Chord Estimation Using Trained Model for Audio Signal Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for recognizing chords from audio signals struggle with accuracy when the frequency spectrum information differs significantly from pre-prepared chord patterns, leading to inaccurate estimation of chords.
Innovation Solution
A chord estimation method that involves estimating a first chord from an audio signal and inputting it into a trained model that has learned chord modification tendencies to estimate a second chord, utilizing feature amounts and temporal changes to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pattern matching method is used to recognize chords by comparing frequency spectrum information with pre-prepared chord patterns, then the recognition process is simple and fast, but the accuracy deteriorates when the frequency spectrum information differs greatly from the pre-prepared patterns
Solution Approach 1:
The system performs preliminary chord estimation to obtain a first chord result, then uses this as input for a trained model to perform secondary estimation. This two-stage preliminary action approach allows the system to first get a quick estimate and then refine it with more complex processing only when needed.
Solution Approach 2:
The trained model acts as an intermediary between the raw frequency spectrum analysis and the final chord recognition. It takes the first chord estimation as input and produces a refined second chord estimation, mediating the transformation from rough to accurate recognition.
2Measurement precision
If a trained model with chord modification tendency learning is used to estimate chords, then the accuracy is improved, but the device complexity increases
Solution Approach 1:
The chord recognition system is segmented into two distinct modules: a first chord estimation module that provides quick initial results, and a trained model module that refines the estimation. This segmentation allows each module to be optimized independently and reduces overall system complexity.
Solution Approach 2:
The system applies partial complex processing by using the trained model only for refinement of the first chord estimation, rather than applying full complex processing to all cases. This partial action approach maintains accuracy while reducing unnecessary computational overhead.
Data Source
AI summary
A chord estimation apparatus estimates a first chord from an audio signal, and estimates a second chord by inputting the estimated first chord to a trained model that has learned a chord modification tendency.