Music Item Relation Mapping Across Audio and Symbolic Formats
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to determine relations between music items of different formats and lengths, such as audio and symbolic files, within a database, limiting the ability to identify related tracks and submixes effectively.
Innovation Solution
Utilizing conditional similarity networks (CSNs) to map music items into multiple subspaces, each modeling a specific characteristic, enabling cross-domain relation of music files by training neural networks to minimize distance between related items in these subspaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing similarity methods are used to determine relations between music items, then the method is simple to implement, but it cannot handle music items of different formats (audio and symbolic files) and lengths effectively
Solution Approach 1:
The patent segments music items into fixed-length windows or segments, allowing audio and symbolic files of any length to be processed uniformly. Each segment is independently encoded and compared, enabling the system to handle varying lengths without requiring complex alignment algorithms.
Solution Approach 2:
The patent introduces an intermediary embedding space where both audio and symbolic music items are transformed into comparable representations. This intermediate representation layer acts as a mediator that enables cross-format comparison without requiring direct format-specific matching logic.
2Quantity of substance
If the track database grows to include more songs and tracks, then the database becomes more comprehensive, but users cannot listen to all tracks to determine which are related
Solution Approach 1:
The patent pre-computes embeddings for all music items in the database and stores them in an accessible format. This preliminary action allows the system to rapidly retrieve and compare embeddings without requiring users to listen to tracks, significantly reducing the time needed to identify related music items even as the database grows.
3Adaptability or versatility
If conditional similarity networks are applied only to full mixes and audio files, then the implementation is straightforward, but it cannot relate individual tracks, submixes, or cross-domain files (audio to symbolic)
Solution Approach 1:
The patent implements a universal embedding framework that can process audio files, symbolic files, individual tracks, submixes, and full mixes through the same encoding pipeline. This multi-functional approach enables cross-domain relation determination while maintaining a relatively simple implementation by reusing the same core components for different input types.
Data Source
AI summary
A method of determining relations between music items, wherein a music item is a submix of a musical composition comprising one or more music tracks, the method comprising determining a first input representation for at least part of a first music item, mapping the first input representation onto to one or more subspaces derived from a vector space using a first model, wherein each subspace models a characteristic of the music items, determining a second input representation for at least part of a second music item, mapping the second input representation onto the one or more subspaces using a second model, and determining a distance between the mappings of the first and second input representations in each subspace, wherein the distance represents the degree of relation between the first and second input representations with respect to the characteristic modelled by the subspace.