Audio Signal Classification Using Tonal and Spectral Tilt Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal classification methods are complex and require significant calculation, making them inefficient for determining the type of audio signals, especially under mid-to-low bit rates.
Innovation Solution
A method and device for audio signal classification that focuses on obtaining tonal characteristic parameters in sub-bands to determine the type of audio signals, reducing the complexity and calculation required by using a tone obtaining module and classification module.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple characteristic parameters (harmony, noise, tail, drag out, rhythm) are calculated for audio signal classification, then the classification accuracy is improved, but the calculation complexity and processing time increase significantly
Solution Approach 1:
The patent extracts only the essential tonal characteristic parameters from the audio signal, specifically focusing on spectral tilt and other key tonal features, while discarding unnecessary parameters like harmony, noise, tail, drag out, and rhythm. This extraction approach maintains sufficient classification accuracy while significantly reducing calculation complexity and processing requirements.
Solution Approach 2:
The patent applies different analysis depths to different parts of the audio signal processing pipeline. Instead of uniformly analyzing all aspects of the signal with equal detail, it focuses computational resources on extracting locally relevant tonal characteristics in specific frequency sub-bands, thereby achieving efficient classification without exhaustive analysis of all signal properties.
2Reliability
If comprehensive characteristic parameters are used for audio signal classification, then the classification reliability is improved, but the processing time and computational load increase
Solution Approach 1:
The patent extracts only the critical tonal characteristic parameters needed for reliable classification, specifically spectral tilt and other key tonal features in sub-bands. By extracting only these essential parameters and ignoring redundant information, the system achieves reliable classification results with significantly reduced processing time and computational load compared to comprehensive parameter analysis.
3Adaptability or versatility
If multiple aspect characteristic parameters are calculated, then the classification completeness is improved, but the energy consumption and computational resources increase
Solution Approach 1:
The patent extracts only the necessary tonal characteristic parameters from the audio signal, focusing on spectral tilt and other key tonal features in sub-bands. This selective extraction approach ensures sufficient classification completeness for distinguishing voice and music types while minimizing computational energy consumption by avoiding calculation of unnecessary parameters.
Solution Approach 2:
The patent changes the parameter set used for classification from comprehensive multi-aspect parameters to a focused set of tonal characteristic parameters. By transforming the classification approach to use only tonal parameters like spectral tilt, the system maintains effective classification completeness while significantly reducing the computational energy required for parameter calculation.
Data Source
AI summary
The present invention discloses a method and a device for audio signal classification, and relates to the field of communications technologies, which solve a problem of high complexity of type classification of audio signals in the prior art. In the present invention, after an audio signal to be classified is received, a tonal characteristic parameter of the audio signal to be classified, where the tonal characteristic parameter of the audio signal to be classified is in at least one sub-band, is obtained, and a type of the audio signal to be classified is determined according to the obtained characteristic parameter. The present invention is mainly applied to an audio signal classification scenario, and implements audio signal classification through a relatively simple method.