Soundhound, Inc.

Monika Almudafar-Depeyrot

Keyvan Mohajer

Bernard Mont-Reynaud

Mark Stevans

Machine learned models take in vectors representing desired behaviors and generate voice vectors that provide the parameters for text-to-speech (TTS) synthesis. Models may be trained on behavior vectors that include user profile attributes, situational attributes, or semantic attributes. Situational attributes may include age of people present, music that is playing, location, noise, and mood. Semantic attributes may include presence of proper nouns, number of modifiers, emotional charge, and domain of discourse. TTS voice parameters may apply per utterance and per word as to enable contrastive emphasis.

Text-to-speech adapted by machine learning

A speech-enabled dialog system responds to a plurality of wake-up phrases. Based on which wake-up phrase is detected, the system's configuration is modified accordingly. Various configurable aspects of the system include selection and morphine of a text-to-speech voice; configuration of acoustic model, language model, vocabulary, and grammar; configuration of a graphic animation; configuration of virtual assistant personality parameters; invocation of a particular user profile; invocation of an authentication function; and configuration of an open sound. Configuration depends on a target market segment. Configuration also depends on the state of the dialog system, such as whether a previous utterance was an information query.

Integration of third party virtual assistants

Software-based systems perform parametric speech synthesis. TTS voice parameters determine the generated speech audio. Voice parameters include gender, age, dialect, donor, arousal, authoritativeness, pitch, range, speech rate, volume, flutter, roughness, breath, frequencies, bandwidths, and relative amplitudes of formants and nasal sounds. The system chooses TTS parameters based on one or more of: user profile attributes including gender, age, and dialect; situational attributes such as location, noise level, and mood; natural language semantic attributes such as domain of conversation, expression type, dimensions of affect, word emphasis and sentence structure; and analysis of target speaker voices. The system chooses TTS parameters to improve listener satisfaction or other desired listener behavior. Choices may be made by specified algorithms defined by code developers, or by machine learning algorithms trained on labeled samples of system performance.

Parametric adaptation of voice synthesis

A speech-enabled dialog system responds to a plurality of wake-up phrases. Based on which wake-up phrase is detected, the system's configuration is modified accordingly. Various configurable aspects of the system include selection and morphing of a text-to-speech voice; configuration of acoustic model, language model, vocabulary, and grammar; configuration of a graphic animation; configuration of virtual assistant personality parameters; invocation of a particular user profile; invocation of an authentication function; and configuration of an open sound. Configuration depends on a target market segment. Configuration also depends on the state of the dialog system, such as whether a previous utterance was an information query.

Growing community of inventors

Menlo Park, CA, United States of America

Monika Almudafar-Depeyrot