Object-Based Audio Downmix Decoding With Channel Gain Modification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio encoding and decoding techniques struggle to efficiently process object-based audio signals across various environments, as they often require complex decoding and rendering operations, especially when dealing with multiple object signals and channels.
Innovation Solution
The proposed method involves generating a downmix signal by downmixing object signals, extracting object-related information, and creating object-based side information that includes gain information for modifying downmix channel signals on a channel-by-channel basis, allowing for adaptive decoding and rendering by allocating object signals to appropriate positions in a multi-channel space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If object-based audio encoding is performed by treating object signals as separate entities with full decoding operations, then audio quality is improved, but decoding complexity and processing time increase significantly
Solution Approach 1:
The patent merges multiple object signals into a single downmix signal representation, where multiple object signals are combined into one or more downmix channel signals. This reduces the number of separate decoding operations needed while preserving the ability to reconstruct individual objects when required, thus lowering decoding complexity without completely sacrificing audio quality.
Solution Approach 2:
The patent applies different processing qualities to different scenarios: full object-based decoding is applied only when needed (local high quality), while downmix-based decoding is used for other cases (local acceptable quality). The system selectively applies full decoding operations only to specific objects or time periods where high quality is necessary, rather than uniformly applying complex decoding to all objects.
2Measurement precision
If multiple object signals are processed and rendered separately in multi-channel space, then spatial audio accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent segments the processing into two distinct stages: downmixing stage where multiple objects are combined into fewer channels, and rendering stage where downmix channels are converted to multi-channel output. This segmentation allows efficient batch processing of multiple objects in the downmix stage, reducing overall processing time while maintaining spatial accuracy in the rendering stage through proper object position information.
Solution Approach 2:
The patent performs preliminary downmixing operations before final rendering, where multiple object signals are pre-processed and combined into downmix signals with embedded position information. This preliminary action reduces the computational burden of the final rendering stage, as the system only needs to process downmix signals rather than individually processing and positioning multiple separate object signals.
3Productivity
If downmix signals are used without object-based side information, then transmission efficiency is improved, but flexibility in adapting to different playback environments decreases
Solution Approach 1:
The patent introduces object-based side information as an intermediary between the downmix signal and the final rendered output. This side information contains metadata about the original object signals (such as position, energy, and other characteristics) that enables flexible adaptation to different playback environments without requiring transmission of the full object signals, thus maintaining transmission efficiency while improving adaptability.
Data Source
AI summary
An audio decoding method and apparatus and an audio encoding method and apparatus which can efficiently process object-based audio signals are provided. The audio decoding method includes receiving a downmix signal and object-based side information, the downmix signal comprising at least two downmix channel signals; extracting gain information from the object-based side information and generating modification information for modifying the downmix channel signals on a channel-by-channel basis based on the gain information; and modifying the downmix channel signals by applying the modification information to the downmix channel signals.