The thing about the shutup key was a miscommunication between me and one of my parteners when we were discussing what might be wrong. Apparently what happened is that Speakup clobbers certain characters and right before we published our code we changed the SYNTH_CLEAR_CHAR to one of these values as opposed to '!' which we had been using for debugging reasons. Our website now has a link to an updated version of the code which should resolve the shutup problem. The new version should also amoriolate the delay problem because our middleware program now sends the data to the TTS program in one-line chunks instead of the previous one-word chunks, a change that also should improve the intonation of the output. However this also means that, at present, the program will only shutup at the end of a line. A few other bugs were also fixed in this release, mostly related to dealing with full buffers. Please give the new code a try and let us know if it helps to fix the problems. As notes our code should be voice/ALSA/OSS independant and although there is a slightly newer version of festival available (1.4.3) that should not be an issue. As to including Mbrola it is on our list although since it does not have full TTS capability but only phoneme-to-sound an additional layer of software will be required so it may be a while. HtH Owen On Wed, 2 Apr 2003, Jacob Schmude wrote: > Hi Well, keypad minus says "parked/unparked" just like it used to. Also, > keypad enter was the shut-up key for my doubletalk which I ran this same > CVS on until it quit a few days ago, so I don't see how the shut-up key > is keypad minus. In any case it still won't shut up at all, even when > pressing other keys with interrupt enabled. Nothing shuts it up. Not > sure if it's relevant, but I'm using the ALSA drivers as the > kernel-packaged drivers do not support my chip. Festival works fine with > the alsa drivers, but you never know, I'm using it through the OSS > emulation. Also I've noticed that before starting to speak something > new, I here the letter "f". This happens whenever I initiate speech or > when speakup starts talking on its own, it does not happen between > words. I'm running festival 1.4.2 which seems to be the latest version. > Is there a newer version I should be getting? Also, should it matter > what voice I use (not that it made a difference on this system)? I > rather like the mbrola_us2 voice and so have festival running that. But > the regular US english voice (kal16K) produced the same results. Out of > curiosity, is mbrola one of the synths on the todo list? It's smaller > than festival (not as small as flite) but has better sound quality. On
