Audio Encoder Context Mapping Across Time-Frequency Resolutions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional audio coding methods experience a significant loss in coding efficiency when frequently switching between different time/frequency resolutions, as they reset the context, leading to sub-optimal codebook selection and reduced adaptation to signal characteristics.
Innovation Solution
Implementing a context mapping mechanism that derives new contexts for coefficients with changing resolutions by using interpolation, extrapolation, sub-sampling, or down-sampling of previous context information, allowing for more frequent switching between block lengths without compromising coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If context is reset when switching between different time/frequency resolutions, then switching between block lengths can occur, but coding efficiency decreases significantly
Solution Approach 1:
The patent applies preliminary action by pre-processing the old context through mapping (interpolation, extrapolation, sub-sampling, or down-sampling) before switching to new resolution. This prepares the context in advance so that it can be directly used for codebook selection after switching, eliminating the need to reset and rebuild context from scratch.
Solution Approach 2:
The patent changes the parameters of the context by applying mathematical transformations (interpolation, extrapolation, sub-sampling, down-sampling) to adapt the old context to the new resolution. This allows the context to maintain its usefulness across different time/frequency resolutions without resetting.
2Loss of energy
If context mapping mechanism is implemented, then coding efficiency is maintained during switching, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary context mapping mechanism that acts as a bridge between the old context and the new resolution requirements. This intermediary process (interpolation, extrapolation, sub-sampling, or down-sampling) transforms the old context into a form suitable for the new resolution, avoiding both complete reset and direct reuse.
3Adaptability or versatility
If frequent switching between resolutions occurs, then adaptation to signal characteristics improves, but coding efficiency loss increases with conventional methods
Solution Approach 1:
By pre-processing the context through mapping operations before switching, the system maintains coding efficiency even when switching frequently. The context is prepared in advance in a resolution-agnostic form that can be quickly adapted to any new resolution without loss of efficiency.
Solution Approach 2:
The patent enables dynamic switching between different time/frequency resolutions by making the context adaptable through mapping. The system can freely switch between block lengths (128 or 1024 coefficients) based on signal characteristics without being constrained by efficiency losses from context resets.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An audio encoder (100) for encoding segments of coefficients, the segments of coefficients representing different time or frequency resolutions of a sampled audio signal, the audio encoder (100) comprising a processor (110) for deriving a coding context for a currently encoded coefficient of a current segment based on a previously encoded coefficient of a previous segment, the previously encoded coefficient representing a different time or frequency resolution than the currently encoded coefficient. The audio encoder (100) further comprises an entropy encoder (120) for entropy encoding the current coefficient based on the coding context to obtain an encoded audio stream.