Synthesis methods using identical transcripts for each speaker.
Models and datasets: https://huggingface.co/resproj007