Automatic extraction and discrimination of vocal main melody based on quadratic wave equation

Abstract

Abstract An algorithm is proposed to derive the quadratic wave equation and to generate a musical theme based on the basic frequency differentiation of singing. The algorithm can effectively reduce the false alarm speed of the melody localization and improve the overall accuracy. The Metric Distance (DIS) algorithm uses the dispersion method to segment musical notes and to detect vocal parts; The amplitude compression frequency filter (PEFAC) uses multi-frequency frequency separation technology. vocal range by calculating the melody. Basic frequency of multiple candidates. Finally, the Viterbi algorithm is used to control the predominant basic frequency trajectory of the vocal segment and is used to distinguish the basic melody of the basic frequency differentiation model. Tests on the MIR-1K data set show that all performance parameters of the improved algorithm are higher than the original algorithm, tone localization call speed (VRR), initial sound frequency accuracy (RPA), and initial color accuracy level (RCA). ) and the Total Accuracy Level (OA) improved by different signal and interference conditions, while the pseudo-localization (VFAR) decreased by about 2%, indicating that the improved algorithm could more accurately identify the accompaniment and vocals. This is because the accuracy of recognizing the basic frequency discrimination model in an improved algorithm is higher than in the original algorithm. In addition, the AET of the improved algorithm is reduced by approximately 0.12 seconds compared to the original algorithm, which ensures that the process calculation and complexity of the frequency function function of the improved algorithm is lower than that of the original algorithm. The octave error detection, the false localization signal is much lower than other algorithms, and the overall accuracy level is higher than other algorithms, which can effectively produce the basic melody of the song.