![]() |
|
| Speech Synthesis | |
|
E-Speech offers a full-featured American English text-to-speech synthesizer, based on LPC and RELP-encoded demisyllable sound inventories, for unrestricted text applications. We also offer custom speech generation systems based on concatenation of large speech units, providing near-natural speech quality for limited domain applications, such as names, addresses, or timetables. For information about our custom systems, visit our Custom Products page. The E-Speech synthesizer provides the highest quality text-to-speech for American English, with a footprint less than 3MB. Because of its superior technology, the E-Speech synthesizer allows high quality, highly accurate access to database-driven information services either over the telephone or directly from a computer. The E-Speech synthesizer's features include:
Applications. The E-Speech synthesizer can be used to read email over the phone, to speak driving directions and traffic reports, or to speak names and addresses for applications like Directory Assistance, Yellow Pages, and Automated Attendant. The E-Speech synthesizer makes an excellent companion for automatic speech recognition (ASR) or speaker verification applications. Also, the E-Speech synthesizer can be used to build name pronunciation dictionaries of unparalleled accuracy, which can be incorporated into ASR software. Platforms. The E-Speech synthesizer version 1.2 is available for Unix (Solaris, Solaris x86, SunOS, DECstation, Linux, FreeBSD, BSDI, HP/UX, Motorola, RS6000 AIX) and Windows NT/95/98/2000/NT/XP/Vista. Arrangements can be made for ports to other operating systems or platforms. |
| © 1998-2010 E-Speech Corporation, Princeton, NJ. All Rights Reserved. |