Method and system for training a text-to-speech synthesis system using a domain-specific speech database
AssigneeAT&T Intellectual Property II, L.P.
InventorHorst Juergen Schroeter
Life of this patent
Estimating the term…
Abstract
A system, method and computer readable medium that trains a text-to-speech synthesis system for use in speech synthesis is disclosed. The method may include recording audio files of one or more live voices speaking language used in a specific domain, the audio files being recorded using various prosodies, storing the recorded audio files in a speech database; and training a text-to-speech synthesis system using the speech database, wherein the text-to-speech synthesis system selects audio selects audio segments having a prosody based on at least one dialog state and one speech act.
Claims20
A method for training a text-to-speech synthesis system, comprising:
- recording audio files of one or more live voices speaking language used in a specific domain, the audio files being recorded using various prosodies;
- storing the recorded audio files in a speech database;
- and
- training a text-to-speech synthesis system using the speech database, wherein the text-to-speech synthesis system selects audio segments having a prosody based on at least one dialog state and one speech act.
A computer-readable storage medium storing instructions for controlling a computing device for use in training a text-to-speech speech synthesis, the instructions comprising:
- recording audio files of one or more live voices speaking language used in a specific domain, the audio files being recorded using various prosodies;
- storing the recorded audio files in a speech database;
- and
- training a text-to-speech synthesis system using the speech database, wherein the text-to-speech synthesis system selects audio segments having a prosody based on at least one dialog state and one speech act.
A text-to-speech synthesis training system, comprising:
- a speech database;
- and
- a specific domain speech knowledge module that record audio files of one or more live voices speaking language used in a specific domain, the audio files being recorded using various prosodies, stores the recorded audio files in the speech database, and trains a text-to-speech synthesis system using the speech database, wherein the text-to-speech synthesis system selects audio segments having a prosody based on at least one dialog state and one speech act.
Dependencies are inferred from the claim text.