio channels within the
2137 SoX effects processing chain; it should not be confused with the
2138 -m global option (where multiple files are mix-combined before
2139 entering the effects chain).
2140 An out-spec contains comma-separated input channel-numbers and
2142 hyphen-delimited channel-number ranges; alternatively, 0 may be
2143 given to create a silent output channel. For example,
2144 sox input.wav output.wav remix 6 7 8 0
2145 creates an output file with four channels, where channels 1, 2,
2146 and 3 are copies of channels 6, 7, and 8 in the input file, and
2147 channel 4 is silent. Whereas
2148 sox input.wav output.wav remix 1-3,7 3
2149 creates a (somewhat bizarre) stereo output file where the left
2150 channel is a mix-down of input channels 1, 2, 3, and 7, and the
2151 right channel is a copy of input channel 3.
2152 Where a range of channels is specified, the channel numbers to
2154 the left and right of the hyphen are optional and default to 1
2155 and to the number of input channels respectively. Thus
2156 sox input.wav output.wav remix -
2157 performs a mix-down of all input channels to mono.
2158 By default, where an output channel is mixed from multiple (n)
2160 input channels, each input channel will be scaled by a factor of
2161 ¹/n. Custom mixing volumes can be set by following a given
2162 input channel or range of input channels with a vol-spec (volume
2163 specification). This is one of the letters p, i, or v, followed
2164 by a volume number, the meaning of which depends on the given
2165 letter and is defined as follows:
2166 Letter Volume number Notes
2168 p power adjust in dB 0 = no change
2169 i power adjust in dB As `p', but invert
2171 the audio
2172 v voltage multiplier 1 = no change, 0.5
2173 ≈ 6dB attenuation,
2174 2 ≈ 6dB gain, -1 =
2175 invert
2176 If an out-spec includes at least one vol-spec then, by default,
2178 ¹/n scaling is not applied to any other channels in the same
2179 out-spec (though may be in other out-specs). The -a (automatic)
2180 option however, can be given to retain the automatic scaling in
2181 this case. For example,
2182 sox input.wav output.wav remix 1,2 3,4v0.8
2183 results in channel level multipliers of 0.5,0.5 1,0.8, whereas
2184 sox input.wav output.wav remix -a 1,2 3,4v0.8
2185 results in channel level multipliers of 0.5,0.5 0.5,0.8.
2186 The -m (manual) option disables all automatic volume adjust‐
2188 ments, so
2189 sox input.wav output.wav remix -m 1,2 3,4v0.8
2190 results in channel level multipliers of 1,1 1,0.8.
2191 The volume number is optional and omitting it corresponds to no
2193 volume change; however, the only case in which this is useful is
2194 in conjunction with i. For example, if input.wav is stereo,
2195 then
2196 sox input.wav output.wav remix 1,2i
2197 is a mono equivalent of the oops effect.
2198 If the -p option is given, then any automatic ¹/n scaling is
2200 replaced by ¹/√n (`power') scaling; this gives a louder mix but
2201 one that might occasionally clip.
2202 * * *
2204 One use of the remix effect is to split an audio file into a set
2206 of files, each containing one of the constituent channels (in
2207 order to perform subsequent processing on individual audio chan‐
2208 nels). Where more than a few channels are involved, a script
2209 such as the following (Bourne shell script) is useful:
2210 #!/bin/sh
2211 chans=`soxi -c "$1"`
2212 while [ $chans -ge 1 ]; do
2213 chans0=`printf %02i $chans` # 2 digits hence up to 99 chans
2214 out=`echo "$1"|sed "s/\(.*\)\.\(.*\)/\1-$chans0.\2/"`
2215 sox "$1" "$out" remix $chans
2216 chans=`expr $chans - 1`
2217 done
2218 If a file input.wav containing six audio channels were given,
2219 the script would produce six output files: input-01.wav,
2220 input-02.wav, ..., input-06.wav.
2221 See also the swap effect.
2223 repeat [count(1)|-]
2225 Repeat the entire audio count times, or once if count is not
2226 given. The special value - requests infinite repetition.
2227 Requires temporary file space to store the audio to be repeated.
2228 Note that repeating once yields two copies: the original audio
2229 and the repeated audio.
2230 reverb [-w|--wet-only] [reverberance (50%) [HF-damping (50%)
2232 [room-scale (100%) [stereo-depth (100%)
2233 [pre-delay (0ms) [wet-gain (0dB)]]]]]]
2234 Add reverberation to the audio using the `freeverb' algorithm.
2236 A reverberation effect is sometimes desirable for concert halls
2237 that are too small or contain so many people that the hall's
2238 natural reverberance is diminished. Applying a small amount of
2239 stereo reverb to a (dry) mono signal will usually make it sound
2240 more natural. See [3] for a detailed description of reverbera‐
2241 tion.
2242 Note that this effect increases both the volume and the length
2244 of the audio, so to prevent clipping in these domains, a typical
2245 invocation might be:
2246 play dry.wav gain -3 pad 0 3 reverb
2247 The -w option can be given to select only the `wet' signal, thus
2248 allowing it to be processed further, independently of the `dry'
2249 signal. E.g.
2250 play -m voice.wav "|sox voice.wav -p reverse reverb -w reverse"
2251 for a reverse reverb effect.
2252 reverse
2254 Reverse the audio completely. Requires temporary file space to
2255 store the audio to be reversed.
2256 riaa Apply RIAA vinyl playback equalisation. The sampling rate must
2258 be one of: 44.1, 48, 88.2, 96 kHz.
2259 This effect supports the --plot global option.
2261 silence [-l] above-periods [duration threshold[d|%]
2263 [below-periods duration threshold[d|%]]
2264 Removes silence from the beginning, middle, or end of the audio.
2266 `Silence' is determined by a specified threshold.
2267 The above-periods value is used to indicate if audio should be
2269 trimmed at the beginning of the audio. A value of zero indicates
2270 no silence should be trimmed from the beginning. When specifying
2271 a non-zero above-periods, it trims audio up until it finds non-
2272 silence. Normally, when trimming silence from beginning of audio
2273 the above-periods will be 1 but it can be increased to higher
2274 values to trim all audio up to a specific count of non-silence
2275 periods. For example, if you had an audio file with two songs
2276 that each contained 2 seconds of silence before the song, you
2277 could specify an above-period of 2 to strip out both silence
2278 periods and the first song.
2279 When above-periods is non-zero, you must also specify a duration
2281 and threshold. duration indicates the amount of time that non-
2282 silence must be detected before it stops trimming audio. By
2283 increasing the duration, burst of noise can be treated as
2284 silence and trimmed off.
2285 threshold is used to indicate what sample value you should treat
2287 as silence. For digital audio, a value of 0 may be fine but for
2288 audio recorded from analog, you may wish to increase the value
2289 to account for background noise.
2290 When optionally trimming silence from the end of the audio, you
2292 specify a below-periods count. In this case, below-period means
2293 to remove all audio after silence is detected. Normally, this
2294 will be a value 1 of but it can be increased to skip over peri‐
2295 ods of silence that are wanted. For example, if you have a song
2296 with 2 seconds of silence in the middle and 2 second at the end,
2297 you could set below-period to a value of 2 to skip over the
2298 silence in the middle of the audio.
2299 For below-periods, duration specifies a period of silence that
2301 must exist before audio is not copied any more. By specifying a
2302 higher duration, silence that is wanted can be left in the
2303 audio. For example, if you have a song with an expected 1 sec‐
2304 ond of silence in the middle and 2 seconds of silence at the
2305 end, a duration of 2 seconds could be used to skip over the mid‐
2306 dle silence.
2307 Unfortunately, you must know the length of the silence at the
2309 end of your audio file to trim off silence reliably. A work‐
2310 around is to use the silence effect in combination with the
2311 reverse effect. By first reversing the audio, you can use the
2312 above-periods to reliably trim all audio from what looks like
2313 the front of the file. Then reverse the file again to get back
2314 to normal.
2315 To remove silence from the middle of a file, specify a below-
2317 periods that is negative. This value is then treated as a posi‐
2318 tive value and is also used to indicate that the effect should
2319 restart processing as specified by the above-periods, making it
2320 suitable for removing periods of silence in the middle of the
2321 audio.
2322 The option -l indicates that below-periods duration length of
2324 audio should be left intact at the beginning of each period of
2325 silence. For example, if you want to remove long pauses between
2326 words but do not want to remove the pauses completely.
2327 duration is a time specification with the peculiarity that a
2329 bare number is interpreted as a sample count, not as a number of
2330 seconds. For specifying seconds, either use the t suffix (as in
2331 `2t') or specify minutes, too (as in `0:02').
2332 threshold numbers may be suffixed with d to indicate the value
2334 is in decibels, or % to indicate a percentage of maximum value
2335 of the sample value (0% specifies pure digital silence).
2336 The following example shows how this effect can be used to start
2338 a recording that does not contain the delay at the start which
2339 usually occurs between `pressing the record button' and the
2340 start of the performance:
2341 rec parameters filename other-effects silence 1 5 2%
2342 sinc [-a att|-b beta] [-p phase|-M|-I|-L] [-t tbw|-n taps] [freqHP]
2344 [-freqLP [-t tbw|-n taps]]
2345 Apply a sinc kaiser-windowed low-pass, high-pass, band-pass, or
2346 band-reject filter to the signal. The freqHP and freqLP parame‐
2347 ters give the frequencies of the 6dB points of a high-pass and
2348 low-pass filter that may be invoked individually, or together.
2349 If both are given, then freqHP less than freqLP creates a band-
2350 pass filter, freqHP greater than freqLP creates a band-reject
2351 filter. For example, the invocations
2352 sinc 3k
2353 sinc -4k
2354 sinc 3k-4k
2355 sinc 4k-3k
2356 create a high-pass, low-pass, band-pass, and band-reject filter
2357 respectively.
2358 The default stop-band attenuation of 120dB can be overridden
2360 with -a; alternatively, the kaiser-window `beta' parameter can
2361 be given directly with -b.
2362 The default transition band-width of 5% of the total band can be
2364 overridden with -t (and tbw in Hertz); alternatively, the number
2365 of filter taps can be given directly with -n.
2366 If both freqHP and freqLP are given, then a -t or -n option
2368 given to the left of the frequencies applies to both frequen‐
2369 cies; one of these options given to the right of the frequencies
2370 applies only to freqLP.
2371 The -p, -M, -I, and -L options control the filter's phase
2373 response; see the rate effect for details.
2374 This effect supports the --plot global option.
2376 spectrogram [options]
2378 Create a spectrogram of the audio; the audio is passed unmodi‐
2379 fied through the SoX processing chain. This effect is optional
2380 - type sox --help and check the list of supported effects to see
2381 if it has been included.
2382 The spectrogram is rendered in a Portable Network Graphic (PNG)
2384 file, and shows time in the X-axis, frequency in the Y-axis, and
2385 audio signal magnitude in the Z-axis. Z-axis values are repre‐
2386 sented by the colour (or optionally the intensity) of the pixels
2387 in the X-Y plane. If the audio signal contains multiple chan‐
2388 nels then these are shown from top to bottom starting from chan‐
2389 nel 1 (which is the left channel for stereo audio).
2390 For example, if `my.wav' is a stereo file, then with
2392 sox my.wav -n spectrogram
2393 a spectrogram of the entire file will be created in the file
2394 `spectrogram.png'. More often though, analysis of a smaller
2395 portion of the audio is required; e.g. with
2396 sox my.wav -n remix 2 trim 20 30 spectrogram
2397 the spectrogram shows information only from the second (right)
2398 channel, and of thirty seconds of audio starting from twenty
2399 seconds in. To analyse a small portion of the frequency domain,
2400 the rate effect may be used, e.g.
2401 sox my.wav -n rate 6k spectrogram
2402 allows detailed analysis of frequencies up to 3kHz (half the
2403 sampling rate) i.e. where the human auditory system is most sen‐
2404 sitive. With
2405 sox my.wav -n trim 0 10 spectrogram -x 600 -y 200 -z 100
2406 the given options control the size of the spectrogram's X, Y & Z
2407 axes (in this case, the spectrogram area of the produced image
2408 will be 600 by 200 pixels in size and the Z-axis range will be
2409 100 dB). Note that the produced image includes axes legends
2410 etc. and so will be a little larger than the specified spectro‐
2411 gram size. In this example:
2412 sox -n -n synth 6 tri 10k:14k spectrogram -z 100 -w kaiser
2413 an analysis `window' with high dynamic range is selected to best
2414 display the spectrogram of a swept triangular wave. For a smi‐
2415 lar example, append the following to the `chime' command in the
2416 description of the delay effect (above):
2417 rate 2k spectrogram -X 200 -Z -10 -w kaiser
2418 Options are also available to control the appearance (colour-
2419 set, brightness, contrast, etc.) and filename of the spectro‐
2420 gram; e.g. with
2421 sox my.wav -n spectrogram -m -l -o print.png
2422 a spectrogram is created suitable for printing on a `black and
2423 white' printer.
2424 Options:
2426 -x num Change the (maximum) width (X-axis) of the spectrogram
2428 from its default value of 800 pixels to a given number
2429 between 100 and 200000. See also -X and -d.
2430 -X num X-axis pixels/second; the default is auto-calculated to
2432 fit the given or known audio duration to the X-axis size,
2433 or 100 otherwise. If given in conjunction with -d, this
2434 option affects the width of the spectrogram; otherwise,
2435 it affects the duration of the spectrogram. num can be
2436 from 1 (low time resolution) to 5000 (high time resolu‐
2437 tion) and need not be an integer. SoX may make a slight
2438 adjustment to the given number for processing quantisa‐
2439 tion reasons; if so, SoX will report the actual number
2440 used (viewable when the SoX global option -V is in
2441 effect). See also -x and -d.
2442 -y num Sets the Y-axis size in pixels (per channel); this is the
2444 number of frequency `bins' used in the Fourier analysis
2445 that produces the spectrogram. N.B. it can be slow to
2446 produce the spectrogram if this number is not one more
2447 than a power of two (e.g. 129). By default the Y-axis
2448 size is chosen automatically (depending on the number of
2449 channels). See -Y for alternative way of setting spec‐
2450 trogram height.
2451 -Y num Sets the target total height of the spectrogram(s). The
2453 default value is 550 pixels. Using this option (and by
2454 default), SoX will choose a height for individual spec‐
2455 trogram channels that is one more than a power of two, so
2456 the actual total height may fall short of the given num‐
2457 ber. However, there is also a minimum height per channel
2458 so if there are many channels, the number may be
2459 exceeded. See -y for alternative way of setting spectro‐
2460 gram height.
2461 -z num Z-axis (colour) range in dB, default 120. This sets the
2463 dynamic-range of the spectrogram to be -num dBFS to
2464 0 dBFS. Num may range from 20 to 180. Decreasing
2465 dynamic-range effectively increases the `contrast' of the
2466 spectrogram display, and vice versa.
2467 -Z num Sets the upper limit of the Z-axis in dBFS. A negative
2469 num effectively increases the `brightness' of the spec‐
2470 trogram display, and vice versa.
2471 -q num Sets the Z-axis quantisation, i.e. the number of differ‐
2473 ent colours (or intensities) in which to render Z-axis
2474 values. A small number (e.g. 4) will give a
2475 `poster'-like effect making it easier to discern magni‐
2476 tude bands of similar level. Small numbers also usually
2477 result in small PNG files. The number given specifies
2478 the number of colours to use inside the Z-axis range; two
2479 colours are reserved to represent out-of-range values.
2480 -w name
2482 Window: Hann (default), Hamming, Bartlett, Rectangular,
2483 Kaiser or Dolph. The spectrogram is produced using the
2484 Discrete Fourier Transform (DFT) algorithm. A signifi‐
2485 cant parameter to this algorithm is the choice of `window
2486 function'. By default, SoX uses the Hann window which
2487 has good all-round frequency-resolution and dynamic-range
2488 properties. For better frequency resolution (but lower
2489 dynamic-range), select a Hamming window; for higher
2490 dynamic-range (but poorer frequency-resolution), select a
2491 Dolph window. Kaiser, Bartlett and Rectangular windows
2492 are also available.
2493 -W num Window adjustment parameter. This can be used to make
2495 small adjustments to the Kaiser or Dolph window shape. A
2496 positive number (up to ten) increases its dynamic range,
2497 a negative number decreases it.
2498 -s Allow slack overlapping of DFT windows. This can, in
2500 some cases, increase image sharpness and give greater
2501 adherence to the -x value, but at the expense of a little
2502 spectral loss.
2503 -m Creates a monochrome spectrogram (the default is colour).
2505 -h Selects a high-colour palette - less visually pleasing
2507 than the default colour palette, but it may make it eas‐
2508 ier to differentiate different levels. If this option is
2509 used in conjunction with -m, the result will be a hybrid
2510 monochrome/colour palette.
2511 -p num Permute the colours in a colour or hybrid palette. The
2513 num parameter, from 1 (the default) to 6, selects the
2514 permutation.
2515 -l Creates a `printer friendly' spectrogram with a light
2517 background (the default has a dark background).
2518 -a Suppress the display of the axis lines. This is some‐
2520 times useful in helping to discern artefacts at the spec‐
2521 trogram edges.
2522 -r Raw spectrogram: suppress the display of axes and leg‐
2524 ends.
2525 -A Selects an alternative, fixed colour-set. This is pro‐
2527 vided only for compatibility with spectrograms produced
2528 by another package. It should not normally be used as it
2529 has some problems, not least, a lack of differentiation
2530 at the bottom end which results in masking of low-level
2531 artefacts.
2532 -t text
2534 Set the image title - text to display above the spectro‐
2535 gram.
2536 -c text
2538 Set (or clear) the image comment - text to display below
2539 and to the left of the spectrogram.
2540 -o file
2542 Name of the spectrogram output PNG file, default `spec‐
2543 trogram.png'. If `-' is given, the spectrogram will be
2544 sent to standard output (stdout).
2545 Advanced Options:
2547 In order to process a smaller section of audio without affecting
2548 other effects or the output signal (unlike when the trim effect
2549 is used), the following options may be used.
2550 -d duration
2552 This option sets the X-axis resolution such that audio
2553 with the given duration (a time specification) fits the
2554 selected (or default) X-axis width. For example,
2555 sox input.mp3 output.wav -n spectrogram -d 1:00 stats
2556 creates a spectrogram showing the first minute of the
2557 audio, whilst
2558 the stats effect is applied to the entire audio signal.
2559 See also -X for an alternative way of setting the X-axis
2561 resolution.
2562 -S position(=)
2564 Start the spectrogram at the given point in the audio
2565 stream. For example
2566 sox input.aiff output.wav spectrogram -S 1:00
2567 creates a spectrogram showing all but the first minute of
2568 the audio (the output file, however, receives the entire
2569 audio stream).
2570 For the ability to perform off-line processing of spectral data,
2572 see the stat effect.
2573 speed factor[c]
2575 Adjust the audio speed (pitch and tempo together). factor is
2576 either the ratio of the new speed to the old speed: greater than
2577 1 speeds up, less than 1 slows down, or, if appended with the
2578 letter `c', the number of cents (i.e. 100ths of a semitone) by
2579 which the pitch (and tempo) should be adjusted: greater than 0
2580 increases, less than 0 decreases.
2581 Technically, the speed effect only changes the sample rate
2583 information, leaving the samples themselves untouched. The rate
2584 effect is invoked automatically to resample to the output sample
2585 rate, using its default quality/speed. For higher quality or
2586 higher speed resampling, in addition to the speed effect, spec‐
2587 ify the rate effect with the desired quality option.
2588 See also the bend, pitch, and tempo effects.
2590 splice [-h|-t|-q] { position(=)[,excess[,leeway]] }
2592 Splice together audio sections. This effect provides two things
2593 over simple audio concatenation: a (usually short) cross-fade is
2594 applied at the join, and a wave similarity comparison is made to
2595 help determine the best place at which to make the join.
2596 One of the options -h, -t, or -q may be given to select the fade
2598 envelope as half-cosine wave (the default), triangular (a.k.a.
2599 linear), or quarter-cosine wave respectively.
2600 Type Audio Fade level Transitions
2602 t correlated constant gain abrupt
2603 h correlated constant gain smooth
2604 q uncorrelated constant power smooth
2605 To perform a splice, first use the trim effect to select the
2607 audio sections to be joined together. As when performing a tape
2608 splice, the end of the section to be spliced onto should be
2609 trimmed with a small excess (default 0.005 seconds) of audio
2610 after the ideal joining point. The beginning of the audio sec‐
2611 tion to splice on should be trimmed with the same excess (before
2612 the ideal joining point), plus an additional leeway (default
2613 0.005 seconds). Any time specification may be used for these
2614 parameters. SoX should then be invoked with the two audio sec‐
2615 tions as input files and the splice effect given with the posi‐
2616 tion at which to perform the splice - this is length of the
2617 first audio section (including the excess).
2618 The following diagram uses the tape analogy to illustrate the
2620 splice operation. The effect simulates the diagonal cuts and
2621 joins the two pieces:
2622 length1 excess
2624 -----------><--->
2625 _________ : : _________________
2626 \ : : :\ `
2627 \ : : : \ `
2628 \: : : \ `
2629 * : : * - - *
2630 \ : : :\ `
2631 \ : : : \ `
2632 _______________\: : : \_____`____
2633 : : : :
2634 <---> <----->
2635 excess leeway
2636 where * indicates the joining points.
2638 For example, a long song begins with two verses which start (as
2640 determined e.g. by using the play command with the trim (start)
2641 effect) at times 0:30.125 and 1:03.432. The following commands
2642 cut out the first verse:
2643 sox too-long.wav part1.wav trim 0 30.130
2644 (5 ms excess, after the first verse starts)
2645 sox too-long.wav part2.wav trim 1:03.422
2646 (5 ms excess plus 5 ms leeway, before the second verse starts)
2647 sox part1.wav part2.wav just-right.wav splice 30.130
2648 For another example, the SoX command
2649 play "|sox -n -p synth 1 sin %1" "|sox -n -p synth 1 sin %3"
2650 generates and plays two notes, but there is a nasty click at the
2651 transition; the click can be removed by splicing instead of con‐
2652 catenating the audio, i.e. by appending splice 1 to the command.
2653 (Clicks at the beginning and end of the audio can be removed by
2654 preceding the splice effect with fade q .01 2 .01).
2655 Provided your arithmetic is good enough, multiple splices can be
2657 performed with a single splice invocation. For example:
2658 #!/bin/sh
2659 # Audio Copy and Paste Over
2660 # acpo infile copy-start copy-stop paste-over-start outfile
2661 # No chained time specifications allowed for the parameters
2662 # (i.e. such that contain +/-).
2663 e=0.005 # Using default excess
2664 l=$e # and leeway.
2665 sox "$1" piece.wav trim $2-$e-$l =$3+$e
2666 sox "$1" part1.wav trim 0 $4+$e
2667 sox "$1" part2.wav trim $4+$3-$2-$e-$l
2668 sox part1.wav piece.wav part2.wav "$5" \
2669 splice $4+$e +$3-$2+$e+$l+$e
2670 In the above Bourne shell script, two splices are used to `copy
2671 and paste' audio.
2672 * * *
2674 It is also possible to use this effect to perform general cross-
2676 fades, e.g. to join two songs. In this case, excess would typi‐
2677 cally be an number of seconds, the -q option would typically be
2678 given (to select an `equal power' cross-fade), and leeway should
2679 be zero (which is the default if -q is given). For example, if
2680 f1.wav and f2.wav are audio files to be cross-faded, then
2681 sox f1.wav f2.wav out.wav splice -q $(soxi -D f1.wav),3
2682 cross-fades the files where the point of equal loudness is 3
2683 seconds before the end of f1.wav, i.e. the total length of the
2684 cross-fade is 2 × 3 = 6 seconds (Note: the $(...) notation is
2685 POSIX shell).
2686 stat [-s scale] [-rms] [-freq] [-v] [-d]
2688 Display time and frequency domain statistical information about
2689 the audio. Audio is passed unmodified through the SoX process‐
2690 ing chain.
2691 The information is output to the `standard error' (stderr)
2693 stream and is calculated, where n is the duration of the audio
2694 in samples, c is the number of audio channels, r is the audio
2695 sample rate, and xk represents the PCM value (in the range -1 to
2696 +1 by default) of each successive sample in the audio, as fol‐
2697 lows:
2698 Samples read n×c
2700 Length (seconds) n÷r
2701 Scaled by See -s below.
2702 Maximum amplitude max(xk) The maximum sample
2703 value in the audio;
2704 usually this will
2705 be a positive num‐
2706 ber.
2707 Minimum amplitude min(xk) The minimum sample
2708 value in the audio;
2709 usually this will
2710 be a negative num‐
2711 ber.
2712 Midline amplitude ½min(xk)+½max(xk)
2713 Mean norm ¹/nΣ│xk│ The average of the
2714 absolute value of
2715 each sample in the
2716 audio.
2717 Mean amplitude ¹/nΣxk The average of each
2718 sample in the
2719 audio. If this
2720 figure is non-zero,
2721 then it indicates
2722 the presence of a
2723 D.C. offset (which
2724 could be removed
2725 using the dcshift
2726 effect).
2727 RMS amplitude √(¹/nΣxk²) The level of a D.C.
2731 signal that would
2732 have the same power
2733 as the audio's
2734 average power.
2735 Maximum delta max(│xk-xk-1│)
2736 Minimum delta min(│xk-xk-1│)
2737 Mean delta ¹/n-1Σ│xk-xk-1│
2738 RMS delta √(¹/n-1Σ(xk-xk-1)²)
2739 Rough frequency In Hz.
2740 Volume Adjustment The parameter to
2741 the vol effect
2742 which would make
2743 the audio as loud
2744 as possible without
2745 clipping. Note:
2746 See the discussion
2747 on Clipping above
2748 for reasons why it
2749 is rarely a good
2750 idea actually to do
2751 this.
2752 Note that the delta measurements are not applicable for multi-
2754 channel audio.
2755 The -s option can be used to scale the input data by a given
2757 factor. The default value of scale is 2147483647 (i.e. the max‐
2758 imum value of a 32-bit signed integer). Internal effects always
2759 work with signed long PCM data and so the value should relate to
2760 this fact.
2761 The -rms option will convert all output average values to `root
2763 mean square' format.
2764 The -v option displays only the `Volume Adjustment' value.
2766 The -freq option calculates the input's power spectrum (4096
2768 point DFT) instead of the statistics listed above. This should
2769 only be used with a single channel audio file.
2770 The -d option displays a hex dump of the 32-bit signed PCM data
2772 audio in SoX's internal buffer. This is mainly used to help
2773 track down endian problems that sometimes occur in cross-plat‐
2774 form versions of SoX.
2775 See also the stats effect.
2777 stats [-b bits|-x bits|-s scale] [-w window-time]
2779 Display time domain statistical information about the audio
2780 channels; audio is passed unmodified through the SoX processing
2781 chain. Statistics are calculated and displayed for each audio
2782 channel and, where applicable, an overall figure is also given.
2783 For example, for a typical well-mastered stereo music file:
2785 Overall Left Right
2787 DC offset 0.000803 -0.000391 0.000803
2788 Min level -0.750977 -0.750977 -0.653412
2789 Max level 0.708801 0.708801 0.653534
2790 Pk lev dB -2.49 -2.49 -3.69
2791 RMS lev dB -19.41 -19.13 -19.71
2792 RMS Pk dB -13.82 -13.82 -14.38
2793 RMS Tr dB -85.25 -85.25 -82.66
2794 Crest factor - 6.79 6.32
2795 Flat factor 0.00 0.00 0.00
2796 Pk count 2 2 2
2797 Bit-depth 16/16 16/16 16/16
2798 Num samples 7.72M
2799 Length s 174.973
2800 Scale max 1.000000
2801 Window s 0.050
2802 DC offset, Min level, and Max level are shown, by default, in
2804 the range ±1. If the -b (bits) options is given, then these
2805 three measurements will be scaled to a signed integer with the
2806 given number of bits; for example, for 16 bits, the scale would
2807 be -32768 to +32767. The -x option behaves the same way as -b
2808 except that the signed integer values are displayed in hexadeci‐
2809 mal. The -s option scales the three measurements by a given
2810 floating-point number.
2811 Pk lev dB and RMS lev dB are standard peak and RMS level mea‐
2813 sured in dBFS. RMS Pk dB and RMS Tr dB are peak and trough val‐
2814 ues for RMS level measured over a short window (default 50ms).
2815 Crest factor is the standard ratio of peak to RMS level (note:
2817 not in dB).
2818 Flat factor is a measure of the flatness (i.e. consecutive sam‐
2820 ples with the same value) of the signal at its peak levels (i.e.
2821 either Min level, or Max level). Pk count is the number of
2822 occasions (not the number of samples) that the signal attained
2823 either Min level, or Max level.
2824 The right-hand Bit-depth figure is the standard definition of
2826 bit-depth i.e. bits less significant than the given number are
2827 fixed at zero. The left-hand figure is the number of most sig‐
2828 nificant bits that are fixed at zero (or one for negative num‐
2829 bers) subtracted from the right-hand figure (the number sub‐
2830 tracted is directly related to Pk lev dB).
2831 For multi-channel audio, an overall figure for each of the above
2833 measurements is given and derived from the channel figures as
2834 follows: DC offset: maximum magnitude; Max level, Pk lev dB,
2835 RMS Pk dB, Bit-depth: maximum; Min level, RMS Tr dB: minimum;
2836 RMS lev dB, Flat factor, Pk count: average; Crest factor: not
2837 applicable.
2838 Length s is the duration in seconds of the audio, and Num sam‐
2840 ples is equal to the sample-rate multiplied by Length.
2841 Scale Max is the scaling applied to the first three measure‐
2842 ments; specifically, it is the maximum value that could apply to
2843 Max level. Window s is the length of the window used for the
2844 peak and trough RMS measurements.
2845 See also the stat effect.
2847 swap Swap stereo channels. If the input is not stereo, pairs of
2849 channels are swapped, and a possible odd last channel passed
2850 through. E.g., for seven channels, the output order will be 2,
2851 1, 4, 3, 6, 5, 7.
2852 See also remix for an effect that allows arbitrary channel
2854 selection and ordering (and mixing).
2855 stretch factor [window fade shift fading]
2857 Change the audio duration (but not its pitch). This effect is
2858 broadly equivalent to the tempo effect with (factor inverted
2859 and) search set to zero, so in general, its results are compara‐
2860 tively poor; it is retained as it can sometimes out-perform
2861 tempo for small factors.
2862 factor of stretching: >1 lengthen, <1 shorten duration. window
2864 size is in ms. Default is 20ms. The fade option, can be `lin'.
2865 shift ratio, in [0 1]. Default depends on stretch factor. 1 to
2866 shorten, 0.8 to lengthen. The fading ratio, in [0 0.5]. The
2867 amount of a fade's default depends on factor and shift.
2868 See also the tempo effect.
2870 synth [-j KEY] [-n] [len [off [ph [p1 [p2 [p3]]]]]] {[type] [combine]
2872 [[%]freq[k][:|+|/|-[%]freq2[k]]] [off [ph [p1 [p2 [p3]]]]]}
2873 This effect can be used to generate fixed or swept frequency
2874 audio tones with various wave shapes, or to generate wide-band
2875 noise of various `colours'. Multiple synth effects can be cas‐
2876 caded to produce more complex waveforms; at each stage it is
2877 possible to choose whether the generated waveform will be mixed
2878 with, or modulated onto the output from the previous stage.
2879 Audio for each channel in a multi-channel audio file can be syn‐
2880 thesised independently.
2881 Though this effect is used to generate audio, an input file must
2883 still be given, the characteristics of which will be used to set
2884 the synthesised audio length, the number of channels, and the
2885 sampling rate; however, since the input file's audio is not nor‐
2886 mally needed, a `null file' (with the special name -n) is often
2887 given instead (and the length specified as a parameter to synth
2888 or by another given effect that has an associated length).
2889 For example, the following produces a 3 second, 48kHz, audio
2891 file containing a sine-wave swept from 300 to 3300 Hz:
2892 sox -n output.wav synth 3 sine 300-3300
2893 and this produces an 8 kHz version:
2894 sox -r 8000 -n output.wav synth 3 sine 300-3300
2895 Multiple channels can be synthesised by specifying the set of
2896 parameters shown between braces multiple times; the following
2897 puts the swept tone in the left channel and adds `brown' noise
2898 in the right:
2899 sox -n output.wav synth 3 sine 300-3300 brownnoise
2900 The following example shows how two synth effects can be cas‐
2901 caded to create a more complex waveform:
2902 play -n synth 0.5 sine 200-500 synth 0.5 sine fmod 700-100
2903 Frequencies can also be given in `scientific' note notation, or,
2904 by prefixing a `%' character, as a number of semitones relative
2905 to `middle A' (440 Hz). For example, the following could be
2906 used to help tune a guitar's low `E' string:
2907 play -n synth 4 pluck %-29
2908 or with a (Bourne shell) loop, the whole guitar:
2909 for n in E2 A2 D3 G3 B3 E4; do
2910 play -n synth 4 pluck $n repeat 2; done
2911 See the delay effect (above) and the reference to `SoX scripting
2912 examples' (below) for more synth examples.
2913 N.B. This effect generates audio at maximum volume (0dBFS),
2915 which means that there is a high chance of clipping when using
2916 the audio subsequently, so in many cases, you will want to fol‐
2917 low this effect with the gain effect to prevent this from hap‐
2918 pening. (See also Clipping above.) Note that, by default, the
2919 synth effect incorporates the functionality of gain -h (see the
2920 gain effect for details); synth's -n option may be given to dis‐
2921 able this behaviour.
2922 A detailed description of each synth parameter follows:
2924 len is the length of audio to synthesise (any time specifica‐
2926 tion); a value of 0 indicated to use the input length, which is
2927 also the default.
2928 type is one of sine, square, triangle, sawtooth, trapezium, exp,
2930 [white]noise, tpdfnoise, pinknoise, brownnoise, pluck;
2931 default=sine.
2932 combine is one of create, mix, amod (amplitude modulation), fmod
2934 (frequency modulation); default=create.
2935 freq/freq2 are the frequencies at the beginning/end of synthesis
2937 in Hz or, if preceded with `%', semitones relative to A
2938 (440 Hz); alternatively, `scientific' note notation (e.g. E2)
2939 may be used. The default frequency is 440Hz. By default, the
2940 tuning used with the note notations is `equal temperament'; the
2941 -j KEY option selects `just intonation', where KEY is an integer
2942 number of semitones relative to A (so for example, -9 or 3
2943 selects the key of C), or a note in scientific notation.
2944 If freq2 is given, then len must also have been given and the
2946 generated tone will be swept between the given frequencies. The
2947 two given frequencies must be separated by one of the characters
2948 `:', `+', `/', or `-'. This character is used to specify the
2949 sweep function as follows:
2950 : Linear: the tone will change by a fixed number of hertz
2952 per second.
2953 + Square: a second-order function is used to change the
2955 tone.
2956 / Exponential: the tone will change by a fixed number of
2958 semitones per second.
2959 - Exponential: as `/', but initial phase always zero, and
2961 stepped (less smooth) frequency changes.
2962 Not used for noise.
2964 off is the bias (DC-offset) of the signal in percent; default=0.
2966 ph is the phase shift in percentage of 1 cycle; default=0. Not
2968 used for noise.
2969 p1 is the percentage of each cycle that is `on' (square), or
2971 `rising' (triangle, exp, trapezium); default=50 (square, trian‐
2972 gle, exp), default=10 (trapezium), or sustain (pluck);
2973 default=40.
2974 p2 (trapezium): the percentage through each cycle at which
2976 `falling' begins; default=50. exp: the amplitude in multiples of
2977 2dB; default=50, or tone-1 (pluck); default=20.
2978 p3 (trapezium): the percentage through each cycle at which
2980 `falling' ends; default=60, or tone-2 (pluck); default=90.
2981 tempo [-q] [-m|-s|-l] factor [segment [search [overlap]]]
2983 Change the audio playback speed but not its pitch. This effect
2984 uses the WSOLA algorithm. The audio is chopped up into segments
2985 which are then shifted in the time domain and overlapped (cross-
2986 faded) at points where their waveforms are most similar as
2987 determined by measurement of `least squares'.
2988 By default, linear searches are used to find the best overlap‐
2990 ping points. If the optional -q parameter is given, tree
2991 searches are used instead. This makes the effect work more
2992 quickly, but the result may not sound as good. However, if you
2993 must improve the processing speed, this generally reduces the
2994 sound quality less than reducing the search or overlap values.
2995 The -m option is used to optimize default values of segment,
2997 search and overlap for music processing.
2998 The -s option is used to optimize default values of segment,
3000 search and overlap for speech processing.
3001 The -l option is used to optimize default values of segment,
3003 search and overlap for `linear' processing that tends to cause
3004 more noticeable distortion but may be useful when factor is
3005 close to 1.
3006 If -m, -s, or -l is specified, the default value of segment will
3008 be calculated based on factor, while default search and overlap
3009 values are based on segment. Any values you provide still over‐
3010 ride these default values.
3011 factor gives the ratio of new tempo to the old tempo, so e.g.
3013 1.1 speeds up the tempo by 10%, and 0.9 slows it down by 10%.
3014 The optional segment parameter selects the algorithm's segment
3016 size in milliseconds. If no other flags are specified, the
3017 default value is 82 and is typically suited to making small
3018 changes to the tempo of music. For larger changes (e.g. a factor
3019 of 2), 41 ms may give a better result. The -m, -s, and -l flags
3020 will cause the segment default to be automatically adjusted
3021 based on factor. For example using -s (for speech) with a tempo
3022 of 1.25 will calculate a default segment value of 32.
3023 The optional search parameter gives the audio length in mil‐
3025 liseconds over which the algorithm will search for overlapping
3026 points. If no other flags are specified, the default value is
3027 14.68. Larger values use more processing time and may or may
3028 not produce better results. A practical maximum is half the
3029 value of segment. Search can be reduced to cut processing time
3030 at the risk of degrading output quality. The -m, -s, and -l
3031 flags will cause the search default to be automatically adjusted
3032 based on segment.
3033 The optional overlap parameter gives the segment overlap length
3035 in milliseconds. Default value is 12, but -m, -s, or -l flags
3036 automatically adjust overlap based on segment size. Increasing
3037 overlap increases processing time and may increase quality. A
3038 practical maximum for overlap is the value of search, with over‐
3039 lap typically being (at least) a little smaller then search.
3040 See also speed for an effect that changes tempo and pitch
3042 together, pitch and bend for effects that change pitch only, and
3043 stretch for an effect that changes tempo using a different algo‐
3044 rithm.
3045 treble gain [frequency[k] [width[s|h|k|o|q]]]
3047 Apply a treble tone-control effect. See the description of the
3048 bass effect for details.
3049 tremolo speed [depth]
3051 Apply a tremolo (low frequency amplitude modulation) effect to
3052 the audio. The tremolo frequency in Hz is given by speed, and
3053 the depth as a percentage by depth (default 40).
3054 trim {position(+)}
3056 Cuts portions out of the audio. Any number of positions may be
3057 given; audio is not sent to the output until the first position
3058 is reached. The effect then alternates between copying and dis‐
3059 carding audio at each position. Using a value of 0 for the
3060 first position parameter allows copying from the beginning of
3061 the audio.
3062 For example,
3064 sox infile outfile trim 0 10
3065 will copy the first ten seconds, while
3066 play infile trim 12:34 =15:00 -2:00
3067 and
3068 play infile trim 12:34 2:26 -2:00
3069 will both play from 12 minutes 34 seconds into the audio up to
3070 15 minutes into the audio (i.e. 2 minutes and 26 seconds long),
3071 then resume playing two minutes before the end of audio.
3072 upsample [factor]
3074 Upsample the signal by an integer factor: factor-1 zero-value
3075 samples are inserted between each pair of input samples. As a
3076 result, the original spectrum is replicated into the new fre‐
3077 quency space (imaging) and attenuated. This attenuation can be
3078 compensated for by adding vol factor after any further process‐
3079 ing. The upsample effect is typically used in combination with
3080 filtering effects.
3081 For a general resampling effect with anti-imaging, see rate.
3083 See also downsample.
3084 vad [options]
3086 Voice Activity Detector. Attempts to trim silence and quiet
3087 background sounds from the ends of (fairly high resolution i.e.
3088 16-bit, 44-48kHz) recordings of speech. The algorithm currently
3089 uses a simple cepstral power measurement to detect voice, so may
3090 be fooled by other things, especially music. The effect can
3091 trim only from the front of the audio, so in order to trim from
3092 the back, the reverse effect must also be used. E.g.
3093 play speech.wav norm vad
3094 to trim from the front,
3095 play speech.wav norm reverse vad reverse
3096 to trim from the back, and
3097 play speech.wav norm vad reverse vad reverse
3098 to trim from both ends. The use of the norm effect is recom‐
3099 mended, but remember that neither reverse nor norm is suitable
3100 for use with streamed audio.
3101 Options:
3103 Default values are shown in parenthesis.
3104 -t [22mnum (7)
3106 The measurement level used to trigger activity detection.
3107 This might need to be changed depending on the noise
3108 level, signal level and other charactistics of the input
3109 audio.
3110 -T num (0.25)
3112 The time constant (in seconds) used to help ignore short
3113 bursts of sound.
3114 -s [22mnum (1)
3116 The amount of audio (in seconds) to search for qui‐
3117 eter/shorter bursts of audio to include prior to the
3118 detected trigger point.
3119 -g num (0.25)
3121 Allowed gap (in seconds) between quieter/shorter bursts
3122 of audio to include prior to the detected trigger point.
3123 -p [22mnum (0)
3125 The amount of audio (in seconds) to preserve before the
3126 trigger point and any found quieter/shorter bursts.
3127 Advanced Options:
3129 These allow fine tuning of the algorithm's internal parameters.
3130 -b num The algorithm (internally) uses adaptive noise estima‐
3132 tion/reduction in order to detect the start of the wanted
3133 audio. This option sets the time for the initial noise
3134 estimate.
3135 -N num Time constant used by the adaptive noise estimator for
3137 when the noise level is increasing.
3138 -n num Time constant used by the adaptive noise estimator for
3140 when the noise level is decreasing.
3141 -r num Amount of noise reduction to use in the detection algo‐
3143 rithm (e.g. 0, 0.5, ...).
3144 -f num Frequency of the algorithm's processing/measurements.
3146 -m num Measurement duration; by default, twice the measurement
3148 period; i.e. with overlap.
3149 -M num Time constant used to smooth spectral measurements.
3151 -h num `Brick-wall' frequency of high-pass filter applied at the
3153 input to the detector algorithm.
3154 -l num `Brick-wall' frequency of low-pass filter applied at the
3156 input to the detector algorithm.
3157 -H num `Brick-wall' frequency of high-pass lifter used in the
3159 detector algorithm.
3160 -L num `Brick-wall' frequency of low-pass lifter used in the
3162 detector algorithm.
3163 See also the silence effect.
3165 vol gain [type [limitergain]]
3167 Apply an amplification or an attenuation to the audio signal.
3168 Unlike the -v option (which is used for balancing multiple input
3169 files as they enter the SoX effects processing chain), vol is an
3170 effect like any other so can be applied anywhere, and several
3171 times if necessary, during the processing chain.
3172 The amount to change the volume is given by gain which is inter‐
3174 preted, according to the given type, as follows: if type is
3175 amplitude (or is omitted), then gain is an amplitude (i.e. volt‐
3176 age or linear) ratio, if power, then a power (i.e. wattage or
3177 voltage-squared) ratio, and if dB, then a power change in dB.
3178 When type is amplitude or power, a gain of 1 leaves the volume
3180 unchanged, less than 1 decreases it, and greater than 1
3181 increases it; a negative gain inverts the audio signal in addi‐
3182 tion to adjusting its volume.
3183 When type is dB, a gain of 0 leaves the volume unchanged, less
3185 than 0 decreases it, and greater than 0 increases it.
3186 See [4] for a detailed discussion on electrical (and hence audio
3188 signal) voltage and power ratios.
3189 Beware of Clipping when the increasing the volume.
3191 The gain and the type parameters can be concatenated if desired,
3193 e.g. vol 10dB.
3194 An optional limitergain value can be specified and should be a
3196 value much less than 1 (e.g. 0.05 or 0.02) and is used only on
3197 peaks to prevent clipping. Not specifying this parameter will
3198 cause no limiter to be used. In verbose mode, this effect will
3199 display the percentage of the audio that needed to be limited.
3200 See also gain for a volume-changing effect with different capa‐
3202 bilities, and compand for a dynamic-range compression/expan‐
3203 sion/limiting effect.
3204
3206 Exit status is 0 for no error, 1 if there is a problem with the com‐
3207 mand-line parameters, or 2 if an error occurs during file processing.
3208
3210 Please report any bugs found in this version of SoX to the mailing list
3211 (sox-users@lists.sourceforge.net).
3212
3214 soxi(1), soxformat(7), libsox(3)
3215 audacity(1), gnuplot(1), octave(1), wget(1)
3216 The SoX web site at http://sox.sourceforge.net
3217 SoX scripting examples at http://sox.sourceforge.net/Docs/Scripts
3218 References
3220 [1] R. Bristow-Johnson, Cookbook formulae for audio EQ biquad filter
3221 coefficients, http://musicdsp.org/files/Audio-EQ-Cookbook.txt
3222 [2] Wikipedia, Q-factor, http://en.wikipedia.org/wiki/Q_factor
3224 [3] Scott Lehman, Effects Explained, http://harmony-cen‐
3226 tral.com/Effects/effects-explained.html
3227 [4] Wikipedia, Decibel, http://en.wikipedia.org/wiki/Decibel
3229 [5] Richard Furse, Linux Audio Developer's Simple Plugin API,
3231 http://www.ladspa.org
3232 [6] Richard Furse, Computer Music Toolkit, http://www.ladspa.org/cmt
3234 [7] Steve Harris, LADSPA plugins, http://plugin.org.uk
3236
3238 Copyright 1998-2013 Chris Bagwell and SoX Contributors.
3239 Copyright 1991 Lance Norskog and Sundry Contributors.
3240 This program is free software; you can redistribute it and/or modify it
3242 under the terms of the GNU General Public License as published by the
3243 Free Software Foundation; either version 2, or (at your option) any
3244 later version.
3245 This program is distributed in the hope that it will be useful, but
3247 WITHOUT ANY WARRANTY; without even the implied warranty of MER‐
3248 CHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
3249 Public License for more details.
3250
3252 Chris Bagwell (cbagwell@users.sourceforge.net). Other authors and con‐
3253 tributors are listed in the ChangeLog file that is distributed with the
3254 source code.
3255sox December 31, 2014 SoX(1)