#style captioning

LibriTTS-P offers an extensive set of annotations for text-to-speech and style captioning, encompassing both human-evaluated speaker traits and machine-generated speaking style prompts. This dataset provides a broader range of annotations for LibriTTS-R speakers, enhancing the naturalness and accuracy of models in TTS and style captioning domains. It also includes comprehensive metadata and style prompt options, facilitating the detailed study of spoken language attributes.