LibriTTS-R [1] is a sound quality improved version of the LibriTTS corpus (http://www.openslr.org/60/) which is a multi-speaker English corpus of approximately 585 hours of read English speech at 24kHz sampling rate, published in 2019.

Overview

This is the LibriTTS-R dataset, adapted for the datasets library.

	Usage




	


	Splits

There are 7 splits (dots replace dashes from the original dataset, to comply with… See the full description on the dataset page: https://huggingface.co/datasets/mythicinfinity/libritts_r.