Audio Fingerprinting Via Mean Normalization Across Frequency Ranges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio fingerprinting technologies often rely on the loudest parts of an audio signal, which can be dominated by noise or bass frequencies, failing to capture the entire signal, especially in noisy environments, and existing methods fail to incorporate samples from all parts of the audio spectrum, especially in higher frequency ranges, leading to ineffective solutions. Existing methods fail to effectively address these limitations by using mean normalization and weighted selection of audio signal components to generate a fingerprint that includes samples from all parts of the audio spectrum, including higher frequency ranges.
Innovation Solution
The method and apparatus generate a fingerprint from an audio signal using mean normalization and weighted selection of audio characteristics, which involves normalizing audio signal characteristics, which involves normalizing the audio signal by an audio characteristic of the surrounding audio region, and selecting points from the normalized audio signal frequency components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If audio fingerprinting relies on the loudest parts of the audio signal, then the fingerprint generation is simple and fast, but the fingerprint fails to capture the entire signal and is dominated by noise or bass frequencies
Solution Approach 1:
The patent applies parameter changes by normalizing the audio signal components based on their frequency characteristics. Instead of using raw amplitude values, the system transforms the signal into the frequency domain and normalizes each frequency component relative to the overall signal energy. This parameter transformation allows the system to capture information from all frequency ranges equally, not just the loudest components, thereby improving measurement precision while maintaining processing efficiency.
Solution Approach 2:
The patent segments the audio signal into distinct frequency components using spectral analysis. By dividing the continuous audio signal into discrete frequency bins and processing each segment independently through normalization, the system can selectively capture representative samples from all frequency ranges. This segmentation enables comprehensive signal representation without being dominated by any single frequency band, resolving the contradiction between simple processing and accurate capture.
2Device complexity
If existing methods use only prominent frequency components, then the processing complexity is low, but higher frequency ranges are not captured effectively
Solution Approach 1:
The patent implements universality by creating a normalization mechanism that universally applies to all frequency components regardless of their prominence. The normalization process treats high frequency components the same as low frequency components, ensuring that no frequency range is excluded from the fingerprint generation. This multi-functional approach allows the system to simultaneously capture information from bass frequencies, mid frequencies, and high frequencies, preventing information loss while maintaining manageable processing complexity through a unified normalization algorithm.
3Ease of manufacture
If audio fingerprinting focuses on dominant frequency components, then the fingerprint is easy to generate, but it fails in noisy environments
Solution Approach 1:
The patent uses parameter changes by transforming the audio signal from the time domain to the frequency domain and applying normalization based on spectral characteristics. This parameter transformation makes the fingerprint generation process robust to noise, as the normalization process identifies and emphasizes genuine signal patterns while suppressing noise artifacts. The system calculates the energy distribution across frequency components and normalizes accordingly, creating a reliable fingerprint that maintains accuracy even in noisy environments while remaining computationally feasible.
Data Source
AI summary
Methods, apparatus, systems, and articles of manufacture are disclosed to fingerprint audio via mean normalization. An example apparatus for audio fingerprinting includes a frequency range separator to transform an audio signal into a frequency domain, the transformed audio signal including a plurality of time-frequency bins including a first time-frequency bin, an audio characteristic determiner to determine a first characteristic of a first group of time-frequency bins of the plurality of time-frequency bins, the first group of time-frequency bins surrounding the first time-frequency bin and a signal normalizer to normalize the audio signal to thereby generate normalized energy values, the normalizing of the audio signal including normalizing the first time-frequency bin by the first characteristic. The example apparatus further includes a point selector to select one of the normalized energy values and a fingerprint generator to generate a fingerprint of the audio signal using the selected one of the normalized energy values.