Provided by: cw_3.5.1-2_amd64 
NAME
cw - sound characters as Morse code on the soundcard or console speaker
SYNOPSIS
cw [-s --system=SYSTEM] [-d --device=DEVICE] [-w --wpm=WPM] [-t --tone=HZ] [-v --volume=PERCENT]
[-g --gap=GAP] [-k --weighting=WEIGHT] [-e --noecho] [-m --nomessages] [-c --nocommands]
[-o --nocombinations] [-p --nocomments] [-f --infile=FILE] [-h --help] [-V --version]
cw installed on GNU/Linux systems understands both short form and long form command line options. cw
installed on other operating systems may understand only the short form options.
There are no mandatory options.
Options may be predefined in the environment variable CW_OPTIONS. If defined, these options are used
first; command line options take precedence.
DESCRIPTION
cw reads characters from an input file, or from standard input, and sounds each valid character as Morse
code on either the system sound card, or the system console speaker. After it sounds a character, cw
echoes it to standard output. The input stream can contain embedded command strings. These change the
parameters used when sounding the Morse code. cw reports any errors in embedded commands on standard
error.
Use 'Ctrl+D' key combination to exit cw.
COMMAND LINE OPTIONS
cw understands the following command line options. The long form options may not be available in non-
LINUX versions.
-s, --system=SYSTEM
Specifies the way that cw generates tones. Valid values are: null for no tones, just timings,
console for tones through the console speaker, alsa for tones generated through the system sound
card using ALSA sound system, oss for tones generated through system sound card using OSS sound
system, pulseaudio for tones generated through system sound card using PulseAudio sound system,
soundcard for tones generated through the system sound card, but without explicit selection of
sound system. These values can be shortened to 'n', 'c', 'a', 'o', 'p', or 's', respectively. The
default value is 'pulseaudio' (on systems with PulseAudio installed), followed by 'oss'.
-d, --device=DEVICE
Specifies the device file to open for generating a sound. cw will use default device if none is
specified. The default devices are: /dev/console for sound produced through console, default for
ALSA sound system, /dev/audio for OSS sound system, a default device for PulseAudio sound system.
See also NOTES ON USING A SOUND CARD below.
-w, --wpm=WPM
Sets the initial sending speed in words per minute. The value must be between 4 and 60. The
default value is 12 WPM.
-t, --tone=HZ
Sets the initial sounder pitch in Hz. This value must be between 0 and 4,000. A value of 0
selects silent operation, and can be used for timing checks or other testing. The default value
is 800Hz,
-v, --volume=PERCENT
Sets the initial sending volume, as a percentage of full scale volume. The value must be between
0 and 100. The default value is 70 %. Sound volumes work fully for sound card tones, but cw
cannot control the volume of tones from the console speaker. In this case, a volume of zero is
silent, and all other volume values are simply sounded.
-g, --gap=GAP
Sets the initial extra gap, in dot lengths, between characters (the 'Farnsworth' delay). It must
be between 0 and 60. The default is 0.
-k, --weighting=WEIGHT
Sets the initial weighting, as a percentage of dot lengths. It must be between 20 and 80. The
default is 50.
-e, --noecho
Stops cw echoing characters on standard output after they are sounded. The default is to have
echoing on.
-m, --nomessages
Stops cw printing error messages on standard error. The default is to print messages.
-c, --nocommands
Stops cw from interpreting commands embedded in the input stream. The default is to interpret
embedded commands.
-o, --nocombinations
Stops cw from treating character strings bracketed by [...] as a single combination character.
The default is to honor combinations.
-p, --nocomments
Stops cw from treating character strings bracketed by {...} as 'comments'. Characters inside
these braces will be echoed to standard output, but not sounded. When comments are being honored,
any embedded commands inside the braces will be ignored. The default is to honor comments.
-f, --infile=FILE
Specifies a text file that cw can read to configure its practice text.
-h, --help
Prints short help message.
-V, --version
Prints information about program's version, authors and license.
SOUNDING CHARACTERS
cw reads characters, one at a time, from its standard input or from its input file. Lowercase letters
are converted internally to uppercase. The following list shows the valid IS0 8859-1 (Latin-1) characters
that can be sounded by cw:
ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"$()+-./:;=?_@ and space
In addition, the program also understands the following ISO 8859-1 and ISO 8859-2 accented characters:
ÜÄÇÖÉÈÀÑŞ (S with cedilla), Ž (Z with caron/hacek),
and accepts the following as single character forms of common procedural signals:
<>!&^~
See cw(7,LOCAL) for more information on the above characters and Morse code.
If cw receives a character not in this set, it prints an error message '?c', where c is the error
character. The only exceptions to this may be the cw command escape character '%', the combination start
and stop characters '[' and ']', and the comment start and stop characters '{' and '}'. See EMBEDDED
COMMANDS and MORSE CODE COMBINATIONS below.
EMBEDDED COMMANDS
cw recognizes special sequences in the input stream as embedded commands. These commands alter the
parameters of the cw while it is running, or query current values. All commands are prefixed by the
command escape character '%', and those which set a value end with a semicolon.
The format of an embedded command to change a parameter value is
%Cvalue;
where C is a command letter indicating what action cw is to take, and value is the argument or value for
the command.
Valid command letters are
T Sets the tone pitch used to sound a character.
W Sets the sending speed.
G Sets the 'Farnsworth' gap between characters.
K Sets the weighting.
E Disables or re-enables echoing of sent characters on standard output.
M Disables or re-enables error messages on standard error.
S Disables or re-enables speaker tone generation.
C Disables processing of embedded commands. Note that once disabled, this command cannot re-enable
them.
O Disables or re-enables recognition of [...] character combinations.
P Disables or re-enables recognition of {...} comments. When comments are being recognized, any
character after an opening '{' and before any closing '}' will be echoed to standard output, but
will not be sounded, or have any other effect.
For example, the embedded command sequence
%W25;%T1200;
will set cw to a speed of 25 WPM, and a tone pitch of 1200Hz.
The 'T', 'W', 'G', and 'A' commands take values along with the command. The limits on values given for
embedded commands are the same as the limits available for command line options, detailed above.
The 'E', 'M', 'S', 'C' and 'O' commands are flags, and treat a value of zero as clear, and any other
value as set. So, for example, the sequence
%M0;%C0;
will turn off error messages, and then turn off the processing of embedded commands.
If a parameter is set successfully, cw reports the new setting on standard error (except if no error
messages is set). If an error is detected in an embedded command, cw reports an error. For the formats
of error messages see the MESSAGE FORMATS section below.
The current values of parameters within cw may be queried, as well as set. The command format
%?C
queries the value of the parameter normally set with command C. cw reports the current value on standard
error, using the same format as when new values are set.
The current values of parameters within cw may also be requested as output in Morse code. The command
format
%>C
will generate Morse output reporting the value of the parameter normally set with command C.
If embedded commands are disabled, '%' characters are treated as any other (in this case, invalid) input
character.
Once processing of embedded commands has been switched off, any command to switch this feature back on
will not be recognized. That is, after '%C0;', an '%C1;' will not be recognized.
There is one additional command, and that is '%Q'. This command closes all open files and terminates cw.
Any characters after this command in the input stream will be lost.
The file cw.h provides a full set of definitions for the commands, special characters, and status codes
of cw.
MESSAGE FORMATS
Where a parameter value is set correctly with an embedded command, the message format
=Cvalue
is returned. C is the command used, and value is the new value.
If an invalid value is supplied for a parameter in an embedded command, a message
?Cvalue
is returned.
Where an invalid command is encountered, the message format
?%C
is used. For an invalid query, the message is
??C
and for an invalid request for a parameter in Morse code the message is
?>C
A character in the input stream that cannot be sounded produces a message
?C
These messages are not intended to be user-friendly, but are designed to be easily and quickly
interpreted by another program. Similarly, the format of embedded commands is more computer-friendly
than user-friendly.
If error messages are disabled, no messages of any type are printed on standard error.
MORSE CODE COMBINATIONS
The standard set of characters offered by cw may not be sufficient for some purposes. For example, some
international characters do not have equivalent ISO 8859-1 and ISO 8859-2 that cw can sound directly.
To help in sounding such characters, cw offers the ability to form combination characters by placing
individual character components between [...] brackets. Cw sounds characters inside a combination
without the usual gap between them. In this way, any missing character in the set can be built.
For example
[VA]
is one way to form the VA procedural signal, though
[SK]
works just as well. The eight-dot error signal can be sounded with
[HSE]
or the C-cedilla in international Morse code with
[CE]
There can be as many valid letters, numbers, or figures inside the [...] brackets as required. For
example, an alternative way of sending the error signal could be
[EEEEEEEE]
Finally, three alternative ways of sending 73 might be
[TTEEE][EEETT]
[TDE][EUT]
[GEE][VT]
Embedded commands may be placed inside [...] combinations if required. Combinations do not nest.
This feature can be disabled by using the -O or --nocombinations command line flags, or with the 'O'
embedded command. If combinations are disabled, '[' and ']' characters are treated as any other
(invalid) input character.
NOTES ON USING A SOUND CARD
By default, cw tries to open default PulseAudio. If PulseAudio server is not accessible, cw tries to open
OSS device "/dev/audio" to access the system sound card. This is generally the correct device to use,
but for systems with special requirements, or those with multiple sound cards, the option -d or --device,
combined with -s or --system can be used to specify the device and audio system for sound card access.
If the sound card device cannot be set up, cw prints the error message
cannot set up soundcard sound
and exits.
Sound card devices, when opened through OSS sound system, are usually single-access devices, so that when
one process has opened the device, other processes are prevented from using it. In such cases cw will of
course conflict with any other programs that expect exclusive use of the system sound card (for example,
MP3 players). If cw finds that the sound card is already busy, it prints the error message
open /dev/audio: Device or resource busy
and exits.
The sound card device is not used if cw is only sending tones on the console speaker.
AUDIO OUTPUT - DEFAULTS AND SELECTION
cw first tries to access sound card using PulseAudio sound system, using default device name, unless user
specifies other audio device with option -d or --device.
cw then tries to access sound card using OSS audio system and default OSS audio device name
('/dev/audio'), unless user specifies other audio device with option -d or --device.
If opening soundcard through OSS fails, cw tries to access the sound card using ALSA audio system, and
default ALSA audio device name ('default'), unless user specifies other audio device with option -d or
--device.
If opening soundcard through ALSA also fails, cw tries to access system console buzzer using default
buzzer device '/dev/console', unless user specifies other audio device with option -d or --device.
It is very common that in order to access the console buzzer device user has to have root privileges.
For that reason trying to open console buzzer almost always fails. This is not a program's bug, this is
a result of operating system's restrictions. Making cw an suid binary bypasses this restriction. The
program does not fork() or exec(), so making it suid should be relatively safe. Note however that this
practice is discouraged for security reasons.
As stated, user can tell cw which device to use, using -d or --device option. Which device files are
suitable will depend on which operating system is running, which system user ID runs cw, and which user
groups user belongs to.
NOTES
Despite the fact that this manual page constantly and consistently refers to Morse code elements as dots
and dashes, DO NOT think in these terms when trying to learn Morse code. Always think of them as 'dit's
and 'dah's.
The Morse code table in the cw(7,LOCAL) man page is provided for reference only. If learning for the
first time, you will be much better off learning by hearing the characters sent, rather than by looking
at the table.
Other programs running in the system may interfere with the timing of the Morse code that cw is sending.
If this is a problem, either try to run on a quiescent system, or try running cw with nice(1L,C,1). UNIX
is not really designed for user-level programs to do the sort of fine timing required to send Morse code.
cw is therefore more sensitive than most programs to other system activity.
cw uses system itimers for its internal timing. On most UNIX flavors, itimers are not guaranteed to
signal a program exactly at the specified time, and they generally offer a resolution only as good as the
normal system 'clock tick' resolution. An itimer SIGALRM usually falls on a system clock tick, making it
accurate to no better than 10mS on a typical 100Hz kernel.
The effect of this is that an itimer period is generally either exactly as specified, or, more likely,
slightly longer. At higher WPM settings, the cumulative effect of this affects timing accuracy, because
at higher speeds, there are fewer 10mS clock ticks in a dot period. For example, at 12 WPM, the dot
length is 100mS, enough to contain five kernel clock ticks. But at 60 WPM, the dot length is 20mS, or
just two kernel clock ticks. So at higher speeds, the effect of itimer resolutions becomes more
pronounced.
To test itimer timing, first try
X="PARIS PARIS PARIS PARIS "
echo "$X" | time cw -w 4
and note the elapsed time, which should be very close to one minute. Next, try
echo "$X$X$X$X$X$X$X$X$X$X$X$X" | time cw -w 48
The elapsed time should be the same. If it has increased, this is the effect of system itimers delaying
for slightly longer than the specified period (higher WPM rates make more itimer calls). That's itimers
for you, not perfect for this job, but the best there is without writing some, and perhaps a lot of,
kernel code.
Except for zero, which is silent, tone values lower than 10Hz may not sound at the expected pitch.
EXAMPLES
Send a string of characters at 25 WPM, 700Hz, with no extra gaps:
echo "UNIX CW SOUNDER" | cw -w 25 -t 700
Send a string at varying speeds and tones on the console speaker, specifying a system console device:
echo "%W12;%T400;400HZ 12WPM %W25;%T1500;1500HZ 25WPM" | cw -m -sc -d /dev/tty2
Send C-cedilla, VA, and a report of the WPM setting, with extra spacing at half volume:
echo "[CE] [VA] %>W" | cw -g 10 -v 50
ERRORS AND OMISSIONS
Cut numbers are not provided, though they can be emulated, up to a point, by pre-filtering.
An output to an optional external device, for example, keying a line on the parallel port, or a serial
line, might also be useful.
SEE ALSO
Man pages for cw(7,LOCAL), libcw(3,LOCAL), cwgen(1,LOCAL), cwcp(1,LOCAL), and xcwcp(1,LOCAL).