This dataset has been used for the paper:

Chitralekha Gupta, Haizhou Li, and Ye Wang, "Automatic Pronunciation Evaluation of Singing", Interspeech 2018.

Please cite this paper if you happen to use this dataset for your research. Thanks!

Last updated: 13th June 2018

Note: This dataset has been made available only for research purposes.

===========================================================================

• 24 audio files extracted from the DAMP Smule Dataset in the folder "wavfiles".

• The Metadata and MTurk Annotation.csv file contains details of the singers such as gender and zone (obtained from the DAMP dataset metadata). Also it contains the average pronunciation quality ratings out of 5, obtained from MTurk users (please refer to the paper for details of this experiment).

• Word-level annotations folder contains word-level markings of good or bad pronunciation by listeners. When a word is marked 1, means it is incorrectly pronounced, and blank means good pronunciation.

===========================================================================