You're over symplifying here. Good text to speech with real speech synthesis for example rsynth requires a fair bit of floating point to achieve. It would run a tad slugishly on my 486dx/33 with 80% cpu usage. I'd hardly call this minimal cpu cycles. What text to speech used to run on your 386? sbtalker? That's got more than one patent on it and we can't use their technology even if they would give it to us. Yes esound can mix audio streams but now we need to wait not only for the sound drivers to load but we'd better hope esound initializes properly or we get no speech. Note that mixing channels in software with esound is expensive cpu-wise, and we're allready burning cycles in kernel space for speakup and for the synthesis. Also note that we need to retain some degree of responsiveness with this system; how long is it going to take when we hit the shut up key for all the buffers to empty or for outgoing speech to be killed? Do we have a synthesis buffer that must also be flushed? I'm one of the first ones to say I want software speech for speakup one day, but let's firstly get in kernel speech (tuxtalk) working so we have a small free speech engine available for our use. Using IBM Viavoice isn't an option for everyone since there are lisencing restrictions. Regards, Kerry. On Sun, Jun 10, 2001 at 11:21:30PM +1200, Bruce Kingsbury wrote: > > > > 1: aAs far as I'm aware, software synthesys ties up a great deal of system > > > resources including the sound card unless of course you run a sb-live. > > How much is "a great deal" exactly..? software speech was possible on my > old 386 and even the most low-end pentium runs 50 times the speed of that! > tts takes only a tiny fraction of the CPU power required for speech > recongition or a winmodem. > > > Geoff pointed out that you can get a card with multiple streams. There are > > a number able to do this. But there are other considerations. > > Not a consideration at all, 'esd' will happily mix as many different > audio streams as you want and play them all through a simple SB16 or > onboard ES chip.. just make your TTS output through esd rather than > direct to /dev/dsp > > > > > _______________________________________________ > Speakup mailing list > Speakup at braille.uwo.ca > http://speech.braille.uwo.ca/mailman/listinfo/speakup -- -- Kerry Hoath: kerry at gotss.net alternatives: kerry at gotss.eu.org or kerry at gotss.spice.net.au ICQ UIN: 8226547
