Audio Analysis Device Multi-Genre Feature Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing music genre classification technologies struggle to accurately select musical pieces with similar features across multiple music categories, as actual musical pieces often contain features common to several genres, leading to variations in musical features among pieces classified under the same genre.
Innovation Solution
An audio analysis method and device that acquire an audio signal, calculate feature values representing the certainty of a musical piece belonging to various music categories, and select candidate pieces with similar feature values, utilizing non-negative matrix factorization and correlation analysis to determine musical similarity across multiple genres.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If music genre classification is performed using conventional single-category methods, then the classification process is simple, but the accuracy of selecting musically similar pieces deteriorates because pieces in the same genre may have significantly different musical features
Solution Approach 1:
The patent segments the music classification problem into multiple independent music categories (e.g., genre, mood, tempo, instrumentation) rather than using a single monolithic classification. Each category is analyzed separately to extract specific feature values, allowing for nuanced comparison across different dimensional aspects of music, thereby improving the accuracy of finding musically similar pieces.
Solution Approach 2:
The patent transitions from single-dimensional genre classification to multi-dimensional classification by introducing multiple music categories as additional dimensions. This dimensional expansion allows the system to capture the complexity of musical similarity across different aspects (genre, mood, tempo, etc.), resolving the contradiction between simple classification and accurate similarity detection.
2Reliability
If multiple music categories are analyzed to improve selection accuracy, then the relevance of suggested pieces improves, but the computational complexity increases
Solution Approach 1:
The analysis process is segmented into independent category-specific analysis modules, where each music category (genre, mood, tempo, etc.) is processed separately. This segmentation allows for targeted feature extraction and comparison for each category, improving reliability while managing computational complexity through modular design.
Solution Approach 2:
The patent changes the parameters being analyzed by extracting specific feature values for each music category rather than using holistic analysis. This parameter-specific approach allows for more precise comparisons across multiple dimensions, enhancing the relevance of suggestions while maintaining computational efficiency through focused analysis.
Data Source
AI summary
An audio analysis device comprises an electronic controller including at least one processor. The electronic controller is configured to execute a plurality of modules including a signal acquisition module configured to acquire an audio signal representing performance sounds of a musical piece, a signal analysis module configured to calculate, for each of a plurality of music categories, a feature value that includes a degree of certainty that the musical piece belongs to the music category, by analyzing the audio signal, and a music selection module configured to select one or more candidate musical pieces whose feature value is similar to the feature value calculated for the musical piece from among a plurality of candidate musical pieces.