Development of a Guitar Practice Support System Based on Chord Estimation

We propose a real-time guitar practice support system that estimates chords played by users and provides feedback to improve performance accuracy. The system captures an electric guitar signal of each played chord via an audio interface, and calculates 12-dimensional chroma vectors from the time–frequency representation. These vectors are then compared with predefined chord templates to compute a matching score. If the score falls below a predefined threshold, the system identifies the weakest contributing note and suggests which string and fret to adjust. Our chord recognition algorithm achieved a 92% accuracy rate. A user study demonstrated that the system effectively improved user’s ability to correct chord errors.

Images