dectalk will read text from stdin in typing-mode (default),
	non-typing mode or command line or file and speak that
	text through the MME server using the highest quality audio
	device it finds. You can override this option by using the -d 
	command line option to select a particular audio device.

Usage: dectalk [-h] [-s #] [-r #] [-d #] [-fo file] [-a "<text>"] [-fi file] [-]
          -a "text" speak text (quoted, multiple instances supported)
          -d  #      Select the audio device ( 0 or 1 )
          -e  #      Output wave file encoding; specify:
                       1   for PCM, 16 bit Mono 11KHz format
                       2   for PCM,  8 bit Mono 11KHz format
                       3   for MULAW 8 bit Mono  8KHz format
          -fo file   Output wave file name, default: dtmemory.wav
                     Also supports Sun .au and two virtual stdout streams:
                     stdout:au Sun au with unknown length
                     stdout:raw Raw PCM or MULAW samples in host byte order
          -h         This help message
          -r  #      Speaking rate ( 75 - 600 )
          -s  #      Speaker number (0-9)
          -v  #      Volume (0-100)
          -fi file   Speak from a specified text file
          -l lang    Use specific language (us,uk,gr,sp,la,fr)
          -pre text  Text to be passed to DECtalk before the normal input.
                     This is useful for passing initializing commands to
                     DECtalk that would normally not be part of the input.
                     If the prefix text has spaces, it must be enclosed in
                     quotes.  An example would be "[:phoneme on]" or
                     "[:nb :ra200]".
                     The prefix text is "forced" out before the input text
                     is read.
          -post text Text to be passed to DECtalk after the normal input.
                     This is useful for passing terminating commands to
                     DECtalk that would normally not be part of the input.
                     If the postfix text has spaces, it must be enclosed
                     in quotes.  An example would be "[:phoneme off]" or
                     "The End".
                     The "normal" input is "forced" out before the postfix
                     text is read.
          -c         Switch to clause-mode from line-mode (for stdin - and -fi)
          -          Read from stdin in line- or clause-mode (non-typing)

Modes:
 * typing: Fast pronounciation of single letters on a line.
           Playing output is aborted when a new input-line is sent.
 * line:   Output is produced/queued after each newline.
           Playing output continues.
 * clause: Output is produced/queued on clause end (with delimiters like
           period (.) or comma (,)) or when the force-character ^K is
           received.
           Playing output continues. Output is always produced on EOF.

Input/mode switch table:
       | typing | line | clause
 ------+--------+------+-------
 stdin |        | -    | -c -
 ------+--------+------+-------
 file  | N/A    | -fi  | -c -fi
 ------+--------+------+-------
 cli   | N/A    | N/A  | -a