Context-Aware Audio Generation for AR Content Interaction

Solution Overview

Problem

Conventional techniques for incorporating audio in digital content, such as augmented reality, are time-consuming and computationally intensive, requiring manual selection from a vast array of options, which is inefficient and costly.

Innovation Solution

A content-aware audio data acquisition system using a machine-learning based audio generation system that automatically generates audio data based on monitored context, employing a programming by demonstration pipeline and large language models to process contextual information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual selection of audio is used, then audio can be selected for digital content, but the process is time consuming and computationally resource intensive

Engineering Contradiction:

Improveaudio selection processVSAvoidtime for audio selection

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The system enables self-service by automatically generating audio data through machine learning models based on contextual information from digital content, eliminating the need for manual audio selection and significantly reducing the time and computational resources required for the process

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of audio selection with an automated machine learning-based audio generation system that processes contextual data and generates appropriate audio data automatically, substituting human effort with intelligent automated systems

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If manual selection of audio is used, then audio can be selected for digital content, but it is computationally resource intensive and expensive

Engineering Contradiction:

Improveaudio selection processVSAvoidcomputational resources for audio selection

Core Design Contradiction:

Ease of operationVSUse of energy by moving object

Solution Approach 1:

The system enables self-service by automatically generating audio data through machine learning models based on contextual information from digital content, eliminating the need for manual audio selection and significantly reducing the time and computational resources required for the process

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent changes the fundamental parameter of audio creation from manual selection to automated generation using machine learning models, transforming the process into an efficient computational system that reduces overall resource consumption by eliminating redundant manual operations

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If manual navigation through audio options is used, then audio of interest can be selected, but the process is inefficient when confronted with billions of potential uses for audio

Engineering Contradiction:

Improveaudio selection capabilityVSAvoidaudio selection efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent replaces the manual mechanical process of audio selection with an automated machine learning-based audio generation system that processes contextual data and generates appropriate audio data automatically, substituting human effort with intelligent automated systems

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system introduces an intermediary machine learning model that acts as a mediator between contextual information from digital content and the final audio generation, automatically filtering and processing billions of potential audio uses to generate only the most relevant audio data

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20260023522A1Context aware audio data aquisition

Publication Date: 2026.01.22 ADOBE INC

AI summary

Context aware audio data acquisition techniques are described. In one or more examples, an event is detected from one or more inputs defining interaction of a virtual object in a user interface with a depiction of a real-world physical environment captured by frames of a digital video. A context of the event in the user interface is monitored and used to generate a prompt to initiate acquisition of audio data based on the context using one or more machine-learning models. The audio data generated by the one or more machine-learning models is presented for output via the user interface.