Translation system and method of the voice signal in the transcriptional representation with metadata.
It refers to the technology of speech recognition, ie voice signal transfer, in transcriptional representation consisting of the characters of individual sounds, words, transcription, proposals and their metadata visualization of the physical characteristics of the speech signal, create the sensation of of speech sounds. Evaluating of pronunciation for training system of speech pronunciation.
Method of speech translation in transcriptional presentation with metadata, which allows to recognize speech sounds with high accuracy in real time, regardless of the speaker's voice, his individual characteristics, environmental noise, the specific features of different language systems. Based on the measurement of only the physical characteristics of the acoustic signal, which directly creates the feeling of the sounds of language.
Recognition sounds of voice , their visualization, pronunciation evaluation of sounds, words concerning the standard voice of the respective language to learn the correct pronunciation. Education of deaf for speaking.
The technology allows to select sounds of speech, to base the authors' theory spectral bands. This characteristics power and dynamic of the spectral signal of voice, wich create a sense speech sounds.
04-Aug-2015
2014134745/08(056276)
PCT/RU2015/000459
Translating an voice signal into transcriptional representation with the metadata. Visualization of the physical characteristics of the voice signal, create the sensation of of speech sounds. Pronunciation accuracy assessment included for pronunciation training system.
Russia
