Multiple Description Audio Encoding for Packet Loss Resistance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio encoders, such as Opus, suffer from poor packet loss resistance, leading to degraded audio quality when network packet loss occurs, as they rely solely on packet loss concealment techniques.

Innovation Solution

Implement Multiple Description Coding (MDC) by generating multiple bitstreams from a single frame of input signal, ensuring any received bitstream can decode correctly, reducing reliance on packet loss concealment and improving audio quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If Multiple Description Coding is implemented to improve packet loss resistance, then reliability is improved, but device complexity increases

Engineering Contradiction:
Improvepacket loss resistanceVSAvoidencoding complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The encoding process is segmented into multiple independent description streams (first description stream, second description stream, etc.), each capable of being decoded independently. This segmentation allows the system to improve reliability through redundancy while managing complexity by processing each stream separately using standard encoding tools.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The encoding apparatus is designed to generate multiple description streams that can serve universal decoding purposes. Each stream can be decoded independently using standard decoders, making the system multi-functional and adaptable to different network conditions without requiring specialized decoding equipment for each stream type.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If Multiple Description Coding generates multiple bitstreams, then packet loss resistance improves, but loss of information increases due to redundant encoding

Engineering Contradiction:
Improvepacket loss resistanceVSAvoiddata redundancy
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The system generates multiple description streams with varying levels of redundancy. Under normal conditions, all streams are transmitted providing full protection. However, the system can adaptively reduce redundancy by transmitting fewer streams when network conditions improve, thus reducing information loss while maintaining packet loss resistance when needed.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The encoding process allows dynamic adjustment of parameters such as the number of description streams, quantization precision, and bit allocation across streams. By changing these parameters based on network conditions and source characteristics, the system optimizes the balance between reliability and information efficiency, reducing redundant encoding when full redundancy is not required.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS20250226851A1Encoding method, decoding method, encoding apparatus, decoding apparatus, electronic device, and storage medium
Publication Date: 2025.07.10 DOUYIN VISION CO LTD
  • US20250226851A1 patent drawing
  • US20250226851A1 patent drawing
  • US20250226851A1 patent drawing

AI summary

An encoding method, a decoding method, an encoding apparatus, a decoding apparatus, an electronic device, and a storage medium. The encoding method includes: obtaining at least two multiple description signals according to a to-be-coded signal, the at least two multiple description signals including a first multiple description signal and a second multiple description signal; performing first quantization on the first multiple description signal to obtain a first candidate quantized signal set, and performing second quantization on the second multiple description signal to obtain a second candidate quantized signal set; obtaining a first quantized signal corresponding to the first multiple description signal and a second quantized signal corresponding to the second multiple description signal according to the first candidate quantized signal set and the second candidate quantized signal set; and encoding the first quantized signal and the second quantized signal to generate a target bitstream corresponding to the to-be-coded signal.