Scalable LSP Encoding via Pre-emphasis Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for scalable LSP encoding in voice communication systems, such as those using CELP systems, inadequately utilize narrowband LSP information for wideband encoding, resulting in inefficient quantization and encoding performance.
Innovation Solution
The implementation of a scalable encoding apparatus and method that performs predictive quantization of wideband LSP parameters using pre-emphasized narrowband quantized LSP parameters, enhancing quantization efficiency by adaptive encoding and multistage vector quantization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple multiplication of narrowband LSP parameters by a constant is used for wideband prediction, then the encoding process is simple, but quantization efficiency and encoding performance are inadequate
Solution Approach 1:
The patent transforms the narrowband LSP parameters through a series of parameter changes including pre-emphasis filtering, spectral transformation, and adaptive scaling. Instead of simple multiplication, the system applies a pre-emphasis filter to enhance high-frequency components, then uses spectral transformation to map narrowband parameters to wideband frequency ranges, finally applying adaptive scaling based on signal characteristics to optimize quantization accuracy.
Solution Approach 2:
The patent introduces intermediary processing stages between narrowband LSP parameters and wideband prediction. These intermediaries include a pre-emphasis filter that modifies the spectral characteristics, a spectral transformation module that bridges frequency domains, and an adaptive scaling mechanism that adjusts parameters based on signal conditions, thereby improving the overall quantization efficiency.
2Device complexity
If narrowband LSP information is inadequately utilized for wideband encoding, then the encoding apparatus is simpler, but encoding performance and quantization efficiency deteriorate
Solution Approach 1:
The patent applies pre-emphasis filtering to the narrowband LSP parameters before they are used for wideband prediction. This preliminary action enhances the high-frequency components of the narrowband parameters, making them more representative of the wideband signal characteristics. By preparing the parameters in advance with improved spectral distribution, the system achieves better encoding performance without significantly increasing apparatus complexity.
Solution Approach 2:
The patent introduces dynamic adaptation in the parameter transformation process. The scaling factors and transformation characteristics are adjusted based on the actual signal conditions and statistical properties of the input speech signal. This dynamic approach allows the system to optimize encoding performance for different speech contexts while maintaining a relatively simple apparatus structure.
3Productivity
If conventional LSP encoding is used without pre-emphasis, then the system uses less processing, but it cannot maximize narrowband LSP information for wideband encoding
Solution Approach 1:
The patent applies pre-emphasis filtering to transform the narrowband LSP parameters, changing their spectral distribution to better represent wideband characteristics. This parameter transformation extracts more useful information from the narrowband parameters by enhancing high-frequency components that are otherwise attenuated. The process maintains processing efficiency by using a simple filter structure while significantly improving information utilization.
Solution Approach 2:
The patent replaces the conventional direct mapping approach with a signal processing-based transformation system. Instead of mechanically multiplying parameters by constants, the system uses pre-emphasis filtering and spectral transformation to extract and transform information, thereby maximizing the utilization of narrowband LSP information for wideband encoding purposes.
Data Source
AI summary
A scalable encoding apparatus, a scalable decoding apparatus and the like are disclosed which can achieve a band scalable LSP encoding that exhibits both a high quantization efficiency and a high performance. In these apparatuses, a narrow band-to-wide band converter receives and converts a quantized narrow band LSP to a wide band, and then outputs the quantized narrow band LSP as converted (i.e., a converted wide band LSP parameter) to an LSP-to-LPC converter. The LSP-to-LPC converter converts the quantized narrow band LSP as converted to a linear prediction coefficient and then outputs it to a pre-emphasizer. The pre-emphasizer calculates and outputs the pre-emphasized linear prediction coefficient to an LPC-to-LSP converter. The LPC-to-LSP converter converts the pre-emphasized linear prediction coefficient to a pre-emphasized quantized narrow band LSP as wide band converted, and then outputs it to a prediction quantizer.