Gain-Shape Vector Coding Under Segment Bit Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gain-shape vector quantization methods face challenges in determining a suitable vector length, which affects quantization error and perceived quality in audio codecs, and also struggle with efficient bit allocation between gain and shape coding.
Innovation Solution
The method involves analyzing the shape of a target vector to determine an optimal number of sections or splits, allowing for efficient gain shape vector encoding and decoding by dynamically adjusting the number of segments based on energy variations and bit budget constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the vector length is increased, then the shape quantizer can handle larger variations within the vector, but the overhead for gain coding increases and fewer bits are left for shape coding
Solution Approach 1:
The patent divides the target vector into multiple sub-vectors or segments, allowing independent quantization of each segment. This segmentation enables the system to handle larger overall vector variations while maintaining efficient bit allocation per segment, resolving the contradiction between handling large variations and maintaining bit budget efficiency.
Solution Approach 2:
The patent dynamically adjusts the number of segments and bit allocation based on the characteristics of the target vector. By making the segmentation structure adaptive rather than fixed, the system can optimize the balance between handling vector variations and maintaining efficient coding, directly addressing the technical contradiction.
2Device complexity
If the vector length is decreased, then the dynamics within the vector are reduced, but the lower dimensionality of the shape VQ has less capability to exploit sample correlation
Solution Approach 1:
By segmenting the vector into multiple smaller sub-vectors, the patent maintains manageable complexity for each segment while collectively preserving the ability to exploit sample correlation across the entire vector. Each segment can be quantized independently with reduced dynamics, yet the overall structure maintains high quantization capability.
Solution Approach 2:
The patent transforms the problem by introducing a temporal or structural dimension through segmentation. Instead of dealing with the full vector dimensionality at once, the system processes multiple lower-dimensional segments, effectively distributing the quantization task across multiple dimensions while maintaining overall precision.
3Measurement precision
If the number of partitions is increased, then larger variations within the vector can be handled, but the overhead for gain coding increases
Solution Approach 1:
The patent segments the vector into multiple partitions, enabling each partition to handle local signal variations efficiently. By distributing the dynamics handling across multiple smaller partitions rather than one large partition, the gain coding overhead per partition is reduced while collectively maintaining the ability to handle large overall variations.
Solution Approach 2:
The patent applies different quantization strategies or parameters to different local partitions based on their specific characteristics. This local optimization allows each partition to handle its local variations efficiently with minimal overhead, while the collection of partitions handles global variations effectively.
Data Source
Figure 1~2
Figure 3a~4
Figure 5~6
AI summary
An audio encoder and an audio decoder and methods therein for gain shape vector encoding and decoding. The method performed by an encoder, where the encoding of each vector segment is subjected to a constraint related to a maximum number of bits, BMAX, allowed for encoding a vector segment. The method comprises determining an initial number, Np_init, of segments for a target vector x and determining energy of each Np_init target vector segment. The method further comprises determining an average number of bits per segment, BAVG, based on a vector bit budget and Np_init. The method further comprises determining a final number of segments to be used in the gain shape vector encoding based on energies of the Np_init segments and a difference between BMAX and BAVG. The performing of the method enables an efficient allocation of the bits of the bit budget over the target vector.