What is this?

This is a dataset extracted from brazilian cartoon called Fudêncio (by MTV), which is somehow similar to South Park.

This dataset has three features

rttm string ( to identify speakers ), episode_name ( for reference ) and the audio voicelines Acapella made using Demucs (no_voice files are not included)

The separation of Demucs is not perfect, but can help people to train RVC voices for different characters.

Plan to future upgrades

I plan to use this later to split voicelines by characters names into another dataset.