Skip to content

Dataset · ccmusic-database

Vocal audio recordings: chest vs. falsetto

ccmusic-database/chest_falsetto

An annotated database of monophonic singing for training neural networks to distinguish chest voice and falsetto.

What's inside

The dataset includes 1,280 monophonic vocal audio recordings in .wav format, recorded and annotated by students of the vocal department of the China Conservatory of Music (labels 'chest' and 'falsetto'). In addition to the audio itself, the set includes mel-spectrograms and mel-frequency cepstral coefficients (MFCC), and metadata indicates Chinese and English languages.

What you can do with it

  • Training classifiers for vocal techniques and voice registers
  • Development of training tools and singing voice analysis systems
  • Testing audio models on mel-spectrograms and MFCC features
Tasks
Audio classification
Languages
Chinese, English
Size
1K–10K rows
Downloads this month
411
Likes
31

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("ccmusic-database/chest_falsetto")