Adaptive Audio Quantization for Error-Resilient Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Predictive audio quantization in codecs is prone to error propagation, leading to significant losses in frames during transmission or storage errors, as it heavily relies on preceding frames, limiting the usage of non-predictive quantization to prevent long error runs.
Innovation Solution
A method that dynamically selects between predictive and non-predictive quantization based on error thresholds and counters, allowing non-predictive quantization when errors are below a certain inaudible level, thereby increasing robustness against errors while maintaining performance in clean channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If predictive quantization is used to improve compression ratio and quantization performance, then coding efficiency is improved, but error propagation occurs leading to loss of up to five frames in case of transmission errors
Solution Approach 1:
The system dynamically adjusts the prediction coefficient based on the reliability of the previous frame. When error propagation is detected or suspected, the prediction coefficient is reduced or set to zero, effectively switching from predictive to non-predictive quantization for that frame. This dynamic adaptation allows the system to maintain high coding efficiency when conditions permit while ensuring error resilience when transmission errors occur.
Solution Approach 2:
The invention changes the prediction parameter (prediction coefficient) based on error conditions. By monitoring frame reliability and adjusting the prediction coefficient accordingly, the system can transition between predictive and non-predictive quantization modes, resolving the contradiction between coding efficiency and error resilience.
2Reliability
If non-predictive quantization is used to prevent error propagation, then error resilience is improved, but quantization performance deteriorates for the majority of frames
Solution Approach 1:
Instead of applying non-predictive quantization uniformly to all frames, the system applies it locally only to frames where error propagation is detected or suspected. The prediction coefficient is adjusted on a per-frame basis, allowing predictive quantization to be used for most frames (maintaining high quantization performance) while switching to non-predictive quantization only when necessary for error resilience.
3Reliability
If forced selection of non-predictive quantization is used to limit consecutive predictive frames, then error propagation is reduced, but overall quantization performance decreases in clean channels
Solution Approach 1:
Rather than forcing a fixed limit on consecutive predictive frames, the system dynamically determines when to switch to non-predictive quantization based on actual error conditions detected in the received frames. This allows the system to maintain predictive quantization for as long as possible in clean channels (preserving performance) while automatically switching to non-predictive mode when errors are detected (controlling propagation).
Data Source
Figure 1
Figure 2
Figure 3
AI summary
For supporting a selection of a predictive or non-predictive quantization in the scope of an audio signal coding, it is determined whether an error resulting with a non-predictive quantization of an audio signal segment lies below a predetermined threshold value. An audio signal segment quantized with the non-predictive quantization is provided as a part of an encoded audio signal at least in case it is determined that the error resulting with the non-predictive quantization of the audio signal segment lies below a predetermined threshold value. Otherwise, an audio signal segment quantized with predictive quantization is provided as a part of an encoded audio signal.