Audio Object Extraction via Sub-band Probability Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting audio objects from traditional channel-based audio content often result in audible artifacts and instability, especially when dealing with complex multi-channel final-mixes that need to be up-mixed from 2D to 3D, as they fail to accurately separate multiple sources and preserve artistic intent.
Innovation Solution
A method and system for soft audio object extraction that analyzes each sub-band of an audio signal to determine a sub-band object probability, softly assigning it to either an audio object or residual audio portion, using factors like spatial position, channel correlation, panning rules, and frequency range to minimize switching artifacts and instability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional channel-based audio extraction methods are used, then the extraction process is simple, but audible artifacts and instability occur
Solution Approach 1:
The patent applies parameter changes by transitioning from binary hard decisions to continuous probability values in the range [0, 1] for sub-band object assignments. This probabilistic parameter transformation allows smooth transitions between audio object and residual audio portions, eliminating the abrupt switching artifacts that cause instability while maintaining extraction simplicity through automated probability-based classification.
2Measurement precision
If hard decision methods are used to assign sub-bands, then the classification is clear and definite, but switching artifacts and instability occur
Solution Approach 1:
The patent introduces probability values as an intermediary between hard decisions and final classification. Instead of directly assigning sub-bands to audio objects or residual audio, the system first computes object probabilities and then uses these intermediate probabilistic values to determine assignments. This intermediary layer smooths transitions and eliminates the harmful switching artifacts while preserving classification precision through the probabilistic framework.
3Reliability
If soft assignment with probability estimation is used, then switching artifacts are reduced, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the audio spectrum into multiple sub-bands and processing each sub-band independently with individual probability estimations. This segmentation approach allows the complex probabilistic computation to be distributed across frequency segments, reducing the overall computational burden while maintaining extraction stability through the soft assignment mechanism in each sub-band.
Data Source
AI summary
Embodiments of the example embodiment relate to audio object extraction. A method for audio object extraction from audio content is disclosed. The method comprises determining a sub-band object probability for a sub-band of the audio signal in a frame of the audio content, the sub-band object probability indicating a probability of the sub-band of the audio signal containing an audio object. The method further comprises splitting the sub-band of the audio signal into an audio object portion and a residual audio portion based on the determined sub-band object probability. Corresponding system and computer program product are also disclosed.