





|
TTS (Text-To-Speech) Technology
After much research and hard work, ECTACO, Inc. is ready to unveil its new Text-To-Speech (TTS) technology for a number of added languages. We invite companies and institutions to learn more about this innovative technology and to make use of it in their hardware and software applications. Interested parties should contact us for more information.
The TTS engine now supports 17 new languages, as outlined below. It allows the pronunciation of any word or phrase written in a format understandable by the system. Initially, only English was used, but the universality of the technology has allowed us to broaden the range of available languages significantly.
The engine is coded in ANSI C, which makes for a high degree of portability to many different platforms. The system can be supplemented with additional dictionaries and can be fine-tuned to any thematic vocabulary.
The engine utilizes formant synthesis at the phoneme and demisyllable level. Sound is generated in three sample rates: 8, 10, and 11 kHz. More technical details are given below.
TTS technology features |
Languages supported
|
English, French, Spanish, German, Japanese,Italian, Russian, Portuguese, Dutch, Arabic, Korean, Mandarin Chinese, Malay, Hebrew, Cantonese Chinese, Greek
|
Quality reproduction of the following is stipulated
|
Phrases
Words
Numbers
Separate letters of the alphabet
|
System requirements
|
Processor productivity - less than 10 MIPS
Less than RAM 300 Kbytes
|
Application size
|
Less than 500 Kbytes (including DB)
|
Quality attributes
|
Natural tone quality
Parametric representation of voice timbre
Real time synthesis
System functionality is represented by playback of phrases through a built-in speaker in real time (i.e. without interruption) starting not later than 0.5 seconds after the start of the synthesis function.Rate of accuracy: not less than 99%
|
Procedure for qualitative analysis of characteristics
|
A text consisting of 100 words and 10 phrases is played back for three listeners. Accuracy of synthesis is defined as the average percentage of words and phrases recognized.
|
Please contact us for more information about applying this technology in your company's business.
|
 |