Gain Quantization in Variable Bit Rate Speech Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech coding techniques face challenges in achieving a good trade-off between subjective quality and bit rate, especially in wideband speech applications, where existing methods struggle to efficiently encode onsets and transient signals at lower bit rates, leading to degraded speech performance in certain modes of operation.
Innovation Solution
A gain quantization method that calculates an initial pitch gain based on multiple subframes, selects a portion of a gain quantization codebook, and jointly quantizes pitch and fixed-codebook gains, restricting the codebook search to the selected portion to optimize bit allocation and improve speech quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional gain quantization methods are used, then the speech coding system can operate at various bit rates, but the speech quality degrades at lower bit rates due to insufficient representation of gain variations
Solution Approach 1:
The gain quantization codebook is divided into multiple portions, each representing different gain variation characteristics. The encoder selects appropriate portions based on the speech frame type (stationary voiced, transient, unvoiced), allowing efficient representation at lower bit rates while maintaining quality for specific speech conditions
Solution Approach 2:
Different codebook portions are designed with different quantization characteristics tailored to specific speech segments. Stationary voiced frames use portions optimized for smooth gain variations, while transient frames use portions with finer quantization resolution, achieving local optimization of speech quality at each bit rate
2Manufacturing precision
If the gain quantization codebook search is performed over the entire codebook, then accurate gain representation is achieved, but the computational complexity and bit rate increase
Solution Approach 1:
The codebook is segmented into multiple portions, and the search is restricted to only the relevant portion based on the speech frame classification. This reduces the search space and computational complexity while maintaining accurate gain representation for the specific speech condition
Solution Approach 2:
Instead of searching the entire codebook, the method performs a partial search over only the necessary codebook portion. This partial action is sufficient to achieve accurate gain representation for each speech frame type without the excessive complexity of a full codebook search
3Reliability
If separate quantization tables are designed for different speech modes, then optimal performance is achieved for each mode, but the device complexity and design difficulty increase
Solution Approach 1:
A single gain quantization codebook is designed to serve multiple speech modes and bit rates. By organizing the codebook into portions with different characteristics and selecting appropriate portions based on speech frame type, the system achieves mode-specific optimization without requiring separate quantization tables for each mode
Solution Approach 2:
The codebook portions are designed with varying quantization parameters (resolution, distribution) to accommodate different speech modes. The encoder dynamically changes which codebook portion is used based on the speech frame characteristics, achieving adaptive optimization without complex table designs
Data Source
AI summary
The present invention relates to a gain quantization method and device for implementation in a technique for coding a sampled sound signal processed, during coding, by successive frames of L samples, wherein each frame is divided into a number of subframes and each subframe comprises a number N of samples, where N<L. In the gain quantization method and device, an initial pitch gain is calculated based on a number f of subframes, a portion of a gain quantization codebook is selected in relation to the initial pitch gain, and pitch and fixed-codebook gains are jointly quantized. This joint quantization of the pitch and fixed-codebook gains comprises, for the number f of subframes, searching the gain quantization codebook in relation to a search criterion. The codebook search is restricted to the selected portion of the gain quantization codebook and an index of the selected portion of the gain quantization codebook best meeting the search criterion is found.