Bi-stage gain shape estimation for high-band audio tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current super wideband (SWB) coding techniques for audio signals face challenges in accurately characterizing the high-band portion due to energy disparities between the low-band and high-band signals, leading to inaccurate side information and audible artifacts during signal reconstruction.
Innovation Solution
The system employs bi-stage gain shape estimation to adjust the harmonically extended low-band excitation to mimic the temporal characteristics of the high-band residual signal, using first and second gain shape parameters to refine the high-band signal reconstruction, ensuring improved energy correlation and reducing artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If SWB coding techniques use signal modeling to predict the high-band from low-band, then coding efficiency is improved, but energy disparities between low-band and high-band result in inaccurate side information
Solution Approach 1:
The patent divides the high-band signal characterization into two separate estimation stages: a first gain shape estimator that provides initial parameters, and a second gain shape estimator that refines them. This segmentation allows each stage to focus on specific aspects of energy correlation, improving overall accuracy while maintaining coding efficiency.
Solution Approach 2:
The first gain shape estimator performs preliminary estimation of high-band energy characteristics before the second estimation stage. This preliminary action provides initial gain shape parameters that guide subsequent refinement, reducing the complexity of the second stage while ensuring accurate tracking of temporal energy variations.
2Quantity of substance
If the high-band is not fully encoded and transmitted, then bandwidth usage is reduced, but the receiver cannot accurately reconstruct the high-band portion
Solution Approach 1:
The patent employs a feedback mechanism where the first gain shape estimator analyzes the relationship between low-band and high-band energy, generates initial parameters, and this information feeds into the second estimator which further refines the gain shape parameters. This iterative feedback process enables accurate high-band reconstruction from compressed side information.
Solution Approach 2:
The system transforms the high-band signal characteristics into compact gain shape parameters that capture temporal energy variations. By changing the representation from full high-band encoding to parameter-based description, the system achieves both bandwidth reduction and accurate reconstruction through precise parameter estimation.
3Measurement precision
If gain shape estimation is used to adjust harmonically extended low-band excitation, then temporal characteristics are improved, but computational complexity increases
Solution Approach 1:
The gain shape estimation process is segmented into two distinct stages, each performing specific computational tasks. The first stage handles initial parameter estimation from harmonically extended excitation, while the second stage refines these parameters. This segmentation distributes computational load and optimizes the balance between accuracy and complexity.
Solution Approach 2:
The system performs partial estimation in the first stage, generating sufficient initial parameters without complete refinement, then applies targeted refinement in the second stage. This partial action approach achieves adequate temporal characteristics tracking with reduced overall computational complexity compared to full estimation in a single stage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method includes determining, at a speech encoder, first gain shape parameters based on a harmonically extended signal and/or based on a high-band residual signal associated with a high-band portion of an audio signal. The method also includes determining second gain shape parameters based on a synthesized high-band signal and based on the high-band portion of the audio signal. The method further includes inserting the first gain parameters and the second gain shape parameters into an encoded version of the audio signal to enable gain adjustment during reproduction of the audio signal from the encoded version of the audio signal.