The TalkingArch script that determines the audio device to use is heavily dependent on alsa functionality, so it may not be possible to add non snd_* devices to try to make them speak, as that is an entirely different subsystem, non-alsa related. Sent from under the cherry moon