Auditory Event Boundary Detection in Real-Time Audio Gain Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio dynamic range control methods are limited in effectively managing auditory events, leading to audible artifacts and inefficiencies in processing complex audio signals, as they often rely on computationally demanding techniques that are not practical for real-time applications.
Innovation Solution
The method involves performing auditory scene analysis by identifying auditory events through spectral content changes and applying dynamic gain modifications based on these events, using techniques such as spectral analysis and transformation into a perceptual loudness domain to control audio dynamics processing parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If auditory scene analysis is performed using traditional computational methods, then auditory events can be identified, but the computational complexity becomes too high for practical real-time applications
Solution Approach 1:
The patent replaces traditional mechanical/computational signal processing methods with a neural network-based system. The neural network is trained offline to perform auditory scene analysis, substituting complex real-time computations with pre-learned patterns that can be evaluated much more efficiently during actual audio processing.
Solution Approach 2:
The neural network is trained in advance on large datasets of audio signals and corresponding auditory event annotations. This preliminary training phase allows the system to learn complex auditory patterns beforehand, so that during real-time operation, only inference is required rather than full computational analysis.
2Object-affected harmful factors
If dynamic gain modifications are applied based on auditory events, then audible artifacts are reduced, but the processing time and computational load increase
Solution Approach 1:
The patent uses a neural network to rapidly classify audio segments and identify auditory events, replacing traditional time-consuming signal processing algorithms. Once events are identified by the neural network, gain modifications are applied efficiently based on the classified event types, reducing overall processing time while maintaining artifact reduction.
Solution Approach 2:
The audio signal is divided into overlapping frames, and the neural network processes each frame independently to identify auditory events. This segmentation allows for parallel processing of different time segments, reducing total processing time while still applying gain modifications at the appropriate granular level to minimize artifacts.
3Measurement precision
If spectral analysis is performed on the entire frequency band, then comprehensive auditory event detection is achieved, but the computational demand increases significantly
Solution Approach 1:
The patent applies different processing strategies to different frequency regions based on their characteristics. The neural network is trained to identify that certain frequency bands contain more relevant auditory events than others, allowing the system to focus computational resources on critical frequency regions while using simpler processing for less important bands.
Solution Approach 2:
The frequency spectrum is divided into multiple bands, and the neural network processes each band separately rather than analyzing the entire spectrum uniformly. This frequency-domain segmentation reduces the computational complexity of spectral analysis while maintaining detection accuracy by allowing parallel processing of frequency bands.
Data Source
AI summary
In some embodiments, a method for processing an audio signal in an audio processing apparatus is disclosed. The method includes receiving an audio signal and a parameter, the parameter indicating a location of an auditory event boundary. An audio portion between consecutive auditory event boundaries constitutes an auditory event. The method further includes applying a modification to the audio signal based in part on an occurrence of the auditory event. The parameter may be generated by monitoring a characteristic of the audio signal and identifying a change in the characteristic.