Techniques for the generation of dubbed audio for an audio/video are described. An exemplary approach is to receive a request to generate dubbed speech for an audio/visual file; and in response to the request to: extract speech segments from an audio track of the audio/visual file associated with identified speakers; translate the extracted speech segments into a target language; determine a machine learning model per identified speaker, the trained machine learning models to be used to generate a spoken version of the translated, extracted speech segments based on the identified speaker; generate, per translated, extracted speech segment, a spoken version of the translated, extracted speech segments using a trained machine learning model that corresponds to the identified speaker of the translated, extracted speech segment and prosody information for the extracted speech segments; and replace the extracted speech segments from the audio track of the audio/visual file with the spoken versions spoken version of the translated, extracted speech segments to generate a modified audio track.

A method (400) for representing an intended prosody in synthesized speech includes receiving a text utterance (310) having at least one word (240), and selecting an utterance embedding (204) for the text utterance. Each word in the text utterance has at least one syllable (230) and each syllable has at least one phoneme (220). The utterance embedding represents an intended prosody. For each syllable, using the selected utterance embedding, the method also includes: predicting a duration (238) of the syllable by decoding a prosodic syllable embedding (232, 234) for the syllable based on attention by an attention mechanism (340) to linguistic features (222) of each phoneme of the syllable and generating a plurality of fixed-length predicted frames (260) based on the predicted duration for the syllable.

A method (400) for representing an intended prosody in synthesized speech includes receiving a text utterance (310) having at least one word (240), and selecting an utterance embedding (204) for the text utterance. Each word in the text utterance has at least one syllable (230) and each syllable has at least one phoneme (220). The utterance embedding represents an intended prosody. For each syllable, using the selected utterance embedding, the method also includes: predicting a duration (238) of the syllable by decoding a prosodic syllable embedding (232, 234) for the syllable based on attention by an attention mechanism (340) to linguistic features (222) of each phoneme of the syllable and generating a plurality of fixed-length predicted frames (260) based on the predicted duration for the syllable.

Technical solution relates to the fields of voice technologies and knowledge graph technologies. A technical solution includes: acquiring script matched with a scenario from a speech package, and acquiring a broadcast template configured for the scenario in advance; and filling the broadcast template with the script to generate the broadcast speech.

Technical solution relates to the fields of voice technologies and knowledge graph technologies. A technical solution includes: acquiring script matched with a scenario from a speech package, and acquiring a broadcast template configured for the scenario in advance; and filling the broadcast template with the script to generate the broadcast speech.

A speech recognition module receives training data of speech and creates a representation for individual words, non-words, phonemes, and any combination. A set of speech processing detectors analyze the training data of speech from humans communicating. The set of speech processing detectors detect speech parameters that are indicative of paralinguistic effects on top of enunciated words, phonemes, and non-words in the audio stream. One or more machine learning models undergo supervised machine learning on their neural network to train on how to associate one or more mark-up markers with a textual representation, for each individual word, individual non-word, individual phoneme, and any combinations of these, that was enunciated with a particular paralinguistic effect. Each mark-up marker can correspond to its own paralinguistic effect.

A speech recognition module receives training data of speech and creates a representation for individual words, non-words, phonemes, and any combination. A set of speech processing detectors analyze the training data of speech from humans communicating. The set of speech processing detectors detect speech parameters that are indicative of paralinguistic effects on top of enunciated words, phonemes, and non-words in the audio stream. One or more machine learning models undergo supervised machine learning on their neural network to train on how to associate one or more mark-up markers with a textual representation, for each individual word, individual non-word, individual phoneme, and any combinations of these, that was enunciated with a particular paralinguistic effect. Each mark-up marker can correspond to its own paralinguistic effect.

A method for controlling an electronic device includes obtaining a text, obtaining, by inputting the text into a first neural network model, acoustic feature information corresponding to the text and alignment information in which each frame of the acoustic feature information is matched with each phoneme included in the text, identifying an utterance speed of the acoustic feature information based on the alignment information, identifying a reference utterance speed for each phoneme included in the acoustic feature information based on the text and the acoustic feature information, obtaining utterance speed adjustment information based on the utterance speed of the acoustic feature information and the reference utterance speed for each phoneme, and obtaining, based on the utterance speed adjustment information, speech data corresponding to the text by inputting the acoustic feature information into a second neural network model.

A method for controlling an electronic device includes obtaining a text, obtaining, by inputting the text into a first neural network model, acoustic feature information corresponding to the text and alignment information in which each frame of the acoustic feature information is matched with each phoneme included in the text, identifying an utterance speed of the acoustic feature information based on the alignment information, identifying a reference utterance speed for each phoneme included in the acoustic feature information based on the text and the acoustic feature information, obtaining utterance speed adjustment information based on the utterance speed of the acoustic feature information and the reference utterance speed for each phoneme, and obtaining, based on the utterance speed adjustment information, speech data corresponding to the text by inputting the acoustic feature information into a second neural network model.

A speech to text system includes a text and labels module receiving a text input and providing a text analysis and a label with a phonetic description of the text. A label buffer receives the label from the text and labels module. A parameter generation module accesses the label from the label buffer and generates a speech generation parameter. A parameter buffer receives the parameter from the parameter generation module. An audio generation module receives the text input, the label, and/or the parameter and generates a plurality of audio samples, A scheduler monitors and schedules the text and label module, the parameter generation module, and/or the audio generation module. The parameter generation module is further configured to initialize a voice identifier with a Voice Style Sheet (VSS) parameter, receive an input indicating a modification to the VSS parameter, and modify the VSS parameter according to the modification.