Multichannel Audio Coding ITD Compensation via Frequency Domain Circular Shift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multichannel audio coding methods, such as parametric stereo in MPEG-4 Part 3, struggle to accurately compensate and synthesize inter-channel time differences (ITDs) in the frequency domain, leading to negative effects on spatial parameter estimates and audio quality.
Innovation Solution
A comparison device and method for multi-channel audio signals that derive ITD parameters, compensate ITDs in the frequency domain through circular shifts, and compute comparison parameters to mitigate the effects of window offsets on spatial parameter estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If channels are aligned in the frequency domain by circular shift to compensate ITD, then ITD compensation accuracy is improved, but window offset occurs which negatively affects spatial parameter estimates
Solution Approach 1:
The patent segments the frequency domain processing into distinct stages: first performing ITD compensation through circular shift, then separately correcting the resulting window offset effects on spatial parameter estimates. This segmentation allows each stage to be optimized independently, resolving the contradiction between ITD compensation accuracy and spatial parameter estimation accuracy.
Solution Approach 2:
The patent applies preliminary ITD compensation through circular shift before extracting spatial parameters, and then applies a corrective action specifically针对 the window offset introduced by the circular shift. This preliminary action followed by correction resolves the contradiction by anticipating and compensating for the negative effect.
2Productivity
If time-to-frequency transform is applied for ITD estimation, then spectral filtering efficiency and computational efficiency are improved, but complexity increases due to additional transforms
Solution Approach 1:
The patent uses the same time-to-frequency transform (e.g., DFT or MDCT) for multiple purposes: ITD estimation, spectral filtering, and spatial parameter extraction. By making the transform universal across different processing stages, the patent achieves computational efficiency through reuse while managing complexity through standardization.
Solution Approach 2:
The patent combines multiple functions (ITD estimation, spectral filtering, spatial parameter extraction) into a single integrated frequency domain processing framework. This merging eliminates the need for separate transform operations, improving productivity while containing complexity through unified processing.
3Loss of substance
If parametric stereo encoding is used to reduce channels, then bitrate efficiency is improved, but capability to compensate ITD and synthesize spatial audio is lost
Solution Approach 1:
The patent extends traditional parametric stereo by adding ITD parameters (along with existing ILD and ICC parameters) to the side information transmitted to the decoder. This parameter change enables the decoder to perform ITD compensation and reconstruct spatial audio effects, resolving the contradiction between bitrate efficiency and spatial audio synthesis capability.
Solution Approach 2:
The patent introduces ITD parameters as intermediary information that mediates between the downmixed stereo signal and the original spatial audio. These parameters carry the essential temporal spatial information needed for accurate ITD compensation, enabling high-quality spatial reconstruction at low bitrates.
Data Source
AI summary
In multichannel audio coding, improved computational efficiency is achieved by computing comparison parameters for ITD compensation between any two channels in the frequency domain for a parametric audio encoder. This may mitigate negative effects on encoder parameter estimates.