kevinj@boulder.Colorado.EDU (Kevin M. Jackson) (09/27/89)
I'm trying to decide on a chip to use for a voice synthesis project I'm
about to undertake. Does anyone have a suggestion regarding some good
quality voice synthesis chips?
I've looked into the NS DigiTalker chip set, but it is a vocabulary based
chip, which makes it a poor choice for my application (I'd like to
have an unlimited vocabulary). Other possibilities I've looked at are
chips by Hitachi and Synergetics.
I am also looking for a good phoneme analysis algorithm (or a short
sample program). If anyone has such an algorithm (even if it's
chip specific), I'd be interested in seeing it.
Please send me mail regarding this. If there's interest, I'll post
a summary of responses to the net.
Thanks,
kevinj (Kevin Jackson)
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
// o
This space intentionally left (( -|- Kevin M. Jackson
blank | \\ //
| -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
V \\ kevinj@boulder.Colorado.EDU
)) kevinj@matrix.Colorado.EDU
//
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
// o
This space intentionally left (( -|- Kevin M. Jackson
blank | \\ //
gbell@pnet12.cts.com (Greg Bell) (09/28/89)
kevinj@boulder.Colorado.EDU (Kevin M. Jackson) writes: >I'm trying to decide on a chip to use for a voice synthesis project I'm >about to undertake. Does anyone have a suggestion regarding some good >quality voice synthesis chips? > This topic has been hashed many-a-time, but its always worth reviving, because sometimes people have new ideas, or new chips have come out. If you need unlimited vocabulary, you have two choices: Go with one of the phoneme synthesizers like GI's SPO256 or Votrax. I've heard of people talk about an SC-01, but I'm not sure who makes it. The main advantages of this approach are: unlimited vocabulary, and low data rate. Disadvantages: sppech is robotic sounding and sometimes hard to understand. (Oh yeah, SSI used to make one... the SSI-263, but now Artic technologies sells the chip). Or, you can record your own speech (digitizing) and have exactly the vocabulary you need. Advantages: excellent reproduction quality. Disadvantages: high data rate. OKI semiconductor makes several ADPCM chips that both digitize and play back ADPCM data. But, due to the high data rate, you need a large amount of RAM. See the June '83 Circuit Cellar in Byte and the December '88 issue of QST for projects using OKI chips. Another possibility for digitizing is by making your own circuit to do this. I have a delta modulation circuit that uses a few op-amps and works well. By the way, OKI has a demo of their MSM6258 by phone: 800-832-6654 ext. 462. Greg Bell_________________________________________________________ Hardware hacker | Electronics hobbyist | UUCP: uunet!serene!pnet12!gbell EE major at UC San Diego |
elliott@optilink.UUCP (Paul Elliott x225) (09/29/89)
In article <976@serene.UUCP>, gbell@pnet12.cts.com (Greg Bell) writes: > > Go with one of the phoneme synthesizers like GI's SPO256 or Votrax. I've > heard of people talk about an SC-01, but I'm not sure who makes it...... I recall that the SC-01 is (was?) made by Votrax, and was their first I.C. systhesizer. Prior to that they made potted modules, using op-amp filters, sources, etc. In a previous incarnation, I was an engineer at a company that made speech synthesizers for the speech-impaired (pretty neat products for their time, if I say so myself). We started with the Votrax modules, and used the SC-01 when it came out, about 7 years ago. Votrax provided an impressive text-to-speech program, running on a P.C., for use with the SC-01. I have no idea what the current state of affairs re: Votrax is, but could make a few phone calls if anyone is interested. ... Paul -- Paul M. Elliott Optilink Corporation (707) 795-9444 {pyramid,pixar,tekbspa}!optilink!elliott "I used to think I was indecisive, but now I'm not so sure."
fritz@mit-caf.MIT.EDU (Frederick Herrmann) (10/01/89)
In article <2427@optilink.UUCP> elliott@optilink.UUCP (Paul Elliott x225) writes: >I recall that the SC-01 is (was?) made by Votrax, and was their first >I.C. systhesizer. Prior to that they made potted modules, using op-amp > >Votrax provided an impressive text-to-speech program, running on a P.C., >for use with the SC-01. I have no idea what the current state of affairs NTIS also distributes a text-to-speech program for Votrax synthesizers, developed at NRL. I believe the principal author's name was Honey Sue Elovitz. I can look up the reference if anybody cares. The program worked pretty well, the only catch was that the source code was in SNOBOL. I converted it to 8088 assembler and defined it as an MS-DOS device: C> echo Hello > voice: I'd be happy to send the source code to anyone who asks, but it was pretty specific to a homebrew project based on a Ciarcia article in Byte. - Fritz fritz@caf.mit.edu