Annotated-VocalSet: A Singing Voice Dataset

There are insufficient datasets of singing files that adequately annotated. One the available includes a variety vocal techniques (n = 17) and several singers (m 20) with WAV (p 3560) is VocalSet dataset. However, although categories, including techniques, singers, tempo, loudness, in dataset, they not Therefore, this study aims to annotate make it more powerful dataset for researchers. The annotations generated audio include fundamental frequency contour, note onset, offset, transition between notes, F0, duration, Midi pitch, lyrics. This paper describes explains our approaches creating testing annotations. Moreover, four different methods define onset/offset compared.