Harmonic Source Separation with Time-Frequency Masking for Audio ID
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current audio processing technologies face challenges in effectively separating and enhancing harmonic sources from audio signals, which is crucial for audio identification, authentication, and improving audio fingerprinting, as they often struggle with mitigating background noise and efficiently categorizing media.
Innovation Solution
The method involves an audio analyzer that processes media signals to extract and enhance harmonic sources using techniques like magnitude spectrogram analysis, Fourier transforms, and time-frequency masking, allowing for the separation of harmonic and percussive components, and a database-driven system for classification and identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If audio signals are processed to separate harmonic and percussive components, then audio identification accuracy is improved, but processing complexity increases
Solution Approach 1:
The audio signal is segmented into harmonic and percussive components using spectral decomposition. The magnitude spectrogram is divided into harmonic regions (periodic patterns) and percussive regions (transient events), allowing separate processing and enhancement of each component to improve identification accuracy.
Solution Approach 2:
A time-frequency mask serves as an intermediary tool to selectively enhance harmonic components while attenuating percussive components. This mask is derived from autocorrelation analysis and spectral characteristics, enabling precise control over which frequency-time regions are enhanced for identification purposes.
2Measurement precision
If harmonic source enhancement is applied to mitigate background noise, then audio fingerprinting quality is improved, but computational resources required increase
Solution Approach 1:
The method changes parameters of the audio signal in the time-frequency domain by applying a time-frequency mask that selectively amplifies harmonic components. This involves modifying the magnitude spectrogram by enhancing specific frequency bins associated with harmonic patterns while suppressing others, thereby improving fingerprinting quality with controlled computational effort.
Solution Approach 2:
Autocorrelation analysis is performed preliminarily to identify periodic patterns and determine the time-frequency mask before actual enhancement. This preliminary characterization of the audio signal's harmonic structure enables efficient subsequent enhancement without requiring exhaustive computational search during the enhancement phase.
3Measurement precision
If audio signals are decomposed into harmonic and percussive components, then media classification accuracy is improved, but processing time increases
Solution Approach 1:
The audio signal is segmented into harmonic and percussive components through spectral decomposition and autocorrelation analysis. This segmentation enables independent processing of each component type, allowing the system to quickly identify media type based on the proportion and characteristics of harmonic versus percussive content without processing the entire signal uniformly.
Solution Approach 2:
Instead of processing all frequency components equally, the method applies partial action by focusing computational resources only on frequency-time regions identified as harmonic through autocorrelation analysis. This selective processing reduces overall computation time while maintaining or improving classification accuracy through targeted enhancement of diagnostically useful components.
Data Source
AI summary
Methods and apparatus for harmonic source enhancement are disclosed herein. An example apparatus includes an interface to receive a media signal. The example apparatus also includes a harmonic source enhancer to determine a magnitude spectrogram of audio corresponding to the media signal; generate a time-frequency mask based on the magnitude spectrogram; and apply the time-frequency mask to the magnitude spectrogram to enhance a harmonic source of the media signal.