Hi-Fi Multi-Speaker English TTS Dataset.

This paper introduces a new multi-speaker English dataset for training text-to-speech models. The dataset is based on public audiobooks from LibriVox and texts from Project Gutenberg. The new dataset contains about 300 hours of speech from 11speakers with at least 16 hours per speaker sampled at 44.1 kHz. To select speech samples with high quality, we considered audio recordings with a signal bandwidth of at least 13 kHz and a signal-to-noise ratio (SNR) of at least 32 dB. The dataset will be released publicly at OpenSLR.