Critical Speech Frame FEC Using Channel Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital voice communication systems over packet-switched networks face challenges in maintaining speech quality due to packet loss, with traditional sender-based forward error correction (FEC) schemes increasing bandwidth and delay, and receiver-based methods failing to effectively mask important frame losses.
Innovation Solution
A source and channel-controlled FEC scheme that dynamically adjusts to identify critical speech frames and create redundant copies only when necessary, using channel state feedback to optimize data rate and minimize delay, ensuring high speech quality without additional latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sender-based forward error correction (FEC) schemes are used to protect speech frames, then speech quality is improved, but bandwidth consumption and delay increase
Solution Approach 1:
The patent applies partial action by implementing selective FEC protection only for critical speech frames (such as unvoiced frames and voiced-unvoiced transition frames) rather than protecting all frames. This selective approach provides necessary error protection for important frames while avoiding the bandwidth overhead of protecting every frame, thus resolving the contradiction between speech quality and bandwidth consumption.
Solution Approach 2:
The patent implements dynamic adaptation by using channel state feedback to adjust the level of FEC protection based on current channel conditions. When channel quality is good, less protection is applied; when channel quality degrades, more protection is applied to critical frames. This dynamic approach optimizes the trade-off between speech quality and bandwidth usage according to real-time network conditions.
2Reliability
If sender-based forward error correction (FEC) schemes are used to protect speech frames, then speech quality is improved, but end-to-end delay increases
Solution Approach 1:
By applying FEC protection selectively only to critical frames rather than all frames, the patent reduces the overall processing and transmission overhead. This partial protection strategy maintains speech quality for important frames while minimizing the delay introduced by FEC operations, thus resolving the contradiction between speech quality and end-to-end delay.
Solution Approach 2:
The patent identifies critical frames in advance using channel state feedback and applies protection only to those frames before transmission. This preliminary identification and selective protection approach avoids the delay of processing all frames uniformly, allowing critical frames to be protected efficiently while non-critical frames are transmitted with minimal overhead.
3Quantity of substance
If receiver-based packet loss concealment is used, then bandwidth is saved, but important frame losses cannot be effectively masked
Solution Approach 1:
The patent introduces forward error correction codes as an intermediary mechanism between the sender and receiver. Instead of relying solely on receiver-based concealment, the FEC codes provide redundant information that enables the receiver to recover lost critical frames. This intermediary approach enhances speech quality for important frames while maintaining bandwidth efficiency through selective protection.
4Reliability
If redundant copies of all speech frames are transmitted, then packet loss resilience is improved, but data rate increases
Solution Approach 1:
The patent applies partial action by transmitting redundant copies only for critical speech frames identified through channel state feedback, rather than creating redundant copies of all frames. This selective redundancy strategy improves packet loss resilience for important frames while keeping the data rate increase moderate, thus resolving the contradiction between reliability and data rate.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for preventing the loss of information within a speech frame is described. A first speech frame to be encoded is selected. A determination is made as to whether or not a second speech frame is a critical speech frame based on the information within the second speech frame and one or more adjacent speech frames. At least a part of an encoded version of the second speech frame is created according to a selected forward error correction (FEC) mode if the second speech frame is a critical speech frame. The first speech frame and the at least a part of the encoded version of the second speech frame are transmitted.