An investigation into several pitch detection algorithms for singing phrases analysis

Abstract

This article provides an investigation into four different pitch detection algorithms: pYin, Praat, Phase Lock Loops (PLL)-based, and an Extended Complex Kalman Filter approach. The first two algorithms compute the pitch on a frame-by-frame basis while the latter two work on a sample-bysample basis. Only in recent years has there been a noticeable increase in the number of papers applying pitch detection techniques to sung phrases. This investigation is done on a dataset contained 76 files of singers. To create a ground truth from the data an alternative approach using the Spear analysis program is applied. The algorithms are compared using a new Singing Data Analyser tool. It was observed that the pYin and Praat are the most reliable algorithms while the PLL and Kalman filter algorithms are very dependent on the userselected parameters.

Smart-Median: A New Real-Time Algorithm for Smoothing Singing Pitch Contours

“…In addition, to compare the accuracy of the algorithms in conditions where the pitch contours included no or only a few errors, an offline pitch-detector algorithm provided in the Praat tool [28] based on the Boersma algorithm [29] was used. According to Faghih and Timoney's studies [4,5], the Praat and Pyin accuracies tend to be similar.…”

“…In other words, in this study, the best smoothing algorithm was considered the one that produced contours most similar to the ground truth. According to studies by Faghih and Timoney [4,5], a reliable offline pitch detector algorithm called PYin [24] was used. The pitch contours estimated by PYin were saved in several CSV files with two columns, time in seconds and F0.…”

“…However, it is common for pitch detector algorithms to make some incorrect estimations, resulting in a pitch contour that is not smooth and includes errors. These errors are often due to doubling or halving estimates of the true pitch value, and are therefore impulsive in appearance rather than random [1][2][3][4][5]. Furthermore, incorrect pitch estimation often happens in real-time pitch detection, especially when the sound source is a human voice [5].…”