Audio Fingerprint Alignment Using Offset Count Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently matching audio fingerprints to identify audio content, particularly in network-based environments, where precise alignment and classification of audio segments are required for accurate identification.
Innovation Solution
A matching machine is employed to generate and align audio fingerprints, utilizing a count vector to determine the best-matching subset of reference sub-fingerprints by evaluating total matches and additional features, such as continuity, noisiness, and symmetry, to classify matches with a predetermined threshold score.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If audio fingerprints are matched using traditional methods, then audio content identification can be performed, but the matching precision and efficiency are insufficient for accurate identification in network-based environments
Solution Approach 1:
The patent segments audio content into multiple audio portions and generates separate fingerprints for each segment. This segmentation allows for more precise local matching while maintaining overall identification efficiency, as the system can process and compare individual segments independently rather than handling entire audio files as single units.
Solution Approach 2:
The patent introduces a count vector dimension that stores total counts of matches between query sub-fingerprints and reference sub-fingerprints at different offsets. This additional dimensional information enables the system to evaluate matches more precisely by considering both the quality and quantity of matches across multiple dimensions, thereby improving matching precision without proportionally increasing processing time.
2Reliability
If the system evaluates multiple features (continuity, noisiness, symmetry) for fingerprint matching, then classification accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-computing and storing the count vector that contains total counts of matches at different offsets. This pre-computed information is then reused during the classification phase, allowing the system to evaluate multiple features (continuity, noisiness, symmetry) without performing redundant computations, thereby improving classification accuracy while managing computational complexity.
Solution Approach 2:
The count vector structure serves multiple purposes within the system - it is used both for determining total matches and for evaluating additional features like continuity and symmetry. This self-service approach allows the same data structure to support multiple classification criteria, reducing the need for separate computational processes for each feature evaluation.
3Measurement precision
If the system uses a count vector to store total counts of matches at different offsets, then the best-matching subset can be identified more accurately, but the data storage and processing requirements increase
Solution Approach 1:
The count vector is constructed by segmenting the matching process into discrete offset positions, where each position stores the total count of matches at that specific offset. This segmentation allows the system to store information in a compact, organized manner that can be efficiently processed while maintaining high matching accuracy through precise offset-based alignment of audio segments.
Data Source
AI summary
Methods, apparatus, systems and articles of manufacture are disclosed to select reference sub-fingerprints for comparison to query sub-fingerprints based on a determination that a query sub-fingerprint is a match with a reference sub-fingerprint, generate a count vector that stores total counts of matches between the query sub-fingerprints and different subsets of the reference sub-fingerprints, each of the different subsets being aligned to the query sub-fingerprints at a different offset from a reference point, each of the different offsets being mapped by the count vector to a different total count, calculate a maximum count among the total counts, a median of the total counts, and a difference between the maximum count and the median of the total counts, and classify the reference sub-fingerprints as a match with the query sub-fingerprints based on the difference between the maximum count in the count vector and the median.