About Us
 











Talktronics, Inc. - Speech Products

HISTORICAL PERSPECTIVE -- 1984
Beginning in 1984, Talktronics produced Text-To-Speech (TTS) synthesizers for the Commodore VIC-20, C64 and C128 personal computers.  These synthesizers were in the form of plug-in modules that installed in the computer's "expansion port".  At the time, the features of these TTS products were quite sophisticated for the meager price tag ( USD $89 ! ).  The Commodore computers were based on the Rockwell R6502 8-bit CMOS microprocessor  family, and used a version of the CPU designated '6510', and made by MOS Technology.  The TTS synthesizers used a "phoneme generator" chip from Silicon Systems, designated SSI263.  A reprint of the original data sheet for the "VIC-Talker" appears below.

VIC TALKER

  • Advanced English text-to-speech rulesimg1.gif (32151 bytes)
  • Numeric, decimal and simple fraction rules
  • Rules for dollars and cents
  • Two user-definable voices
  • Automatic or direct intonation control
  • Text, proofreading, spell & say and spell-only modes
  • User-programmable exceptions memory
VIC-TALKER  provides unlimited vocabulary translation of text to synthesized speech using advanced English language pronunciation rules and a user-expandable exception memory. The cartridge incorporates an internal audio mixer to blend the synthesized voice output with sounds generated by the "VIC" sound chip for simultaneous sound effects with voice.

The English text to speech rules have been developed to maximize performance of the phoneme synthesizer chip, producing an intelligible synthesized voice from unedited text.  In addition to the over 360 English rules, interpretation of numeric strings involving decimal fractions, simple fractions, dollars and cents is accomplished to permit proper voice synthesis of virtually any word processor output.  BASIC commands such as GOTO, GOSUB and PRINT# are translated to allow easy interpretation of BASIC program listings. 

Some sample translations:

INPUT TEXT 

VIC TALKER SAYS

126933

one hundred twenty six thousand nine hundred thirty three

12.785

twelve point seven eight five

$37.95

thirty seven dollars and ninety five cents

$.08

eight cents

(949) 555-1212

area code nine four nine (pause) five five five one two one two

1/2

one half

13/64

thirteen sixty fourths

34A785XXX

three four A seven eight five X X X

GOTO 45

go to forty five

PRINT #3, "My name is VIC"

print on three my name is vic


Special modes are provided to define and select two different "voices", either one to be used on a word by word basis.  In a "proofreading" mode, the second voice is used to call out punctuation and symbols.  The primary voice is then used to synthesize the words and numbers in the text.  The VIC-TALKER may also be programmed to work in "spell and say" or "spell-only" modes.  Changing the two pre-programmed voices is extremely easy for the user, and once set, each of the two voices can be selected by control commands. 

VIC-TALKER
uses sentence-weighted intonation to provide a variable pitch voice to emphasize statements, questions, and exclamations.  In the monotonic mode, intonation is controlled directly by commands on a word by word or letter by letter basis.  Using direct control of pitch, the user can synthesize a singing voice, with the range of seven octaves on an even-tempered scale.

Programming of the user-defined exception words is easily accomplished in the "learning" mode.  The user commands VIC-TALKER to produce the desired pronunciation of a special word by sending the word followed by a phonetic spelling of the word.  The letter string is pronounced according to the English rules, producing an output phoneme string for the synthesizer chip.  This output string is then stored, upon user command, in the exception dictionary, along with the original word.  In this way, personal names, city and street names, or any otherwise special translations of words will subsequently produce the desired pronunciation.  The exception dictionary resides in EEPROM (electrically erasable programmable read-only memory), such that the exception dictionary is never lost during power-down, and becomes a permanent extension of the internal rules.  Facilities are provided to edit the exceptions and to prevent redundant storage.

Home   About Talktronics   Imaging Products    Design Services   Speech Products   Request Information/Get A Catalog   Contact Us

Copyright 2007 Talktronics, Inc., 23400 Peralta Drive, Ste. D, Laguna Hills, CA 92653 USA.  
All rights reserved.
    Last Updated:  12/7/07