DEVELOPMENT OF A DATA-TO-SPEECH CONVERTER
Abstract
A method is presented for converting sequence of digital information directly into corresponding sequences of spoken words. The long range design objective for this device is to operate on command of a computer at speech rates and fidelity approximating natural speech with a 10,000 to 50,000 word vocabulary. A design study was made, an optimal approach selected, and a feasibility model constructed. This model contains a randomACCESS PHOTO-OPTICAL MEMORY WHICH STORES 500 PRERECORDED WORDS ON FILM FOR PHOTO-ELECTRIC READOUT ON COMMAND FROM A MANUAL KEYBOARD OR PAPER TAPE. The specific design configuration is demonstrably extendable to a full size vocabulary at acceptable data rates and fidelity. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 27, 1962
- Accession Number
- AD0290679
Entities
People
- Harold Johnson