Encoding Apparatus Using Parameter Eta for Adaptive Bit Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques lack a method to identify and configure an appropriate encoding or decoding process based on the parameter η for efficient encoding and decoding of time-series signals, particularly in the context of sound signals.
Innovation Solution
An encoding apparatus and decoding apparatus that utilize the parameter η as a shape parameter for the generalized Gaussian distribution to adjust bit allocation in the encoding process, allowing for adaptive encoding and decoding by selecting the appropriate configuration based on the parameter η for each time section.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LSP parameter is used directly for sound classification and section estimation, then the parameter can indicate characteristics of time-series signals, but it is difficult to perform threshold-based processes due to multiple values
Solution Approach 1:
The patent extracts the essential characteristic from the multi-value LSP parameter by introducing a single-value parameter η that represents the overall shape characteristic of the signal. This extracted parameter can be easily used for threshold-based classification while preserving the important signal characteristics.
Solution Approach 2:
Instead of directly using the complex multi-value LSP parameter for classification, the patent inverts the approach by deriving a simplified parameter η from the LSP values that captures the essential shape characteristic, making threshold-based operations feasible.
2Ease of manufacture
If a fixed encoding process configuration is used for all time sections, then the encoding process is simple to implement, but encoding efficiency decreases when signal characteristics vary across time sections
Solution Approach 1:
The patent introduces dynamic adaptability to the encoding process by using the parameter η to identify appropriate encoding configurations for different time sections. The encoding process can dynamically adjust its parameters based on the signal characteristics represented by η, optimizing efficiency while maintaining implementation feasibility through automated identification.
Solution Approach 2:
The patent changes the encoding process parameters based on the identified parameter η for each time section. By adjusting encoding parameters according to the signal's shape characteristics, the system achieves higher encoding efficiency for varying signal types while keeping the overall process manageable through systematic parameter adaptation.
3Device complexity
If bit allocation is fixed in the encoding process, then the encoding process is straightforward, but compression efficiency is suboptimal when signal characteristics change over time
Solution Approach 1:
The patent makes bit allocation dynamic by linking it to the parameter η that characterizes the signal shape in each time section. The bit allocation automatically adapts to signal characteristics, improving compression efficiency while maintaining relatively simple process complexity through systematic adaptation rules.
Solution Approach 2:
The patent changes the bit allocation parameters based on the identified signal characteristics represented by η. By adjusting the number of bits allocated to different parameters according to the signal type, the system achieves optimal compression efficiency for varying signal characteristics while keeping the encoding process manageable.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An encoding apparatus is an encoding apparatus for encoding a time-series signal for each of predetermined time sections in a frequency domain, wherein a parameter η is a positive number, the parameter η corresponding to a time-series signal is a shape parameter of generalized Gaussian distribution that approximates a histogram of a whitened spectral sequence, which is a sequence obtained by dividing a frequency domain sample sequence corresponding to the time-series signal by a spectral envelope estimated by regarding the η-th power of absolute values of the frequency domain sample sequence as a power spectrum, and any of a plurality of parameters η is selective or the parameter η is variable for each of the predetermined time sections; and the encoding apparatus comprises an encoding portion encoding the time-series signal for each of the predetermined time sections by an encoding process with a configuration identified at least based on the parameter η for each of the predetermined time sections.