Global Pulse Replacement Fixed Codebook Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fixed codebook search methods in speech encoding, such as the full search, focused search, and depth first tree search methods, face high computational complexity and degrade speech quality due to inefficient pulse replacement procedures.
Innovation Solution
A global pulse replacement method is introduced, where initial codebook vectors are determined by magnitudes of pulse-position likelihood-estimator vectors, and pulses are replaced at each track to find an optimal codebook vector with reduced computation, using a mathematical equation to assess improvements in the search process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If full search method is used, then speech quality is improved, but computational complexity increases
Solution Approach 1:
The codebook search is divided into multiple stages: first searching pulse positions in tracks 0-2, then selectively searching track 3 based on threshold conditions. This segmentation reduces the search space while maintaining quality.
Solution Approach 2:
Instead of searching all possible pulse positions (full search), the method searches only the most promising positions identified by correlation thresholds, performing partial search that balances quality and complexity.
2Device complexity
If focused search method is used, then computational complexity is reduced, but speech quality degrades
Solution Approach 1:
The method uses correlation values as feedback to dynamically determine the search range. By comparing correlation sums against thresholds, the system adapts the search depth to maintain quality while reducing complexity.
3Device complexity
If depth first tree search method is used, then computational complexity is reduced, but speech quality still degrades
Solution Approach 1:
The method changes the search parameter from sequential track ordering to magnitude-based pulse position selection. By ordering pulse positions according to the magnitude of their likelihood-estimator vectors, the search prioritizes the most promising candidates.
4Device complexity
If pulse replacement procedure is used, then computational complexity is reduced, but speech quality degrades
Solution Approach 1:
The method performs preliminary sorting of pulse positions based on likelihood-estimator vector magnitudes before the search begins. This preliminary action identifies the most promising pulse positions in advance, guiding the search to high-quality candidates more efficiently.
Data Source
AI summary
The present research can decrease the amount of computation and enhance speech quality by using a global pulse replacement method in a fixed codebook search. The fixed codebook search method in a speech encoder based upon global pulse replacement, includes the steps of: (a) computing absolute values of the pulse-position likelihood-estimator vectors; (b) temporarily obtaining a codebook vector; (c) computing a mathematical equation by replacing a pulse; (d) determining whether a value computed based upon the mathematical equation is increased after pulse replacement; (e) obtaining a new codebook vector by replacing the pulse; and (f) maintaining a previous codebook vector.