Neural Network Sound Categorization for Faster SFX Search

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing sound databases for movies and videogames require extensive manual categorization and familiarity with the database scheme, making it difficult for new content creators to find and add sounds efficiently.

Innovation Solution

A multi-tiered sound categorization system using neural networks that learns sound categorization based on acoustic properties, enabling automatic classification and search of sound effects (SFX) through hierarchical classification and haptic feedback.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual categorization of sound databases is used, then sound files can be organized and accessed, but extensive time and familiarity with the categorization scheme are required

Engineering Contradiction:
ImproveAccessibility of sound filesVSAvoidTime to find and add sounds
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical categorization with an automated neural network system. The neural network automatically analyzes audio features of sound files and assigns them to appropriate categories and subcategories, eliminating the need for manual classification while maintaining organized access to sound files.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The sound database system performs self-categorization through the neural network, which automatically processes new sound files upon upload, determines their categories and subcategories based on acoustic properties, and organizes them without human intervention. This self-service mechanism continuously maintains the database organization.

Inventive Principle:
Principle #25Self-service

2Reliability

If new custom sounds are added to the database manually, then the database remains comprehensive, but extensive time and expertise are required

Engineering Contradiction:
ImproveComprehensiveness of sound databaseVSAvoidSpeed of adding new sounds
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The manual process of adding and categorizing new sounds is replaced by an automated neural network system. When new sound files are uploaded, the neural network automatically analyzes their acoustic properties, determines appropriate categories and subcategories, and integrates them into the database, maintaining comprehensiveness while dramatically increasing productivity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If detailed hierarchical categorization is implemented, then sound files can be precisely classified, but the complexity of the categorization scheme increases

Engineering Contradiction:
ImprovePrecision of sound classificationVSAvoidComplexity of categorization scheme
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The categorization scheme is segmented into multiple hierarchical levels: broad categories (e.g., environmental sounds, mechanical sounds) and more specific subcategories (e.g., nature sounds, animal sounds). This segmentation allows precise classification while managing complexity by breaking down the classification task into manageable hierarchical stages that the neural network processes sequentially.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3857448B1Sound categorization system
Publication Date: 2025.10.29 SONY INTERACTIVE ENTERTAINMENT LLC
  • EP3857448B1 patent drawingFigure 1
  • EP3857448B1 patent drawingFigure 2A~2C
  • EP3857448B1 patent drawingFigure 2D

AI summary

A system method and computer program product for hierarchical categorization of sound comprising one or more neural networks implemented on one or more processors. The one or more neural networks are configured to categorize a sound into a two or more tiered hierarchical course categorization and a finest level categorization in the hierarchy. The categorization sound may be used to search a database for similar or contextually related sounds.