Musical Instrument Onset Estimation Model Construction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for analyzing musical instrument performance suffer from insufficient onset estimation accuracy, making it difficult to accurately evaluate a performer's skills.
Innovation Solution
A computer-based method involving the construction of an estimation model using machine learning, which prepares training data that include feature amount data from musical instrument sounds and onset data, as well as data from different sound sources, to estimate onset points in musical performance sounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional techniques are used for analyzing musical instrument performance, then the analysis process is simple, but the onset estimation accuracy is insufficient
Solution Approach 1:
The patent applies preliminary action by preparing training data in advance that includes both musical instrument sounds and non-musical instrument sounds. This pre-prepared diverse training data enables the machine learning model to learn onset detection patterns more effectively, improving onset estimation accuracy without requiring complex real-time processing during actual performance analysis
Solution Approach 2:
The patent changes the parameter composition of training data by incorporating not only musical instrument sounds but also non-musical instrument sounds as training data. This parameter change in data diversity enables the estimation model to better distinguish true onsets from other sound events, thereby improving measurement precision while maintaining manageable model complexity
2Measurement precision
If machine learning with diverse training data is used, then onset estimation accuracy is improved, but data preparation complexity increases
Solution Approach 1:
The patent applies universality by creating a training data system that serves multiple functions: musical instrument sounds provide positive examples of onsets, while non-musical instrument sounds provide negative examples. This multi-functional training data approach improves onset estimation accuracy while the systematic methodology keeps data preparation manageable through reusable data sets
3Adaptability or versatility
If only musical instrument sounds are used for training, then data preparation is simple, but the model cannot distinguish sounds from other sources
Solution Approach 1:
The patent applies segmentation by dividing training data into distinct categories: musical instrument sounds and non-musical instrument sounds. This segmentation enables the model to learn discriminative features for different sound sources, improving adaptability while the organized categorical structure makes managing the quantity of training data more systematic and manageable
Data Source
AI summary
An estimation model construction method realized by a computer includes preparing a plurality of training data that include first training data that include first feature amount data that represent a first feature amount of a performance sound of a musical instrument and first onset data that represent a pitch at which an onset exists, and second training data that include second feature amount data that represent a second feature amount of sound generated by a sound source of a type different than the musical instrument, and second onset data that represent that an onset does not exist, and constructing, by machine learning using the plurality of training data, an estimation model that estimates, from a feature amount data that represent a feature amount of a performance sound of the musical instrument, estimated onset data that represent a pitch at which an onset exists.