vgetty voice shell commands

                           vgetty voice shell commands
                          -----------------------------

              The voice extensions for the mgetty package

                    Copyright (C) 1995, 1997 Marc Eberhard.
                    Many contributions by others, including
                    Marc SCHAEFER <schaefer@alphanet.ch>, the
                    current vgetty maintainer.
                       http://www-internal.alphanet.ch/~schaefer/vgetty.html

                          Voice shell command guide

             Writen by Mizser Krisztian <christo@dundi.sch.bme.hu>
           (please email me corrections, additional informations ...)



 1.) Introduction
 2.) Default behaviour of vgetty (if you don't specify "call_program"
     in voice.conf)
 3.) Custom behaviour of vgetty (if you specify "call_program" in voice.conf)
 4.) The structure of "call_program" (and the other "*_programs" 's)
  4.1) General things
  4.2) Example script
  4.3) Another script
 5.) Commands / answers to/from the voice library and explanation of them
 6.) Possible events, detected tones
 7.) Converting voice files to/from your modem's format
 8.) Logging



 1.) Introduction


  Vgetty is a great tool for creating an answering machine or any other
  system that uses your voice modem's capabilities. It is fully
  customizable, so you can create your own system with features you
  would like.
  Vgetty can play, record messages, can detect DTMF tones, data/fax
  calling tones.
  Vgetty works with mgetty closely, so it calls mgetty when data or fax
  connection is needed to establish.

  You can control the behaviour of vgetty in voice.conf. The options
  are well explained in voice.conf, please read them to have an idea
  what vgetty can do and how it does them.



 2.) Default behaviour of vgetty (if you don't specify "call_program"
     in voice.conf):


  NOTE: Stuff within double quotes "" refers to variables in voice.conf

  Vgetty answers the phone after "rings" rings, then plays the greeting
  message specified in "message_list" (if specified). During playing the
  modem reports detected DTMF tones, vgetty handles these as follows:
  A DTMF command is in the form of *numbers# , if vgetty detects such a
  sequence, it stops playing, then executes "dtmf_program" if specified
  with 'numbers' as argument (without * and #).
  If vgetty detects a single number only, it stops playing, then tries to
  make a data connection with the remote end, if it doesn't succeed, then
  it tries a fax connection with the remote end.
  If vgetty detects an empty # character (without *numbers  before it), then
  it simply hangs up the phone.
  If no DTMF was received during playing, vgetty issues a beep command, then
  it switches to recording mode, when the remote end can say the message to
  be recorded. Recording ends if DTMF tones detected as described above (in
  this case vgetty removes the recorded message),
  or if "record_max_len" time exceeded, or if the remote end hangs up.
  When finished, vgetty executes "message_program" if specified with the
  recorded filename as argument.

  NOTE: In this case vgetty gives different names to the recorded messages,
        you don't need to bother with this question.


 3.) Custom behaviour of vgetty (if you specify "call_program" in voice.conf):


  Vgetty answers the phone after "rings" rings, then executes the shell script
  defined by "call_program". After the "call_program" terminates, vgetty does
  different actions depending on the exit code as follows:
     Exit code 1: vgetty tries to make a DATA connection with the remote end
     Exit code 2: vgetty tries to make a FAX connection with the remote end
     Exit code 3: vgetty first tries to make a DATA, then in case of
                  unsuccess a FAX connection with the remote end
     If none of
     the above  : vgetty closes the device and exits.

  NOTE: All the above mentioned programs (call_program, dtmf_program,
        message_program, button_program) are standard shell scripts executed
        by /bin/sh (Marc, is it true?)
  NOTE: In this case, you have to give different names to your recorded
        messages, vgetty cannot do this for you since you control everything
        in your script.


 4.) The structure of "call_program" (and the other "*_programs" 's:

 4.1) General things

  Most of you who want to do something else than vgetty's default behaviour
  will have to define "call_program" in voice.conf and write a nice shell
  script that handles everything after the modem picks up the phone.
  The principles: A shell script communicates with the voice library through
  unique file descriptors ($VOICE_INPUT and $VOICE_OUTPUT). See the two basic
  function 'receive' and 'send' in scripts/demo.sh . Just define these two
  functions in each of your scripts, and you don't have to deal with them
  anymore. So what do these two functions do?
  Receive: It reads one message from the voice library.
  Send   : It passes one message (command) to the voice library.
   (The voice library is just the interface between your shell script and the
    modem. It translates the modem's messages to normal, human readable strings
    and vice versa.)
  You can consider the messages from the voice_library as members in a FIFO
  (first-in first-out) queue. You can read one message from the queue using
  the receive function (it will return the oldest message in the queue).
  Every time your modem detects something or just executes your commands, it
  adds a message to the queue (for example: PLAYING, DTMF_DETECTED, READY, etc.)



  4.2) Example script


   Now that you understand the basic behaviour of the system, let's see an
   example (scripts/dtmf.sh):

     function receive :
     function send    : we discussed these two functions above

     ANSWER=`receive`

    // Executes the "receive" function and puts the returned message in a
    // variable named ANSWER

     if [ "$ANSWER" != "HELLO SHELL" ]; then
          echo "$0: voice library not answering" >&2
          exit 1
     fi

    // If the message contained in ANSWER is not "HELLO SHELL" then exit,
    // because something's wrong.

     send "HELLO VOICE PROGRAM"

    // Executes the "send" function that passes the "HELLO VOICE PROGRAM" to the
    // voice library

     ANSWER=`receive`

     if [ "$ANSWER" != "READY" ]; then
          echo "$0: initialization failed" >&2
          exit 1
     fi

    // If the ANSWER wasn't "READY" then exit, something's wrong.
    // So far nothing happened, we just made sure that the communication between
    // our script and the voice library is OK. You have to start every script
    // like this one.


     send "DEVICE DIALUP_LINE"
     ANSWER=`receive`
     if [ "$ANSWER" != "READY" ]; then
          echo "$0: could not set output device" >&2
          exit 1
     fi

    // Sets the output device to dialup line.

     if [ -f demo.rmd ]; then
          send "PLAY demo.rmd"

    // Let's play demo.rmd

          ANSWER=`receive`

          if [ "$ANSWER" != "PLAYING" ]; then
               echo "$0: could not start playing" >&2
               exit 1
          fi

    // After the PLAY command, voice library should return "PLAYING"

          ANSWER=`receive`

          if [ "$ANSWER" != "READY" ]; then
               echo "$0: something went wrong on playing" >&2
               exit 1
          fi

    // When the playing is finished, we get "READY"
    // If no events occur during playing, the script just waits until
    // playing is finished (e.g. READY)
    // Function 'receive' won't return until an event happens.

     fi

          send "BEEP"
          ANSWER=`receive`

          if [ "$ANSWER" != "BEEPING" ]; then
               echo "$0: could not send a beep" >&2
               exit 1
          fi

          ANSWER=`receive`

          if [ "$ANSWER" != "READY" ]; then
               echo "$0: could not send a beep" >&2
               exit 1
          fi

    // Send out a BEEP, wait for answer (BEEPING, when finished: READY)

          send "RECORD demo.rmd"
          ANSWER=`receive`

          if [ "$ANSWER" != "RECORDING" ]; then
               echo "$0: could not start recording" >&2
               exit 1
          fi

          ANSWER=`receive`

          if [ "$ANSWER" != "READY" ]; then
               echo "$0: something went wrong on recording" >&2
               exit 1
          fi

    // Send a RECORD command, wait for answer (RECORDING, when finished: READY)

     send "GOODBYE"
     ANSWER=`receive`

     if [ "$ANSWER" != "GOODBYE SHELL" ]; then
          echo "$0: could not say goodbye to the voice library" >&2
          exit 1
     fi

    // We don't want to do anything else, send a GOODBYE to the voice library
    // If the answer is "GOODBYE SHELL", then the library got our request, and
    // stops communicating with us.

     exit 0

    // Leave the shell script with exit code 0 .


    It was a very simple script, it just played a greeting message, then sent
    a beep to the line, then recorded a message.

 4.3) Another script

    Let's see a more complicated example, where we want to handle the events
    during playing the greeting message.


        send "ENABLE EVENTS"
        ANSWER=`receive`
        if [ "$ANSWER" != "READY" ]; then
              echo "$0: something's wrong" >&2
              exit 1
        fi

    // This command enabled events reporting to the script. You need this if
    // you want to detect and handle events (dtmf codes, calling tones)

     if [ -f demo.rmd ]; then
          send "PLAY demo.rmd"
     fi

     while true; do
          ANSWER=`receive`

          case $ANSWER in
               PLAYING)
                    ANSWER=`receive`
                    case $ANSWER in
                         READY)            send "PLAY demo.rmd";;
                         RECEIVED_DTMF)    send "STOP"; proc_dtmf;;
                         LOOP_BREAK|BUSY_TONE) send "STOP"; receive; send "GOODBYE";
                                           receive; exit 0;;
                         FAX_CALLING_TONE) send "STOP"; receive;
                                           send "GOODBYE"; receive; exit 2;;
                    esac;;
          esac
     done

  In this example we play the greeting message in an infinite loop. During
  playing we can check what the modem reports. If we get a dtmf code
  (RECEIVED_DTMF), then we call our dtmf handling function (proc_dmtf). If
  the remote end hangs up (at my place it manifests in BUSY_TONE, but
  normally it should be LOOP_BREAK, signalling remote hangup), then we
  stop playing, say GOODBYE to the voice library, then exit with code 0.
  If the modem detects a fax calling tone (FAX_CALLING_TONE), then we stop
  playing and exit with code 2 (which means that mgetty tries to make a fax
  connection with the remote end). You can see that there is a 'receive' after
  STOP and GOODBYE, they are for getting the READY and "GOODBYE SHELL" answers.
  (in fact we should check if another event happened after issuing STOP)
  It is that simple. We just have to make sure we read all messages from the
  voice library. In order to do that, we need to know the possible answers
  to our commands, and the possible messages, events from the voice library.
  (For a more complicated example see scripts/voice_mail.sh )



 5.) Commands and answers to/from the voice library and explanation of them


  In the following table you can see all the commands and the possible answers
  for that particular command.


 Command:       ""
 Answers:       "HELLO SHELL"
 Explanation:   After the very first 'receive' in your script, you should
                get this answer from the voice library. It means that the
                voice lib is ready for your commands.

 Command:       "HELLO VOICE PROGRAM"
 Explanation:   This should be the first command you issue, it will start
                the conversation between your script and the voice lib.
 Answers:       "READY": If everything's ok, voice lib will return this.

 Command:       "ENABLE EVENTS"
 Explanation:   Enables event reporting to the shell (dtmf codes, data/fax
                calling tones, silence, quiet)
 Answers:       "READY": Voice lib got your request.

 Command:       "DISABLE EVENTS"
 Explanation:   Disables event reporting to the shell.
 Answers:       "READY": Voice lib got your request.

 Command:       "DEVICE argument"
 Possible Args: NO_DEVICE, DIALUP_LINE, EXTERNAL_MICROPHONE, INTERNAL_SPEAKER,
                LOCAL_HANDSET, EXTERNAL_SPEAKER
 Explanation:   This command sets the device that vgetty will use.
 Answers:       "READY": Voice lib got your request.
                (also see below).

 Command:       "BEEP [frequency [length]]"
 Explanation:   Sends a beep through the chosen device (if frequency or
                length are not given, the defaults from voice.conf are used)
 Answers:       "BEEPING": Beeping is executed right now.
                "READY": Beeping ended.

 Command:       "DIAL number"
 Explanation:   Dials the given number.
 Answers:       "DIALING": Current action is dialing.
                "READY": Dialing ended.

 Command:       "GET TTY"
 Explanation:   Returns the modem device currently in use.
 Answers:       DevID: Device id.
                "READY": Voicelib ready for next command.

 Command:       "GET MODEM"
 Explanation:   Returns the voice modem driver name, for example
                "US Robotics", see ../libvoice/*.c, at the end
                of the file (the big structure). Then returns "READY".
                The goal of this function is to for example choose the
                right RMD file format to pass for PLAY.
 Answers:       "READY" after <voice driver name>

 Command:       "AUTOSTOP argument"
 Possible Args: ON, OFF
 Explanation:   With AUTOSTOP on, the voicelib will automatically abort a
                play in progress and return READY. This is useful for faster
                reaction times for voice menus.
 Answers:       "READY": Voicelib ready again.

 Command:       "GOODBYE"
 Explanation:   Stops conversation with the voice lib.
 Answers:       "GOODBYE SHELL": Voice lib stops conversation.

 Command:       "PLAY filename"
 Explanation:   Plays filename through the chosen device (the file must be in
                your modem's format!!! See section 7)
 Answers:       "PLAYING": 'filename' is being played.
                "READY": Playing of 'filename' has ended.
 Notes: if you specify an optional non zero third numerical argument, this
        changes the DTMF semantic: it will make the multi-DTMF detection more
        reliable when used with ``AUTOSTOP''. For details, please look
        at Al Borcher's perl vgetty shell script.

 Command:       "RECORD filename"
 Explanation:   Records filename from the chosen device (recorded file will be
                in your modem's format! See section 7)
 Answers:       "RECORDING": 'filename' is being recorded.
                "READY": Recording of 'filename has ended.

 Command:       "WAIT seconds"
 Explanation:   Waits for 'seconds' seconds. During waiting, events will be
                detected and reported!
 Answers:       "WAITING": Waiting has started.
                "READY": Waiting is over. (Note, that after this READY
                         events will NOT be reported to the shell!)

 Command:       "STOP"
 Explanation:   Stops current action (DIALING, PLAYING, RECORDING or WAITING)
                Current state will be IDLE. ( Please note, that events will NOT
                be reported after "STOP" . If you don't want to do anything
                but to detect events, then try the WAIT command.
 Answers:       "READY": Current action successfully stopped.


 Command:       "DTMF"
 Explanation:   Contributed by <cadams@ro.com>, this adds support for sending
                DTMFs over the voice modem from the voice shell.
                The argument is the DTMFs to dial. This can be used for
                example to use vgetty as a voice-mailbox poller.
 Answers        "DTMFING", then "READY".

 Command:       "QUOTE"
 Explanation:   Contributed by <gonz@ratloop.com>, this adds support for
                sending modem-specific raw AT commands to the voice modem
                from the voice shell. The argument is the command to issue.
                Please do not use this lightly since it makes your voice
                shell script modem dependant.
 Answers        "READY".

 If something fails, you get "ERROR" as answer. That can mean two things,
 either you tried to issue multiple commands, e.g. a second PLAY command
 before stopping the first (vgetty will complain about this, saying
 "Nested command in shell script"), or something's really broken (hardware,
 software).

 You will receive:  ``DEVICE_NOT_AVAILABLE'' when the requested DEVICE
 doesn't exist.


 6.) Possible events, detected tones


 Please note, that events will only be reported if you issue an "ENABLE EVENTS"
 command, and current action is either PLAYING, RECORDING, WAITING or
 DIALING. After finishing these actions ("READY"), events won't be reported.
 (Use WAIT if you don't want to do anything but get events)


 Event:        BONG_TONE
 Description:  The modem detected a bong tone on the line.

 Event:        BUSY_TONE
 Description:  The modem detected busy tone on the line. Note that the
               proper IS-101 way to signal remote hangup is
               LOOP_BREAK.

 Event:        CALL_WAITING
 Description:  Defined in IS-101.

 Event:        DIAL_TONE
 Description:  The modem detected dial tone on the line.

 Event:        DATA_CALLING_TONE
 Description:  The modem detected data calling tone on the line.

 Event:        DATA_OR_FAX_DETECTED
 Description:  The modem detected data or fax calling tones on the line.

 Event:        FAX_CALLING_TONE
 Description:  The modem detected fax calling tone on the line.

 Event:        HANDSET_ON_HOOK
 Description:  Locally connected handset went on hook.

 Event:        HANDSET_OFF_HOOK
 Description:  Locally connected handset went off hook.

 Event:        LOOP_BREAK
 Description:  Defined in IS-101. This usually is for remote hangup.

 Event:        LOOP_POLARITY_CHANGE
 Description:  Defined in IS-101.

 Event:        NO_ANSWER
 Description:  After dialing the modem didn't detect answer for the time
               give in dial_timeout in voice.conf.

 Event:        NO_CARRIER
 Description:  The caller has hung up. For now, this event is detected only
               by the ISDN4Linux driver.

 Event:        NO_DIAL_TONE
 Description:  The modem didn't detect dial tone (make sure your modem is
               connected properly to your telephone company's line, or check
               the ATX command if dial tone in your system differs from
               the standard)

 Event:        NO_VOICE_ENERGY
 Description:  It means that the modem detected voice energy at the
               beginning of the session, but after that there was a
               period of XXX seconds of silence (XXX can be set with
               the AT+VSD command...at least on ZyXELs)

 Event:        RING_DETECTED
 Description:  The modem detected an incoming ring.

 Event:        RINGBACK_DETECTED
 Description:  The modem detected a ringback condition on the line.

 Event:        RECEIVE_DTMF
 Description:  The modem detected a dtmf code, you can get the code with an
               additional 'receive' command.

 Event:        SILENCE_DETECTED
 Description:  The modem detected that there was no voice energy at the
               beginning of the session and after that for XXX seconds.
               (XXX can be set with the AT+VSD command...at least on ZyXELs)

 Event:        SIT_TONE
 Description:  Defined in IS-101.

 Event:        TDD_DETECTED
 Description:  Defined in IS-101.

 Event:        VOICE_DETECTED
 Description:  The modem detected a voice signal on the line. IS-101 does
               not define, how the modem makes this decision, so be careful.

 Event:        UNKNOWN_EVENT
 Description:  None of the above :)



 7.)  Converting voice files to/from your modem's format


   Under directory 'pvftools' there are a number of useful utilities that
   can be used to convert voice files. Vgetty can only handle voice files
   converted to your modem's format. After recording, recorded files are
   in your modem's format. Converting a voice file (say 'wav') to rmd
   (raw modem format, that is the format your modem accepts data) takes
   two steps. First you have to convert your file to pvf (portable voice
   format) with ??topvf (?? = wav in our case), then use pvftormd with
   the proper arguments (see pvtormd -h for a help).
   The other way is similar (rmdtopvf , pvfto??).
   Browse through pvftools, there are other useful utilities there, too.



 8.) Logging

   You can set the desired log level in voice.conf, log path can be set
   in voice/include/paths.c (It defaults to /var/log).