Speech Encoder Error Correction via Adaptive Bit Rate FEC
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech coding technologies face challenges in providing effective error correction for packetized data transmission, particularly in real-time communication, due to increased complexity and delay issues with media independent FEC and the need for parallel encoders and decoders.
Innovation Solution
A method is introduced that generates error correction data by encoding the residual signal at a lower bit rate and using it to create an error correction bitstream, which is delayed and multiplexed with the output bitstream, allowing for adaptive error correction based on frame sensitivity to packet losses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If media independent FEC is used to provide error correction for packetized data, then error correction capability is improved, but device complexity and delay increase
Solution Approach 1:
The patent combines the FEC encoding operation with the existing speech encoder by integrating the FEC generator into the encoder architecture. The FEC generator uses the same analysis results (LPC parameters, pitch, energy) as the main encoder, merging two functions into a unified processing path rather than operating as separate parallel systems.
Solution Approach 2:
The encoder is designed to serve multiple functions: it performs speech encoding for transmission and simultaneously generates FEC data for error correction. The same analysis modules (LPC analysis, pitch detection, energy calculation) serve both the main speech signal and the FEC generation, making the encoder a multi-functional device that reduces overall system complexity.
2Reliability
If parallel encoders and decoders are used for FEC, then error correction is improved, but delay increases
Solution Approach 1:
The patent applies preliminary action by generating FEC data in advance for future packets. The FEC generator creates error correction data for packet n using analysis results from packet n-1, so that when packet n is transmitted and potentially lost, the FEC data is already prepared and can be immediately applied at the receiver without requiring parallel decoding operations.
Solution Approach 2:
The patent segments the encoding process into distinct functional components: speech encoding for the main signal and separate FEC generation for error correction. This segmentation allows each component to operate independently using the same analysis results, avoiding the need for parallel encoders while still providing comprehensive error correction capability.
3Quantity of substance
If lower bit rate encoding is used for FEC, then overhead is reduced, but error correction effectiveness may deteriorate
Solution Approach 1:
The patent applies parameter changes by adapting the encoding rate dynamically based on channel conditions and packet importance. The system can switch between different encoding modes (e.g., full rate for critical packets, reduced rate for less critical packets) and adjusts parameters such as quantization precision and codebook size to optimize the trade-off between overhead and error correction effectiveness.
Solution Approach 2:
The patent applies partial action by providing error correction only where and when needed. Instead of uniformly applying full-rate encoding to all packets, the system selectively applies FEC to packets based on their importance and the current channel conditions, reducing overall overhead while maintaining sufficient error correction effectiveness for critical communication segments.
Data Source
AI summary
A method system and program for encoding and decoding a speech signal including error correction data. The method comprises: receiving a speech signal comprising successive frames, for each of a plurality of frames of the speech signal, analysing the speech signal to determine side information and a residual signal, encoding the residual signal at a first bit rate, and generating an output bitstream based on the residual signal encoded at the first bit rate, and for at least one of the plurality of frames of the speech signal, encoding the residual signal at a second bit rate that is lower than the first bit rate; and generating error correction data based on the residual signal encoded at the second bit rate.