Speech Encoding Quantization Error Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech signal encoding methods face challenges in minimizing spectrum quantization error, which affects the quality of synthesized speech.
Innovation Solution
A method that extracts optimal spectrum vectors, adaptive codebook candidates, and fixed codebook candidates using best information to determine optimal coding parameters through a search process, combining all coding parameters to minimize quantization error and improve speech quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If step-by-step optimization scheme is used to extract coding parameters, then device complexity is reduced, but manufacturing precision (quantization error minimization) deteriorates
Solution Approach 1:
The patent applies preliminary action by extracting spectrum information, adaptive codebook information, and fixed codebook information in advance before the final parameter optimization. These pre-extracted candidates are stored and then used in combination during the search process to determine optimal coding parameters, thereby achieving better quantization error minimization without excessive complexity.
Solution Approach 2:
The patent segments the coding parameter extraction process into three independent stages: spectrum vector extraction (first best information), adaptive codebook extraction (second best information), and fixed codebook extraction (third best information). Each stage produces candidate parameters that are later combined, allowing the system to achieve high precision through comprehensive search while maintaining manageable complexity through modular processing.
2Manufacturing precision
If all coding parameters are combined and searched to determine optimal parameters, then manufacturing precision (quantization error minimization) is improved, but device complexity increases
Solution Approach 1:
The patent reduces search process complexity by performing preliminary extraction of candidate parameters from three separate information sources (spectrum, adaptive codebook, fixed codebook). These pre-prepared candidates significantly reduce the search space compared to exhaustive search, enabling optimal parameter determination with manageable computational complexity.
Solution Approach 2:
The patent merges the results from three separate extraction processes (spectrum vector candidates, adaptive codebook candidates, fixed codebook candidates) into a unified search process. By combining these candidate sets and searching for the optimal combination, the system achieves comprehensive optimization while leveraging the structured organization of candidates to control complexity.
3Adaptability or versatility
If conventional speech encoding technologies are maintained for compatibility, then adaptability is improved, but manufacturing precision (quantization error) may worsen
Solution Approach 1:
The patent implements multi-functionality by designing an encoding apparatus that can operate in multiple modes: it maintains compatibility with conventional speech encoding technologies while simultaneously implementing the enhanced three-stage extraction method. This allows the system to achieve improved quantization error performance when needed while preserving adaptability to work with various conventional encoding standards and technologies.
Data Source
AI summary
According to the present invention, a linear prediction filter coefficient of a current frame is acquired from an input signal using linear prediction, a quantized spectrum candidate vector of the current frame, corresponding to the linear prediction filter coefficient of the current frame, is acquired on the basis of first best information, and the quantized spectrum candidate vector of the current frame and the quantized spectrum vector of the previous frame are interpolated. Accordingly, in contrast to conventional phased optimization techniques, optimum parameters which minimize quantization errors, can be obtained.