Hierarchical Scalable Audio Encoding Band Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hierarchical encoding schemes, the fixed position of sub-bands for quantization limits coding accuracy, as the position of sub-bands with large residual energy cannot be selected as quantization targets, resulting in insufficient quality of decoded speech signals.
Innovation Solution
A coding apparatus and method that dynamically selects the quantization target band in each layer based on the coding results of the lower layer, allowing for improved encoding of perceptually important bands across hierarchies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the position of sub-bands for quantization is fixed in each hierarchy, then the encoding process is simple and structured, but the coding accuracy is limited and decoded speech quality is insufficient
Solution Approach 1:
The patent applies dynamics by making the quantization target band selection adaptive rather than fixed. The second layer coding section dynamically selects the quantization target band based on the decoded signal from the first layer, allowing the system to adapt to different speech characteristics and maximize coding accuracy for each input signal
Solution Approach 2:
The patent changes the parameter of quantization target band position from a fixed predetermined value to a variable determined by signal characteristics. The band selection is adjusted based on the decoded signal energy distribution, enabling optimal coding accuracy while maintaining hierarchical encoding structure
2Device complexity
If the quantization target band position is restricted to predetermined bands, then the encoding structure is simplified, but sub-bands with large residual energy cannot be selected resulting in poor decoded quality
Solution Approach 1:
The patent changes the quantization target band parameter from restricted predetermined positions to any sub-band within the frequency domain. The second layer coding section can select from all available sub-bands based on residual energy distribution in the decoded signal, improving decoded speech quality while maintaining manageable encoding structure
Solution Approach 2:
The system uses its own decoded output to inform subsequent encoding decisions. The decoded signal from the first layer serves as the basis for selecting the second layer quantization target band, creating a self-adaptive mechanism that improves reliability without significantly increasing external complexity
3Ease of operation
If coding results from lower hierarchies are not utilized, then each layer encoding is independent and simple, but overall coding accuracy across hierarchies is not enhanced
Solution Approach 1:
The patent implements feedback by using the decoded signal from the first layer as input for second layer band selection. This feedback loop allows the system to utilize coding results from lower hierarchies to improve overall coding accuracy, while maintaining the operational simplicity of independent layer encoding through the feedback mechanism
Solution Approach 2:
The patent performs preliminary decoding of the first layer signal before using it to guide second layer encoding decisions. This preliminary action enables the system to identify optimal quantization target bands in advance, improving overall coding accuracy while keeping the encoding process structured and manageable
Data Source
AI summary
Disclosed is an encoding device that improves the quality of a decoded signal in a hierarchical coding (scalable coding) method, wherein a band to be quantized is selected for every level (layer). The encoding device (101) is equipped with a second layer encoding unit (205) that selects a first band to be quantized of a first input signal from among a plurality of sub-bands, and that generates second layer encoding information containing first band information of said band; a second layer decoding unit (206) that generates a first decoded signal using the second layer encoding information; an addition unit (207) that generates a second input signal using the first input signal and the first decoded signal; and a third layer encoding unit (208) that selects a second band to be quantized of the second input signal using the first decoded signal, and that generates third layer encoding information.