ADPCM Packet Loss Concealment via Combined Prediction Error
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet loss concealment methods in ADPCM codecs for professional wireless microphones introduce significant latency and audible transients due to time-warping and re-phasing techniques, which are not suitable for real-time wireless audio transmission.
Innovation Solution
A method that combines the difference between a substitute signal and the prediction error with the dequantized prediction error to create a combined prediction error, which is then used to adapt decoder parameters and generate an output signal during transition periods, eliminating additional latency and improving audio quality by minimizing error audibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If time-warping and re-phasing techniques are used to conceal packet loss, then audio quality is improved, but latency increases significantly
Solution Approach 1:
The patent extracts and removes the time-warping and re-phasing processing steps from the packet loss concealment method. By eliminating these specific operations, the system achieves packet loss concealment without the associated latency penalty, directly resolving the contradiction between audio quality improvement and time loss.
Solution Approach 2:
Instead of applying complex time-warping and re-phasing operations to improve audio quality, the patent inverts the approach by using a simpler substitution method that replaces the corrupted packet data with reconstructed data from available information, thereby achieving acceptable audio quality without the latency overhead of complex processing.
2Object-generated harmful factors
If time-warping and re-phasing techniques are applied to re-align signal phases, then transient abruptness is reduced, but processing delay becomes unacceptable
Solution Approach 1:
The patent removes the time-warping and re-phasing operations that cause processing delay. By extracting these specific harmful processing steps, the system achieves transient concealment through a simpler mechanism that does not introduce unacceptable delays.
Solution Approach 2:
The patent performs preliminary reconstruction of the substitute signal using available data before the actual packet loss occurs. By preparing the reconstruction data in advance through adaptive prediction and substitution, the system avoids the need for complex post-processing time-warping operations, thereby reducing processing delay while still concealing abrupt transients.
3Reliability
If ADPCM decoder parameters are adapted independently to encoded prediction error during dropout, then error propagation is minimized, but abrupt transients occur at transition edges
Solution Approach 1:
The patent merges the substitute signal generation with the adaptive parameter adaptation process. By combining the prediction error substitution with continuous adaptation of decoder parameters to match the substitute signal characteristics, the system minimizes abrupt transients at transition edges while maintaining error propagation control.
Solution Approach 2:
The patent implements dynamic adaptation of decoder parameters during the packet loss period. The parameters are continuously adjusted to match the characteristics of the substitute signal, creating a smooth transition that reduces abrupt transients while maintaining reliable error control throughout the dropout period.
Data Source
AI summary
A method of packet loss concealment in an adaptive differential pulse-code modulation (ADPCM) codec with a packet loss compensation (PLC) circuit is provided. The method provides a predetermined transition period between a correct signal (xdec) and a substitute signal (xPLC) and a difference (dPLC,m) between the substitute signal (xPLC,m) and a computed prediction signal (xpred,m) is combined with a dequantized prediction error (ddec,m) to receive a dequantized combined prediction error (dcomb,m) which is added to a predicted signal (xpred,m,) to provide a combined transition signal (xcomb,m) as basis for an output signal (xout−xcomb) during the predetermined transition period for adapting all decoder parameters.