3D Audio Scene Tree Encoding for Low-Bitrate Spatial Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scene graph data structures for describing audio object locations in 3D scenes are inefficient in terms of data usage, especially in low-bitrate streaming scenarios, leading to suboptimal storage and transmission efficiency.
Innovation Solution
An audio codec system encodes scene metadata as a scene tree structure into a bitstream, which includes an origin of a first 3D scene relative to a second 3D scene and the position of sound sources within the first scene, using identifiers to reduce data requirements while maintaining accurate spatial rendering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional scene graph data structures are used to describe audio object locations in 3D scenes, then accurate spatial representation is achieved, but data usage efficiency deteriorates leading to suboptimal storage and transmission efficiency
Solution Approach 1:
The scene graph data structure is segmented into hierarchical levels (global scene, local scenes, audio objects) where each level contains only the information necessary for that level. This segmentation allows the system to represent spatial relationships using compact relative position data rather than complete absolute coordinates for all objects, thereby reducing data usage while maintaining spatial rendering accuracy.
Solution Approach 2:
The patent implements a nested hierarchical structure where local scenes are nested within the global scene, and audio objects are nested within local scenes. Each nested level references its parent scene's origin, allowing efficient encoding of positions through relative offsets. This nesting approach enables the system to achieve accurate spatial representation with reduced data requirements by avoiding redundant information across hierarchical levels.
2Measurement precision
If detailed position information is transmitted for all audio objects, then spatial rendering quality is improved, but transmission bitrate increases which is problematic for low-bitrate streaming
Solution Approach 1:
The patent applies local quality by encoding position information with varying precision based on the object's context. Audio objects within local scenes use relative position encoding with respect to their local scene origin, which requires fewer bits than absolute global coordinates. This localized encoding approach maintains sufficient position accuracy for spatial rendering while significantly reducing the bitrate required for transmission compared to encoding all positions with uniform high precision.
3Stability of the object's composition
If complete scene configuration data is stored for each audio object, then spatial relationships are accurately maintained, but storage efficiency deteriorates
Solution Approach 1:
The patent performs preliminary organization of scene data into a hierarchical structure where the global scene origin and local scene origins are established beforehand. Audio objects then reference their positions relative to these pre-defined origins rather than storing complete scene configuration data. This preliminary structuring ensures scene configuration integrity is maintained while dramatically reducing storage requirements by eliminating redundant configuration information.
Data Source
AI summary
A method that includes receiving a bitstream that comprises: an encoded version of an audio signal that is associated with a sound source that is within a first 3D scene, a scene tree structure that includes an origin of the first scene relative to an origin of a second scene, and a position of the sound source within the first scene relative to the origin of the first scene, wherein the position references the origin of the first scene using an identifier, wherein the scene tree structure defines an initial configuration of the sound source with respect to the first and second scenes; determining a position of a listener; producing a set of spatially rendered audio signals by spatially rendering the audio signal according to the position of the sound source with respect to the position of the listener; and using the spatially rendered audio signals to drive speakers.