Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!bbn!gatech!ncar!stout!cook
From: cook@stout.ucar.edu (Forrest Cook)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Speech Synthesis / Voice Recognition
Keywords: Hearsay 1000
Message-ID: <1391@ncar.ucar.edu>
Date: 11 Feb 89 01:32:33 GMT
References: <449@oucsace.cs.OHIOU.EDU>
Sender: news@ncar.ucar.edu
Reply-To: cook@stout.UCAR.EDU (Forrest Cook)
Organization: Field Observing Facility, NCAR, Boulder, CO
Lines: 27

In article <449@oucsace.cs.OHIOU.EDU> thollowe@oucsace.cs.OHIOU.EDU (Tom Hollowell) writes:
>What I want to know is are there any _good_ sounding synthesizers,
>and how good does the recognition work.

On the synthesis side:  Heathkit (Griefkit :-)  ) sells a PC phoneme synth
board for under $100.  It is based on the SP0256 chip that was featured in
one of Ciarcia's Byte articles a few years ago.  I have not heard the chip
but it probably sounds similar to the Votrax SC-01.
If anybody out there has bought one of these and has made it talk with
Turbo C, I would be very interested in hearing about it.
I don't know if Heath provides any interface routines, but driving the
board should be fairly simple.

Plusses of Phoneme Synthesis:
- Low data rate < 50 B/Sec.
- Easy to write English to Phoneme translator programs.
  (one was posted to the net a few years ago, it is ~probably~ on SIMTEL20)

Minuses:
- VERY mechanical sounding output,  It takes a while to learn its dialect so
  it is not very good for dealing with the general public.
- Usually only one voice type on the chip.

 ^	Forrest Cook   (The preceding comments were my Opinions)	 ^
/|\	cook@stout.ucar.edu						/|\
/|\	{husc6 | rutgers | ames | gatech}!ncar!stout!cook		/|\
/|\	{uunet | ucbvax | allegra | cbosgd}!nbires!ncar!stout!cook	/|\