Audio Decoder Error Concealment Using Time Domain Excitation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio decoding technologies face challenges in maintaining good audio quality when audio frames are lost during transmission over unreliable channels, as they often require extensive buffering, which degrades real-time capabilities and consumes significant memory.
Innovation Solution
The proposed solution involves an audio decoder that uses a time domain excitation signal to perform error concealment by combining an extrapolated time domain excitation signal with a noise signal and filtering it using linear-prediction-coding parameters to generate error concealment audio information, even when audio frames are encoded in a frequency domain representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extensive buffering is used to handle lost audio frames, then audio quality is maintained, but real-time capabilities are degraded and memory consumption increases
Solution Approach 1:
The patent applies preliminary action by using linear prediction to extrapolate audio frames before they are actually lost. The system pre-computes prediction coefficients and prepares concealment signals in advance, so when a frame is lost, the concealment can immediately use the pre-calculated predictions without waiting for buffering or retransmission, thus maintaining real-time capabilities while preserving audio quality
Solution Approach 2:
The patent introduces an intermediary mechanism - the linear prediction model - that acts as a mediator between lost audio frames and the audio output. Instead of directly buffering lost frames, the system uses prediction coefficients and excitation signals as intermediaries to generate concealment signals that approximate the lost content, reducing the need for extensive buffering while maintaining quality
2Reliability
If extensive buffering is used to handle lost audio frames, then audio quality is maintained, but memory consumption increases
Solution Approach 1:
The system pre-computes and stores only essential prediction coefficients and excitation signal parameters rather than buffering entire audio frames. This preliminary preparation of compact representation data reduces memory requirements from storing full frame buffers to storing only the parameters needed for reconstruction, significantly reducing memory consumption while maintaining audio quality
Solution Approach 2:
The patent transforms the audio signal into a parameter-based representation using linear prediction coefficients and excitation signals. By working with these compressed parameters instead of raw audio data, the system reduces the quantity of data that needs to be buffered and stored in memory, achieving efficient error concealment with minimal memory consumption
3Loss of time
If time domain excitation signal is used for error concealment in frequency domain codec, then buffering requirements are reduced, but implementation complexity increases
Solution Approach 1:
The patent introduces a time domain excitation signal as an intermediary that bridges the frequency domain codec structure and time domain error concealment. The excitation signal serves as a mediator that can be processed in time domain for prediction while the main audio processing remains in frequency domain, reducing buffering requirements without requiring a complete system redesign
Solution Approach 2:
The patent segments the audio processing into distinct frequency domain and time domain components. The main audio coding remains in frequency domain while the error concealment functionality is separated into a time domain excitation signal processing module. This segmentation allows each part to operate independently with optimal buffering requirements, reducing overall system complexity despite the dual-domain approach
Data Source
AI summary
An audio decoder for providing a decoded audio information on the basis of an encoded audio information includes an error concealment configured to provide an error concealment audio information for concealing a loss of an audio frame following an audio frame encoded in a frequency domain representation using a time domain excitation signal.