Gain Parameter Estimation Saturation Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In audio signal transmission using bandwidth extension, high band signal energy calculations can lead to saturation, resulting in inaccurate gain parameter estimation and degradation of audio quality due to limited bit representation in fixed-point systems.
Innovation Solution
Scaling the high band signal before energy calculation to prevent saturation, using scaling factors such as 2, 4, or 8, to ensure the energy value fits within available bit representation, thereby accurately determining gain shape and frame parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If energy value is calculated directly without scaling in fixed-point systems, then computation speed is maintained, but saturation occurs when energy value exceeds available bit representation, causing loss of precision in gain parameter estimation
Solution Approach 1:
The patent applies preliminary scaling to the high band signal before energy calculation to prevent saturation. By scaling the signal by a factor (e.g., 2, 4, or 8) before computing the energy value, the system ensures that the energy calculation remains within the representable range of fixed-point arithmetic, thereby avoiding saturation and maintaining measurement precision without requiring complex post-processing corrections
2Measurement precision
If signal scaling is applied before energy calculation, then saturation is prevented and gain parameter accuracy is improved, but additional computational steps and memory operations are required
Solution Approach 1:
The patent changes the scale parameter of the high band signal by applying a scaling factor before energy calculation. This parameter transformation ensures that the signal energy remains within the dynamic range of fixed-point representation, preventing saturation and improving energy value accuracy. The scaling factor is carefully selected to balance precision improvement with minimal impact on encoding speed
3Quantity of substance
If fixed-point representation with limited bits is used, then memory usage and transmission bandwidth are reduced, but saturation occurs at high energy levels causing audio quality degradation
Solution Approach 1:
The patent applies preliminary scaling to the high band signal before energy calculation to prevent saturation. By scaling the signal by a factor (e.g., 2, 4, or 8) before computing the energy value, the system ensures that the energy calculation remains within the representable range of fixed-point arithmetic, thereby avoiding saturation and maintaining measurement precision without requiring complex post-processing corrections
Solution Approach 2:
The scaling factor acts as an intermediary element that bridges the gap between the high energy signal and the limited fixed-point representation. By introducing this intermediate scaling operation, the system can accurately represent high energy values without exceeding the bit representation limits, thus maintaining audio quality consistency while using efficient fixed-point data representation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device including gain shape circuitry configured to determine a number of sub-frames of multiple sub-frames that are saturated, the multiple sub-frames included in a frame of a high band audio signal. The device also includes gain frame circuitry configured to determine, based on the number of sub-frames that are saturated, a gain frame parameter corresponding to the frame.