Non-lexical Cue Insertion Engine for Natural Speech Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current natural language interfaces in computing devices, such as smartphones and laptops, produce synthesized speech that sounds artificial due to the lack of non-lexical cues like prosody, pitch, and disfluencies, making it difficult for users to comprehend the intended meaning.
Innovation Solution
Incorporating non-lexical cues into text-to-speech systems using algorithmic technology to determine insertion points for cues such as disfluencies, phrasal stress, intonation, and breathing noises, which are then synthesized into the output speech to convey intended meaning and improve comprehension.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If non-lexical cues are incorporated into text-to-speech systems, then the naturalness and comprehensibility of synthesized speech is improved, but the device complexity increases
Solution Approach 1:
The patent segments the text-to-speech system into distinct functional modules: a non-lexical cue insertion engine that processes lexical text to generate augmented text with non-lexical cues, and a speech synthesizer that converts the augmented text to speech. This segmentation allows each module to specialize in specific tasks, improving overall comprehensibility while managing complexity through modular design.
Solution Approach 2:
The patent introduces an intermediary component - the non-lexical cue insertion engine - that acts as a mediator between the lexical text input and the speech synthesis process. This intermediary adds non-lexical cues (such as disfluencies, pauses, and emphasis markers) to the lexical text, thereby enhancing the naturalness and comprehensibility of the final speech output without requiring complete redesign of the synthesis system.
2Loss of information
If non-lexical cues are added to convey intended meaning, then the information content of synthesized speech is improved, but the processing time increases
Solution Approach 1:
The non-lexical cue insertion engine performs preliminary action by pre-processing the lexical text to identify and insert appropriate non-lexical cues before the speech synthesis stage. By determining insertion points and adding cues in advance, the system enriches the information content of the speech without adding significant processing time during the actual synthesis and playback phases.
Solution Approach 2:
The patent applies parameter changes by modifying the text representation to include non-lexical cue markers at specific insertion points. These parameter changes (adding metadata tags, modifying text structure) enable the speech synthesizer to incorporate expressive features efficiently, conveying additional information about intended meaning without proportionally increasing processing time.
Data Source
AI summary
Systems and methods are disclosed for providing non-lexical cues in synthesized speech. An example system includes processor circuitry to generate a breathing cue to enhance speech to be synthesized from text; determine a first insertion point of the breathing cue in the text, wherein the breathing cue is identified by a first tag of a markup language; generate a prosody cue to enhance speech to be synthesized from the text; determine a second insertion point of the prosody cue in the text, wherein the prosody cue is identified by a second tag of the markup language; insert the breathing cue at the first insertion point based on the first tag and the prosody cue at the second insertion point based on the second tag; and trigger a synthesis of the speech from the text, the breathing cue, and the prosody cue.