Spatial Background Noise Coding for Smooth Multichannel Transitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies struggle to efficiently model spatial background noise in multi-channel audio signals, particularly in transitions between active and inactive sections, leading to abrupt changes that disrupt the listening experience.
Innovation Solution
A method for encoding and decoding multi-channel audio signals that involves determining active and inactive frames, generating downmix signals, and encoding spatial parameters (SPAR metadata) to reconstruct high-quality spatial comfort noise, using voice activity detection and bitrate control to optimize data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If spatial background noise is modelled using existing technologies, then data transmission requirements are reduced, but transitions between active and inactive sections become abrupt and perceptually disruptive
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing spatial parameters (covariance matrices, eigenvectors, eigenvalues) during active frames before transitioning to inactive frames. These pre-computed parameters are then reused during inactive frames to generate smooth comfort noise, avoiding abrupt transitions while minimizing data transmission.
Solution Approach 2:
The patent uses copying by creating a simplified representation of the spatial background noise characteristics through downmix signals and encoding essential spatial parameters (eigenvectors and eigenvalues) rather than transmitting complete multi-channel noise data. This copied representation is sufficient to regenerate perceptually accurate comfort noise at the decoder.
2Reliability
If complete spatial parameters are transmitted for background noise, then perceptual quality is maintained, but data transmission requirements increase
Solution Approach 1:
The patent extracts only the essential spatial characteristics from the complete multi-channel background noise signal. Specifically, it extracts eigenvectors and eigenvalues from the covariance matrix, and further extracts dominant eigen-directions and eigen-gains. This extracted subset of parameters captures the critical spatial information needed for comfort noise generation while dramatically reducing data transmission requirements.
Solution Approach 2:
The patent transforms the complete spatial parameter set into a reduced parameter representation by changing from transmitting full covariance matrices to transmitting only dominant eigenvectors and eigenvalues. This parameter transformation maintains perceptual quality by focusing on the most significant spatial characteristics while reducing data volume.
3Quantity of substance
If downmix signals with fewer channels are used, then data transmission is optimized, but reconstruction of multi-channel signal becomes more complex
Solution Approach 1:
The patent applies dynamics by making the reconstruction process adaptive to the type of frame being processed. During active frames, the system uses signal-based reconstruction methods, while during inactive frames, it switches to comfort noise generation using pre-computed spatial parameters. This dynamic adaptation optimizes both reconstruction quality and computational efficiency for different operational conditions.
Solution Approach 2:
The patent introduces downmix signals as an intermediary representation between the original multi-channel input and the reconstructed multi-channel output. The downmix signal serves as a compressed intermediate form that retains essential spatial information, which is then expanded back to full multi-channel format using SPAR parameters, simplifying the overall reconstruction process.
Data Source
AI summary
The present document describes a method (600) for encoding a multi-channel input signal (101) which comprises N different channels. The method (600) comprises, for a current frame of a sequence of frames, determining (601) whether the current frame is an active frame or an inactive frame using a signal and/or a voice activity detector, and determining (602) a downmix signal (103) based on the multi-channel input signal (101), wherein the downmix signal (103) comprises N channels or less. In addition, the method (600) comprises determining (603) upmixing metadata (105) comprising a set of parameters for generating, based on the downmix signal (103), a reconstructed multi-channel signal (111) comprising N channels, wherein the upmixing metadata (105) is determined in dependence of whether the current frame is an active frame or an inactive frame. The method (600) further comprises encoding (604) the upmixing metadata (105) into a bitstream.