Thanks! We’re appending a clean 'Director's Notes' block directly into the text compilation payload (mapping explicit Style, Pace, and Accent strings). The speech model reads those staging notes natively to synthesize the dynamic shifts you hear in the clip!