We provide meticulously crafted and validated datasets specifically designed to enhance Text-to-Speech (TTS) model performance. Our datasets incorporate human annotation for emotional and prosodic features, ensuring superior accuracy and natural-sounding speech.