home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.multimedia
- Path: sparky!uunet!usc!cs.utexas.edu!uwm.edu!ux1.cso.uiuc.edu!news.iastate.edu!barrett
- From: barrett@iastate.edu (Marc N. Barrett)
- Subject: Re: C= : Where's the Speech?
- Message-ID: <Bz48IL.2su@news.iastate.edu>
- Sender: news@news.iastate.edu (USENET News System)
- Organization: Iowa State University, Ames, IA
- References: <Bz1v4z.4qI@undergrad.math.waterloo.edu> <Bz276t.n1x@news.iastate.edu> <57432@dime.cs.umass.edu>
- Date: Fri, 11 Dec 1992 22:15:54 GMT
- Lines: 78
-
- In article <57432@dime.cs.umass.edu> barrett@astro.cs.umass.edu (Daniel Barrett) writes:
- >
- >[Followups are to c.s.a.advocacy.]
- >
- >>>In article <Bz09KH.HDp@news.iastate.edu> barrett@iastate.edu (Marc N. Barrett) writes:
- >>>> Why would a DSP be necessary for something as simple as speech synthesis?
- >
- > Good speech synthesis is *not* simple. You don't know what you are
- >talking about. The synthesized speech from personal computers would never
- >fool anyone into thinking it was a real person talking. Maybe that's
- >because it's HARD to create convincing computer voices, eh? (Digital
- >sampling doesn't count.)
- >
- >>In article <Bz1v4z.4qI@undergrad.math.waterloo.edu> pejanes@undergrad.math.waterloo.edu (Peter Janes) writes:
- >>>At a recent U of W users group meeting, a Commodore rep hinted that an
- >>>improved in-house version is being written--one that may take advantage of
- >>>a (then-unannounced) DSP card.
- >
- >In article <Bz276t.n1x@news.iastate.edu> barrett@iastate.edu (Marc N. Barrett) writes:
- >> I still think that would be a stupid move. That would mean that only a
- >>tiny number of Amigas with a DSP would be supported with speech...
- >
- > Marc once again has jumped to an illogical conclusion based on his
- >own paranoia and fantasies. (No surprise.) Just because speech synthesis
- >may "take advantage of... a DSP card" doesn't mean that it won't work on
- >a stock Amiga, idiot.
-
- Then why has Commodore removed the speech capability completely from
- AmigaOS 2.1 and AmigaOS 3.0? The OS additions to support the DSP are not
- planned until AmigaOS 3.1. Logically, if routines were planned to support
- speech with both a DSP and the standard hardware, they would not have changed
- anything with respect to the existing routines until AmigaOS 3.1. That they
- removed it sooner seems to support the idea that from now on speech will only
- be supported with a DSP.
-
- >>With 2.0.4, the speech capability even on an A500 or A1000 is already quite
- >>good. The improvements in moving to a DSP would not be all that great;
- >
- > Marc, you clearly know absolutely nothing about speech synthesis.
- >Don't you think that having over 250 times the sample resolution and
- >who-knows-how-many-times the processing power will make a difference?!?
- >Have you ever compared an Ensoniq Mirage with an EPS-16+? The difference
- >is *staggering*.
-
- Your problem is that you pay too much attention to raw numbers, and too
- little attention to reality. Sure, a DSP can blow away the standard hardware
- in creating speech, ON PAPER. Run the standard 2.0 'say' program sometime
- (you may have to reinstall the 2.0.4 speech libraries to do it), and listen to
- how good the speech is. Sure, it is fairly obviously a computer, but it is
- also quite good and VERY understandable. With some tweeks (using
- SpeechToy 2.0), the speech can be made even better.
-
- In other words, the speech is good enough already that I just don't see all
- that much room for improvement. You can install an array of DSPs or even an
- array of Crays all dedicated to doing speech synthesis, and the speech would
- not be improved all that much.
-
- >>...the speech libraries could be written something like the math libraries;
- >>if enhanced hardware is present, software supporting the enhanced hardware
- >>would be automatically kicked in.
- >
- > The first intelligent thing you've said all day. Your paranoid
- >conclusion that a DSP would be required is total speculation. It's
- >what is known as a "straw man" argument. If you build up something out
- >of nothing, it's easy to kick it down.
-
- Kick it down if you want, but it is the most reasonable explanation to date
- for why Commodore did not bother updating their licesnse to keep the speech
- synthesis libraries in AmigaOS 2.1 and 3.0.
-
- > //////////////////////////////////////\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
- >| Dan Barrett -- Dept of Computer Science, Lederle Graduate Research Center |
- >| University of Massachusetts, Amherst, MA 01003 -- barrett@cs.umass.edu |
- > \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\/////////////////////////////////////
-
- ---
- | Marc Barrett -MB- | email: barrett@iastate.edu
- --------------------------------------------------
-