Yes, text2wav as well as 'festival --tts' seem to pick up the Festival 
default voice.  I've done some more playing and it seems that sound_play 
sometimes gets it right and sometimes not.  For example, the kal_diphone 
voice is the one that sounds very low (bass) and slow.  But sound_play 
does OK with don_diphone and rab_diphone even though they don't sound 
exactly like they do when playing through Festival direclty.  I also 
tried the the Finnish voice suo_fi_lj_diphone which plays fine in 
Festival but rather low and slow like kal_diphone when using sound_play.

And thanks Moritiz for the pointer to other voices.  For now I just want 
something that can be understood and won't scare the cats. ;-)  Either 
don_diphone or rab_diphone should do the trick.

--patrick

> Just out of curiosity, does text2wav pick up your Festival default
> voice? (That's what's being used in soundplay_node.)