Audio Error Concealment Using Unified Comfort Noise Tracing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio signal encoding and decoding technologies face challenges in effectively managing signal fade-out during error concealment in switched audio coding systems, particularly in maintaining a pleasant noise characteristic and avoiding computational overhead during burst packet losses.
Innovation Solution
An apparatus and method for decoding audio signals that trace and apply a common comfort noise level in the excitation domain, ensuring a consistent noise level during burst packet losses, and adjust the LPC synthesis gain to match the excitation domain energy, while fading the TCX LTP gain to zero to prevent tonal component distortion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate noise level tracing is implemented for each core coder (ACELP/TCX), then accurate noise tracking for each domain is achieved, but computational overhead and system complexity increase
Solution Approach 1:
The patent merges separate noise level tracing operations for ACELP and TCX domains into a single unified noise level tracing unit. This unit traces noise level in one domain (e.g., excitation domain) and the traced noise level is commonly applied to both ACELP and TCX core coders, eliminating the need for separate tracing operations while maintaining accurate noise tracking for error concealment in both coding modes
Solution Approach 2:
The unified noise level tracing unit serves multiple functions: it traces noise level for both ACELP and TCX domains, provides common comfort noise level for burst packet loss concealment, and adapts to different coding modes without requiring separate tracing mechanisms. This multi-functional approach reduces system complexity while preserving measurement precision
2Stability of the object's composition
If LPC synthesis gain is adjusted to match excitation domain energy, then consistent noise level is achieved during error concealment, but additional computational steps are required
Solution Approach 1:
The patent implements feedback control where the LPC synthesis gain is adjusted based on the traced noise level from the excitation domain. The traced noise level serves as feedback to modify the LPC synthesis gain, ensuring that the synthesized speech maintains consistent noise characteristics with the original signal during error concealment, while the feedback loop is efficiently integrated into the existing decoding process
3Object-generated harmful factors
If TCX LTP gain is faded to zero during error concealment, then tonal component distortion is prevented, but signal fidelity may be reduced
Solution Approach 1:
The patent applies dynamic control to the TCX LTP gain parameter during error concealment. The LTP gain is faded to zero when error concealment is activated to prevent tonal distortion from incorrect LTP prediction, and restored when normal decoding resumes. This dynamic adjustment adapts the signal processing behavior to the current operational state, preventing harmful artifacts while preserving signal fidelity during normal operation
Data Source
AI summary
An apparatus for decoding an audio signal includes a receiving interface, wherein the receiving interface is configured to receive a first frame and a second frame. Moreover, the apparatus includes a noise level tracing unit for determining noise level information being represented in a tracing domain. Furthermore, the apparatus includes a first reconstruction unit for reconstructing a third audio signal portion of the audio signal depending on the noise level information and a second reconstruction unit for reconstructing a fourth audio signal portion depending on noise level information being represented in the second reconstruction domain.