Beijing Opera Acapella Phonemes Dataset

The dataset consists of excerpts from 15 arias of two female singers, chosen from a CompMusic corpus of Jingju arias. For a given aria were present two versions: a recording with voice plus accompaniment and an
accompaniment-only one. Thus a cappella singing was generated by subtracting the instrumental accompaniment
from the polyphonic version.
Please refer to the following publication if you use this data in your research:
Dzhambazov, G., Yang Y., Caro Repetto R., & Serra X. (2016).  Automatic Alignment of Long Syllables In A cappella Beijing Opera. 6th International Workshop on Folk Music Analysis (FMA 2016). 88-91.
The dataset can be downloaded from:

and the particular snapshot used in the above paper here: