This is a collection of notes on Speech Dispatcher, a place to gather knowledge about Speech Dispatcher which can later be transfered into proper guides, manuals or technical documents.

Nothing special needs to be done, but one must be careful not to use audio fallback. It is necessary to have only one single audio output system specified in configuration.

Users can but do not have to setup their user-specific configurations using the spd-conf tool.

The following setup has been suggested by Halim Sahim:

  1. disable autospawn
  2. Keep communication method at unix_socket
  3. Set socketfile to var/run/speech-dispatcher/speechd.sock
  4. Add user speech-dispatcher
  5. Add the user to group audio.
  6. Insert in etc/profile: SPEECHD_ADDRESS=unix_socket:/var/run/speech-dispatcher/speechd.sock

We have further talked about a dynamic script in etc/profile.d which sets SPEECHD_ADDRESS accordingly based on whether a system Speech Dispatcher is running or not.