Immersive Audio Downmixing With Adaptive Gain Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing immersive voice and audio services (IVAS) codecs face challenges in reconstructing immersive audio scenes accurately due to imperfect decorrelation and limited quantization ranges in passive downmix schemes, leading to audio artifacts and spatial distortion.
Innovation Solution
Implementing adaptive downmix strategies that combine passive and active downmix techniques, adjusting downmixing gains and scaling factors to minimize prediction errors and ensure better parameter estimation, while maintaining energy preservation across channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If passive downmix scheme is used, then implementation is simple, but reconstruction accuracy deteriorates due to imperfect decorrelation and limited quantization ranges
Solution Approach 1:
The patent implements dynamic switching between passive and active downmix schemes based on signal properties and voice activity detection. The system adapts the downmix strategy in real-time, using passive downmix when simple and effective, and active downmix when higher reconstruction accuracy is needed, thus resolving the contradiction between implementation simplicity and reconstruction accuracy.
Solution Approach 2:
The patent changes the downmixing parameters (gains, scaling factors) adaptively based on signal characteristics. By adjusting these parameters dynamically, the system can achieve better reconstruction accuracy when needed while maintaining simplicity when possible, resolving the contradiction between ease of implementation and measurement precision.
2Measurement precision
If active downmix scheme is used, then prediction error is reduced, but computational complexity increases
Solution Approach 1:
The system dynamically selects between passive and active downmix schemes based on voice activity detection and signal properties. Active downmix with its lower prediction error is used selectively when beneficial, while passive downmix is used otherwise, thus reducing overall computational complexity while maintaining prediction accuracy when needed.
Solution Approach 2:
The patent applies active downmix processing only partially - specifically during voice activity periods or when signal characteristics indicate benefit - rather than continuously. This partial application reduces computational complexity while still achieving prediction error minimization when it matters most.
3Measurement precision
If downmixing gains are increased to minimize prediction error, then audio quality improves, but spatial distortion increases due to quantization limits
Solution Approach 1:
The patent adaptively adjusts downmixing gains and scaling factors based on signal properties and voice activity. By optimizing these parameters dynamically rather than using fixed high gains, the system achieves good audio quality while avoiding the spatial distortion that results from excessive quantization of large gain values.
Solution Approach 2:
The system uses voice activity detection and signal property analysis as feedback to adjust downmixing parameters in real-time. This feedback mechanism ensures that gains are increased only when necessary for audio quality, preventing the spatial distortion that would result from consistently high gain values.
4Adaptability or versatility
If adaptive downmix switching is implemented, then audio quality across different conditions improves, but device complexity increases
Solution Approach 1:
The patent implements adaptive switching between downmix schemes based on voice activity detection and signal property analysis. This dynamic adaptation improves audio quality across different conditions by selecting the appropriate downmix mode, while the switching logic is designed to be computationally efficient to minimize the increase in device complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is an audio signal encoding/decoding method that uses an encoding downmix strategy applied at an encoder that is different than a decoding re-mix/upmix strategy applied at a decoder. Based on the type of downmix coding scheme, the method comprises: computing input downmixing gains to be applied to the input audio signal to construct a primary downmix channel; determining downmix scaling gains to scale the primary downmix channel; generating prediction gains based on the input audio signal, the input downmixing gains and the downmix scaling gains; determining residual channel(s) from the side channels by using the primary downmix channel and the prediction gains to generate side channel predictions and subtracting the side channel predictions from the side channels; determining decorrelation gains based on energy in the residual channels; encoding the primary downmix channel, the residual channel(s), the prediction gains and the decorrelation gains; and sending the bitstream to a decoder.