Techniques for training and applying prosody models for speech synthesis are provided. A speech recognition engine processes audible speech to produce text annotated with prosody information. A prosody model is trained with this annotated text. After initial training, the model is applied during speech synthesis to generate speech with non-standard prosody from input text. Multiple prosody models can be used to represent different prosody styles.
G10L 17/00 - Speaker identification or verification techniques
G10L 21/00 - Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
G10L 13/10 - Prosody rules derived from textStress or intonation
G10L 15/18 - Speech classification or search using natural language modelling
G10L 13/08 - Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
G10L 15/06 - Creation of reference templatesTraining of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
2.
Systems and methods for automated media commentary
Techniques for providing automated media commentary are provided. A user agent requests audio commentary for media. In response, a service searches data sources to identify the specified media, finds information related to those entities, generates text that represents those information, combines the text into a textual monologue, and synthesizes speech audio from that textual monologue. The service selects relevant information to be likely unknown to the user while also being desired by the user.
Techniques for training and applying prosody models for speech synthesis are provided. A speech recognition engine processes audible speech to produce text annotated with prosody information. A prosody model is trained with this annotated text. After initial training, the model is applied during speech synthesis to generate speech with non-standard prosody from input text. Multiple prosody models can be used to represent different prosody styles.
G10L 13/08 - Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
G10L 13/00 - Speech synthesisText to speech systems
Methods evaluating propositions about timeseries are provided. A service accepts a proposition about one or more timeseries, and the service monitors those timeseries. A proposition can reference multiple timeseries with different time intervals and different units. When new data is available, the service evaluates the proposition to determine if the proposition is true, false, or neither. The service automatically performs unit conversions and selects relevant, previously observed values before evaluating a proposition when new data is available.
Techniques for training and applying prosody models for speech synthesis are provided. A speech recognition engine processes audible speech to produce text annotated with prosody information. A prosody model is trained with this annotated text. After initial training, the model is applied during speech synthesis to generate speech with non-standard prosody from input text. Multiple prosody models can be used to represent different prosody styles.
G10L 17/00 - Speaker identification or verification techniques
G10L 21/00 - Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
G10L 21/06 - Transformation of speech into a non-audible representation, e.g. speech visualisation or speech processing for tactile aids
G10L 19/14 - Details not provided for in groups ; G10L 19/06-G10L 19/12, e.g. gain coding, post filtering design or vocoder structure
6.
Collaborative, incremental specification of identities
Methods and systems for collaborative, incremental specification of identities are provided. Users of an information processing system collaborate to define and refine identities of entities, and users can create references for those identities. Relationships among the identifiers imply related keywords that can be used to improve search, navigation, and integration. Relationship factoring can be used to maintain efficient logical and physical representations.
An event processor accepts event representations that comprise perdurance and epistemological relationships among entities or entity references, and the event processor stores information derived from event representations in a datastore. A search component searches entities based on local epistemological and perdurance properties, and a perdurance engine traverses the perdurance relationships obtained from the datastore. An epistemological propagator generates derived epistemological relationships using the perdurance engine and local epistemological data obtained from the datastore.
Methods evaluating propositions about timeseries are provided. A service accepts a proposition about one or more timeseries, and the service monitors those timeseries. A proposition can reference multiple timeseries with different time intervals and different units. When new data is available, the service evaluates the proposition to determine if the proposition is true, false, or neither. The service automatically performs unit conversions and selects relevant, previously observed values before evaluating a proposition when new data is available.
Techniques for training and applying prosody models for speech synthesis are provided. A speech recognition engine processes audible speech to produce text annotated with prosody information. A prosody model is trained with this annotated text. After initial training, the model is applied during speech synthesis to generate speech with non-standard prosody from input text. Multiple prosody models can be used to represent different prosody styles.
G10L 13/08 - Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
G10L 13/00 - Speech synthesisText to speech systems
