Multi-object audio file structure for efficient object access

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional object-based audio content services face challenges in efficiently accessing and controlling individual audio objects due to sequential storage and frame-based index creation, leading to high bandwidth usage and complex user interactions.

Innovation Solution

A method for creating, editing, and reproducing multi-object audio content files that groups frames by reproduction time and stores index information based on predetermined units, allowing for efficient storage and transmission, and enabling users to control audio object positions and sound levels through preset information stored within or separate from the content file.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If frames for each object or data of entire objects are stored sequentially according to conventional MPEG-4 specification, then the file structure can be created, but access property to objects becomes remarkably low when a plurality of objects are stored

Engineering Contradiction:
Improveaccess property to objectsVSAvoidfile structure complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent segments the audio content by separating individual audio objects from the combined audio stream. Each audio object is extracted and stored as an independent entity with its own metadata, allowing selective access and control of individual objects without processing the entire audio file. This segmentation enables efficient random access to specific audio objects while maintaining the overall file structure.

Inventive Principle:
Principle #1Segmentation

2Speed

If index information is created on a frame basis using position information and size information of each frame, then random access to objects can be enabled, but a great deal of index information is generated and operations for acquiring index information become huge, taking long time to make random access

Engineering Contradiction:
Improverandom access speedVSAvoidamount of index information
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential index information needed for random access rather than creating comprehensive frame-based indexes. By identifying and storing only the critical position and size metadata for each audio object, the system reduces the volume of index information while maintaining efficient access capabilities. This selective extraction eliminates redundant indexing operations.

Inventive Principle:
Principle #2Taking out (Extraction)

3Loss of energy

If audio signals of diverse sound sources are combined into audio signals of a form and stored, then bandwidth usage is reduced, but viewers cannot control signal strength of audio signals of each specific sound source

Engineering Contradiction:
Improvebandwidth usageVSAvoidcontrol capability of audio signals
Core Design Contradiction:
Loss of energyVSAdaptability or versatility

Solution Approach 1:

The patent implements a dynamic audio object system where individual audio objects can be independently controlled after being extracted from the combined stream. The system maintains the efficiency of combined audio storage while enabling dynamic selection, manipulation, and control of individual audio objects through metadata tags and object identifiers. This allows viewers to adjust signal strength of specific sound sources without requiring separate storage of all audio streams.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP2113112B1Method for creating, editing, and reproducing multi-object audio contents files for object-based audio service, and method for creating audio presets
Publication Date: 2019.07.03 ELECTRONICS & TELECOMM RES INST
  • EP2113112B1 patent drawingFigure 1
  • EP2113112B1 patent drawingFigure 2
  • EP2113112B1 patent drawingFigure 3

AI summary

Provided are a method for creating, editing and reproducing a multi-object audio content file for an object-based audio service and a method for creating audio presets. The multi-object audio content file creating method includes creating a plurality of frames for each audio object forming an audio content; and creating a multi-object audio content file by grouping and storing the frames according to each reproduction time. This invention can enhance functions of the object-based audio service and make it easy to access to each audio object of an audio content file.