VitalVoice meets the basic requirement of users to speech synthesis system: it allows you to voice any (even non-standard) text (SMS, emails, online forums, etc.) so that the listener has the impression that he hears natural human voice.
Text can be read in different synthesized voices. Every voice is based on the use of speaker’s speech base (volume of about 10 hours of speech), marked on 9 levels, including textual interpretation, counting the words, syllables, allophones, pause, markers of words’ and phrasal stresses, types of intonation, non-speech events and other phonetic phenomena.
For the correct intonation and determining the place of stress in words, the powerful module of automated processing of Russian text has been developed, using morphological, syntactic and semantic types of analysis. So «VitalVoice» is the unique technology of Russian speech synthesis due to this module.
Advantages:
- High quality and natural sounding of any text
- Taking into account phonetic, morphological and grammatical peculiarities of the Russian language
- Technology of natural intonation cloning
- Proper placement of accents
- Proper explanation of abbreviations, numbers and special characters
- Simplicity of use and implementation
- Support of standard data exchange protocols and markup languages (MRCP, SAPI, SSML)
- User dictionary
- Ability to change the pitch of the voice and speech rate in wide range
Functional features:
- Explanation of standard abbreviations using semantic analysis (Minsk, Brest, Vitebsk, 2010, 145)
- Correct reading of abbreviations (State Auto Inspection, BSU)
- Explanation of dates, time, correct reading of numbers (02/26/2010, 10:40)
- Explanation of special marks ($ 20, house number 7)
- Correct interpretation of formulas (2 * 3 = 6)
- Withdrawal of homographs (correct pronunciation of words with different meanings and the same spelling)
- 8 different voices of speech synthesizer
- Ability to change the pitch of the voice and speech rate in wide range
- The rate of formation of sound file is 10-12 times higher than actual time
Technical characteristics:
- The format of input data: txt, doc, rtf
- Output data format: wav, mp3
- Wav-file format: 22050 Hz sampling frequency, bit rate of 16, PCM, mono
