Adaptive Frame Loss Concealment for Speech and Music Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current frame loss concealment methods in digital communication systems either cause audible artifacts in speech signals or produce distortion in music signals, failing to perform well for both speech and music, which are dominant in audio signals used in movie, TV, and radio applications.
Innovation Solution
An adaptive audio decoding system that employs multiple frame loss concealment methods, analyzing the previously-decoded audio signal to select the most suitable method for either speech or music, based on classification, to perform frame loss concealment operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the frame repeat method is used for frame loss concealment, then the implementation is simple and works well for busy-sounding audio signals, but it causes audible artifacts (clicks) in speech signals and distortion in nearly periodic signals
Solution Approach 1:
The patent implements a dynamic frame loss concealment system that automatically adapts the concealment method based on real-time analysis of audio signal characteristics. The system switches between different concealment strategies (frame repeat, waveform substitution, or silence insertion) depending on whether the detected signal is speech, music, or other types, thereby resolving the contradiction between simple implementation and avoidance of audible artifacts across different signal types
Solution Approach 2:
The system changes the operational parameters of the frame loss concealment process by selecting different concealment methods based on signal classification. For speech signals, it uses waveform substitution with pitch prediction; for music signals, it uses frame repeat; and for unknown signal types, it uses silence insertion. This parameter adaptation allows the system to maintain simplicity while avoiding artifacts in specific signal contexts
2Object-affected harmful factors
If sophisticated waveform extrapolation methods are used for frame loss concealment, then the audio quality is improved for speech signals, but the system complexity increases and performance deteriorates for music signals
Solution Approach 1:
The patent segments the frame loss concealment process into distinct methods, each optimized for specific signal types. Instead of using a single complex extrapolation method for all signals, the system divides the task: waveform substitution for speech, frame repeat for music, and silence insertion for unknown types. This segmentation reduces overall system complexity while maintaining high audio quality for each signal category
Solution Approach 2:
The system achieves universality by implementing a multi-functional frame loss concealment apparatus that can handle multiple signal types (speech, music, and unknown) using a single integrated framework. The classifier and adaptive selection mechanism allow one system to perform multiple specialized functions, avoiding the need for separate complex systems for each signal type while maintaining optimal performance across all categories
3Adaptability or versatility
If a single frame loss concealment method is used for all audio signals, then the system is simple, but it cannot perform well for both speech and music signals simultaneously
Solution Approach 1:
The patent employs a dynamic adaptive mechanism that continuously monitors audio signal characteristics and adjusts the concealment method accordingly. The signal classifier analyzes features such as periodicity and spectral characteristics to determine whether the signal is speech, music, or other types, and automatically selects the appropriate concealment strategy. This dynamic adaptation provides versatility across signal types while keeping the underlying system architecture relatively simple
Solution Approach 2:
The system incorporates feedback through a signal classification mechanism that continuously analyzes the audio signal properties and uses this information to guide the frame loss concealment process. The classifier provides feedback about the signal type to the concealment algorithm, which then adjusts its behavior accordingly. This feedback loop enables the system to adapt to different signal types without requiring complex manual configuration or multiple fixed systems
Data Source
AI summary
A system and method for performing frame loss concealment (FLC) when portions of a bit stream representing an audio signal are lost within the context of a digital communication system. The system and method utilizes a plurality of different FLC techniques, wherein each technique is tuned or designed for a different kind of audio signal. When a frame is lost, a previously-decoded audio signal corresponding to one or more previously-received good frames is analyzed. Based on the result of the analysis, the FLC technique that is most likely to perform well for the previously-decoded audio signal is chosen to perform the FLC operation for the current lost frame. In one implementation, the plurality of different FLC techniques include an FLC technique designed for music, such as a frame repeat FLC technique, and an FLC technique designed for speech, such as a periodic waveform extrapolation (PWE) technique.