Stereo Audio Encoder Reordering for Low-Bitrate Parameter Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current stereo audio encoders face challenges in achieving high coding efficiency, particularly at lower bitrates, which affects the quality of encoded stereo parameters and downmixed signals, leading to inefficient bit usage and reduced performance.
Innovation Solution
The proposed solution involves storing the order of first-order probabilities for symbols to be encoded and using an adaptive Golomb Rice coding method, where the probability order values are encoded, allowing for better exploitation of intraframe value correlation to reduce bitrate consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If adaptive Golomb Rice coding with probability order storage is used for stereo parameters, then bitrate consumption is reduced, but device complexity increases
Solution Approach 1:
The encoder performs preliminary sorting of quantized stereo parameters by probability order before entropy coding. By pre-organizing the parameters in decreasing order of probability, the encoder exploits the statistical correlation more effectively, resulting in lower bitrate consumption. The sorting is done once per frame before the actual coding process, making the subsequent encoding more efficient.
Solution Approach 2:
The invention changes the ordering parameter of stereo parameters from their original quantization order to probability-based order. This parameter transformation allows the entropy coder to exploit the inherent statistical correlation in stereo parameters more effectively, reducing the number of bits needed to represent the same information while managing complexity through efficient data structures.
2Measurement precision
If more bits are allocated to stereo parameters, then stereo quality is improved, but bits available for downmixed signal encoding are reduced
Solution Approach 1:
The invention transforms the representation of stereo parameters by sorting them according to probability order rather than using fixed quantization levels. This parameter reorganization exploits the statistical correlation between stereo parameters, allowing for more efficient bit allocation. As a result, fewer bits are needed to achieve the same stereo quality, freeing up bits for downmixed signal encoding while maintaining or improving overall quality.
Solution Approach 2:
The encoder creates a sorted copy of the quantized stereo parameters based on probability order, then uses this reordered sequence for entropy coding. This copying approach allows the original quantization structure to be preserved for quality while the sorted copy enables more efficient compression, optimizing the trade-off between stereo parameter quality and bits available for the downmixed signal.
3Ease of manufacture
If fixed quantization levels are used for stereo parameters, then encoding simplicity is maintained, but coding efficiency is reduced
Solution Approach 1:
The encoder performs a preliminary sorting operation on quantized stereo parameters before entropy coding. This pre-processing step organizes parameters in decreasing order of probability, creating a more favorable distribution for entropy coding. The sorting is computationally efficient and maintains relative simplicity while significantly improving coding efficiency by exploiting statistical correlations that would otherwise be lost.
Solution Approach 2:
The invention introduces dynamic reordering of stereo parameters based on their probability distribution within each frame, rather than using a fixed quantization order. This dynamic adaptation allows the encoder to optimize for the actual statistical properties of the input signal, improving coding efficiency while maintaining implementation simplicity through the use of efficient sorting algorithms and data structures.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method comprising: receiving at least two audio channel signals; determining, for a first frame, at least two parameters representing a difference between the at least two channel audio signals; scalar quantising (451 ) the at least two parameters to generate at least two index values; determining an initial index map for reordering (453) one of the at least two index values, and determining at least one further index map for reordering at least one further of the at least two index values, wherein the at least one further index map is determined based on the one of the at least two index values; reordering the one of the at least two index values based on the initial index map; reordering the further of the at least two index values based on the at least one further index map; encoding (455) the reordered one of the at least two index values dependent on an order position of the reordered one of the at least two index values; encoding the reordered further of the at least two index values based on an order position of the reordered further of the at least two index values; generating a single channel representation of the at least two audio channel signals dependent on the at least two parameters; and encoding (456) the single channel representation.