Neural Network Music Analysis for Automated Tagging

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current music search methods, such as text query and class query, require expert intervention and are inefficient for handling new music releases, as they rely on manual classification and bibliographic information input, limiting scalability and accuracy in finding similar music.

Innovation Solution

A music analysis method using an artificial neural network that converts sound source and music property data into embedding vectors, allowing for cross-comparison and analysis in a common space to accurately extract and reflect music characteristics, enabling more precise similar music searches and services.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual classification and bibliographic information input are used for music search, then music can be organized and retrieved, but expert intervention is required and the system cannot efficiently handle new music releases

Engineering Contradiction:

Improvemusic classification efficiencyVSAvoidautomatic music tagging capability

Core Design Contradiction:

ProductivityVSExtent of automation

Solution Approach 1:

The system enables music to be automatically classified and tagged without expert intervention by using neural networks to analyze audio features and generate classifications autonomously. The neural network processes new music releases automatically, extracting features and generating tags without requiring manual input from experts.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of expert classification with an automated neural network system. The neural network analyzes audio signals, extracts features, and performs classification automatically, substituting the need for human experts to manually input bibliographic information and classify music.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Stability of the object's composition

If taxonomy-based classification is used, then music can be organized by genre or atmosphere, but the system lacks adaptability when new items are added

Engineering Contradiction:

Improveclassification structure stabilityVSAvoidclassification system adaptability

Core Design Contradiction:

Stability of the object's compositionVSAdaptability or versatility

Solution Approach 1:

The classification system transitions from a static taxonomy to a dynamic neural network-based system that can adapt to new music. The neural network continuously learns from new data and adjusts its classifications, allowing the system to accommodate new genres, styles, and music types without requiring manual updates to the classification structure.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the parameters of classification by using neural network features and embeddings instead of fixed taxonomy categories. The neural network can identify and classify music based on various parameters such as tempo, mood, instrumentation, and style, providing flexible adaptation to new music types while maintaining organizational stability.

Inventive Principle:

Principle #35Parameter changes

3Ease of operation

If text query or class query methods are used, then music search can be performed using existing information, but the methods require expert intervention and do not accurately capture music characteristics

Engineering Contradiction:

Improvesearch operation simplicityVSAvoidmusic characteristic analysis accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The neural network acts as an intermediary between the user's simple search query and the complex music characteristic analysis. The system automatically extracts relevant features and performs accurate music analysis in the background, returning precise results without requiring the user to understand or specify complex music parameters.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual expert analysis with automated neural network analysis. The neural network accurately extracts music characteristics such as tempo, mood, instrumentation, and style through automated audio feature extraction and analysis, eliminating the need for expert intervention while maintaining high precision in music characteristic identification.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20230351152A1Music analysis method and apparatus for cross-comparing music properties using artificial neural network

Publication Date: 2023.11.02 NEUTUNE CO LTD

AI summary

A music analysis device that cross-compares music properties using an artificial neural network comprises a processor including an artificial neural network module and a memory module storing instructions executable by the processor. The artificial neural network module includes a pre-processing module that outputs stem data that is specific attribute data constituting the audio data according to a preset standard for the input audio data, a first artificial neural network that takes first stem data as first input information and outputs a first embedding vector that is an embedding vector for the first stem data as first output information, a second artificial neural network that takes second stem data as second input information and outputs a second embedding vector that is an embedding vector for the second stem data as second output information and a dense layer that uses information the first output information and the second output information are used as input information and output a first tagging information and a second tagging information as output information which are music tagging information for the first output information and the second output information.