Voice Activity Detection With Dynamic Timeout for Sentence Ending

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing speech-based systems struggle to accurately determine the end of a sentence, leading to frustration for users due to either premature cutoffs for slow speakers or delayed responses for systems with long timeout periods.

Innovation Solution

A method for determining a variable timeout period based on the domain, intent, and user speech rate, using automatic speech recognition and natural language understanding to interpret user utterances and execute instructions only after the timeout period has elapsed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If a short timeout period is used for detecting end of sentence, then the system responds quickly, but slow speakers are cut off prematurely

Engineering Contradiction:

Improveresponse speedVSAvoidsentence detection accuracy

Core Design Contradiction:

SpeedVSReliability

Solution Approach 1:

The timeout period is made dynamic rather than fixed. The system adjusts the timeout period based on the detected speech rate of the user. For slow speakers, the timeout is extended to prevent premature cutoff, while for fast speakers, the timeout remains shorter to maintain quick response. This dynamic adaptation resolves the contradiction between response speed and detection accuracy.

Inventive Principle:

Principle #15Dynamics

2Reliability

If a long timeout period is used for detecting end of sentence, then slow speakers are not cut off, but the system responds slowly

Engineering Contradiction:

Improvesentence detection accuracyVSAvoidresponse speed

Core Design Contradiction:

ReliabilityVSSpeed

Solution Approach 1:

The system changes the parameter of timeout period based on detected speech characteristics. By analyzing the speech rate and adjusting the timeout parameter accordingly, the system ensures that slow speakers are not cut off while fast speakers still receive quick responses. This parameter adaptation resolves the contradiction between detection accuracy and response speed.

Inventive Principle:

Principle #35Parameter changes

3Ease of manufacture

If a fixed timeout period is used, then the system is simple to implement, but it cannot adapt to different speaking styles

Engineering Contradiction:

Improvesystem implementation simplicityVSAvoidadaptation to speaking styles

Core Design Contradiction:

Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The system transitions from a static fixed timeout to a dynamic adaptive timeout that responds to different speaking styles. By detecting speech rate and adjusting the timeout period accordingly, the system adapts to various speaking patterns while maintaining reasonable implementation complexity through the use of speech rate detection algorithms.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS12586576B2Semantically conditioned voice activity detection

Publication Date: 2026.03.24 SOUNDHOUND AI IP LLC

AI summary

A method includes recognizing words comprised by a first utterance; interpreting the recognized words according to a grammar comprised by a domain; from the interpreting of the recognized words, determining a timeout period for the first utterance based on the domain of the first utterance; detecting end of voice activity in the first utterance; executing an instruction following an amount of time after detecting end of voice activity of the first utterance in response to the amount of time exceeding the timeout period, the executed instruction based at least in part on interpreting the recognized words.