Spherical Codebook Audio Coding for Low-Delay Vector Quantisation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio coding methods, such as linear predictive coding, face challenges in achieving low delay and complexity while maintaining high audio quality, especially at lower bit rates, due to high algorithmic delay and computational complexity in vector quantization processes.
Innovation Solution
The method employs a spherical codebook using the apple-peeling code construction principle for vector quantization in a CELP-like scheme, which reduces complexity by pre-selecting code vectors and excluding candidates with higher partial distortion, and uses a coding tree and trigonometric function table for efficient encoding and decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vector quantization with a sparse codebook is applied according to the CELP approach, then speech coding quality is improved, but computational complexity and algorithmic delay increase
Solution Approach 1:
The patent pre-calculates and stores code vectors in a codebook before actual encoding. The codebook is constructed offline using training data, containing pre-optimized excitation vectors that will be used during real-time encoding. This preliminary preparation reduces the computational burden during actual speech coding operations.
Solution Approach 2:
The patent divides the speech signal into frames and processes each frame independently through vector quantization. The codebook is also segmented into multiple code vectors that can be selectively applied to different parts of the speech signal, allowing efficient processing while maintaining quality.
2Measurement precision
If vector quantization with a sparse codebook is applied according to the CELP approach, then speech coding quality is improved, but algorithmic delay increases
Solution Approach 1:
The codebook is pre-computed and stored before encoding operations. During real-time encoding, the system only needs to search and select from pre-existing code vectors rather than generating them on-the-fly, significantly reducing algorithmic delay while maintaining the quality benefits of vector quantization.
3Measurement precision
If the input vector is located between two adjacent quantisation values of each dimension, then quantisation precision is improved, but the number of code vectors to be searched increases
Solution Approach 1:
The patent applies different quantization strategies to different dimensions of the input vector based on their local characteristics. For dimensions where the input vector falls between adjacent quantization values, the system uses refined search methods localized to those specific dimensions rather than exhaustively searching all code vectors, maintaining precision while improving encoding speed.
Solution Approach 2:
The patent performs partial search in the codebook by focusing computational effort only on relevant code vectors that are likely to match the input vector, rather than exhaustively evaluating all code vectors. This selective approach maintains quantization precision for critical dimensions while reducing overall computational burden.
Data Source
Figure 1~3
Figure 4~6
Figure 7~9
AI summary
A new wideband audio coding concept is presented that provides good audio quality at bit rates below 3 bits per sample with an algorithmic delay of less than 10 ms. The new concept is based on the principle of Linear Predictive Coding (LPC) in an analysis-by-synthesis framework, as known from speech coding. A spherical codebook is used for quantisation at bit rates which are higher in comparison to low bit rate speech coding for improved performance for audio signals. For superior audio quality, noise shaping is employed to mask the coding noise. In order to reduce the computational complexity of the encoder, the analysis-by synthesis framework has been adapted for the spherical codebook to enable a very efficient excitation vector search procedure. Furthermore, auxiliary information gathered in advance is employed to reduce a computational encoding and decoding complexity at run time significantly. This auxiliary information can be considered as the SCELP codebook. Due to the consideration of the characteristics of the apple-peeling-code construction principle, this codebook can be stored very efficiently in a read-only-memory.