Dataset for Multilingual and Accent-Aware Transformation of Read Speech to Conversational Speech

This dataset supports the research article \u201cMultilingual and Accent-Aware Transformation of Read Speech to Conversational Speech.\u201dIt contains 300 English speech recordings including read, conversational, and emotionally expressive (happy) speech samples. Each audio file is paired with its text transcript and the output of a part-of-speech (POS) tagger generated using the NLTK toolkit. The dataset provides a foundation for studying prosody, filler word placement, and emotional tone variation in conversational speech. It is intended for research on expressive text-to-speech synthesis, prosody modeling, and emotion-aware human-computer interaction.