Direct Visualization of Bass Guitar Frequency Patterns and Their Fret Fingerings via Combined Fast and Short-Time Fourier Transforms

This study aims to develop an efficient method for transcribing bass guitar notes from polyphonic music, addressing the challenges in creating user-friendly tablature for low-frequency instruments.

Employing Fast Fourier Transform (FFT) and Short-Time Fourier Transform (STFT), the method divides music into sections to extract bass notes, identifying frequencies using FFT with Hamming windowing and isolating fundamental frequencies. STFT and image recognition determine note details, such as lengths, and Euclidean methods with minimum distance estimation are used to generate tablatures reflecting minimal finger movements.

Findings show a Word Error Rate (WER) of 3.13% and a Character Error Rate (CER) of 6.25% in simple polyphonic music, demonstrating the method's effectiveness. Challenges remain in transcribing complex compositions, indicating a need for methodological refinement.

The research proposes a simple and effective transcription method for bass guitar from polyphonic music, generating tablatures that assist amateurs in performing more easily. It opens avenues for further improvement in automatic music transcription technologies and suggests potential applications in detecting mechanical faults in low-frequency domains.