Predictive Coherence Vector Coding for Stable Stereo DTX Audio
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating comfort noise in stereo DTX systems fail to adapt to changes in the stereo image of background noise and lack synchronization between channels, leading to audible artifacts and an inconsistent stereo image.
Innovation Solution
A method is introduced to encode a channel coherence parameter using a predictive scheme with adaptive inter-frame prediction and variable bit rate entropy coding, where coherence vectors are encoded and decoded to maintain channel coherence and adapt to changing environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional comfort noise generation methods are used in stereo DTX systems, then bandwidth is reduced during speech pauses, but the stereo image becomes inconsistent and audible artifacts occur
Solution Approach 1:
The coherence vector is divided into multiple frequency bands, with each band encoded separately using predictive coding. This segmentation allows independent optimization of prediction for each frequency region, improving the accuracy of stereo image representation while maintaining low bit rate.
Solution Approach 2:
The system dynamically adapts the prediction model based on the actual audio signal characteristics. The predictive coding scheme adjusts to changing stereo images in real-time, allowing the comfort noise generator to maintain consistent spatial representation even when the background noise characteristics change.
2Quantity of substance
If predictive coding with variable bit rate is used, then bit rate efficiency is improved, but error propagation increases
Solution Approach 1:
The coherence vector is segmented into frequency bands, and prediction is performed independently in each band. This segmentation limits the scope of error propagation to individual frequency regions rather than affecting the entire signal, while still achieving high bit rate efficiency through predictive coding.
Solution Approach 2:
The system uses a robust prediction model that anticipates and compensates for potential errors. By establishing accurate predictions beforehand based on the natural correlations in audio signals, the system cushions against error propagation when bit rate truncation occurs.
3Measurement precision
If channel coherence parameters are encoded with high precision, then spatial representation quality is improved, but transmission bandwidth increases
Solution Approach 1:
The system changes the representation parameters by using predictive coding to encode coherence vectors. Instead of transmitting raw precision values, the system transmits prediction residuals, which significantly reduces the number of bits needed while maintaining high spatial representation quality through the predictive model.
Solution Approach 2:
The predictive coding scheme creates a copy of the coherence vector based on the natural correlations between frequency bands and temporal patterns. This copied prediction is then used to reconstruct the signal, achieving high precision spatial representation with reduced bit rate by leveraging the redundancy in the data.
Data Source
AI summary
A method, system, and computer program to encode and decode a channel coherence parameter applied on a frequency band basis, where the coherence parameters of each frequency band form a coherence vector. The coherence vector is encoded and decoded using a predictive scheme followed by a variable bit rate entropy coding.