Beijing Infoquick Sinovoice Speech Technology Corp.

 

Solution=>|   Profession Level for CTI  |   Embedded Level Solution  |   for Other Fields  |

 

Core Tech

jTTS, the core technology of SinoVoice, uses the large scale recorded voice library and algorithm based on hierarchical prosody structure matching. The large scale recorded voice library covers various articulation situation almost in all kinds of context, and the algorithm based on hierarchical prosody structure matching ensures that the system can choose the most proper original speech fragment for jointing. Hence, the Chinese speech generated from jTTS technology is much better in such aspects as clearness and naturalness than those from other systems.

Features of jTTS v4.0

Based on jTTS v3.0, jTTS v4.0 has some new functions as mentioned below:

•  Supporting multilingual TTS as well as setting of multi fields;

•  Supporting the search, access and uploading of timbres;

•  Supporting SSML (Speech Synthesis Markup Language);

•  Direct play of text files as well as supporting plus DLL, which can draw out text, for supporting other formats;

•  Supporting various character sets such as GB (including GB2312, GBK, GB18030), Big5, Shift-JIS, ISO-8859-1, Unicode, Unicode Big Endian, and UTF8; automatic recognition of Unicode text with the mark of BOM;

•  Supporting synchronous and asynchronous synthesis of one Session. Transmit data to users by passive callback function in addition to the active acquisition of speech data in the original version;

•  Supporting plus speech format conversion program.

 

The Chinese engine has some new features as mentioned below:

•  Supporting the mixed reading of Chinese and English of the same timbre;

•  Brand new include file analysis and analysis algorithm of hierarchical words which has enhanced greatly the naturalness of words and sentences;

•  Improved analysis algorithm of polyphonic characters;

•  Improved automatic judgment algorithm of figures and symbols articulation; and providing definition and use of external rules of text analysis;

•  Available for multi fields, that is, it provides options of different fields in order to achieve the best synthesis effect through customization;

•  Supporting SSML;

•  Improved resource-maintaining tool of jTTSLib.

>> Features of jTTS v3.0






Copyright@ 2001-2004,beijing InfoQuick SinoVoice speech technology Corp. All Rights Reserved.