Predictive Pre-recording Audio for Voice Input

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users face challenges in formulating effective search queries due to the vast amount of internet information, and existing voice-based search systems require explicit initiation of voice input, leading to potential front truncation of voice input, reduced speech recognition accuracy, and battery life issues in mobile devices.

Innovation Solution

Implementing predictive pre-recording of audio on mobile devices by using sensor data and probabilistic models to anticipate when a user is likely to initiate voice input, even when the device is locked or in sleep mode, thereby commencing background recording and improving speech recognition accuracy while conserving battery life.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If continuous recording buffer is used to capture pre-input audio, then speech recognition accuracy is improved, but battery life is depleted

Engineering Contradiction:
Improvespeech recognition accuracyVSAvoidbattery life
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary action by initiating background recording only when prediction indicators suggest the user is about to provide voice input, rather than continuously recording. This anticipatory approach captures the necessary pre-input audio context only when needed, improving speech recognition accuracy while avoiding continuous battery drain from constant recording operations.

Inventive Principle:
Principle #10Preliminary action

2Ease of operation

If voice input initiation is detected after user speaks, then voice-based search is enabled, but front truncation of voice input occurs

Engineering Contradiction:
Improvevoice-based search enablementVSAvoidfront truncation of voice input
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The system performs preliminary action by starting background recording before the user actually speaks, based on prediction of upcoming voice input. This ensures that the beginning of the user's utterance is captured completely without front truncation, while still requiring explicit user initiation to maintain ease of operation and prevent accidental activation.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If explicit initiation of voice input is required, then device control is maintained, but speech recognition accuracy is reduced

Engineering Contradiction:
Improvedevice controlVSAvoidspeech recognition accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The system performs preliminary action by pre-recording audio in the background when prediction indicators suggest voice input is coming. This allows the system to have the audio already captured and ready by the time the user provides the voice command, improving speech recognition accuracy while maintaining device control through the requirement of explicit user initiation.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2553563B1Predictive pre-recording of audio for voice input
Publication Date: 2019.07.17 GOOGLE LLC
  • EP2553563B1 patent drawingFigure 1
  • EP2553563B1 patent drawingFigure 2
  • EP2553563B1 patent drawingFigure 3

AI summary

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for providing predictive pre-recording of audio for voice input. In one aspect, a method includes establishing, as input data, state data that references a state of a mobile device and sensor data that is sensed by one or more sensors of the mobile device, applying a rule or a probabilistic model to the input data, inferring, based on applying the rule or the probabilistic model to the input data, that a user of the mobile device is likely to initiate voice input, and invoking one or more functionalities of the mobile device in response to inferring that the user is likely to initiate voice input.