How Do Instructions Shape Speech? Cross-Attention Attribution for Style-Captioned Text-to-Speech

Voice interfaces are increasingly governed by natural language instructions — a user might request speech that sounds "warm and conversational" or "brisk and authoritative." But when a text-to-speech system fails to capture that nuance, diagnosing the problem is largely guesswork. This paper borrows the DAAM attribution framework from image generation and applies it to speech diffusion models for the first time, producing heatmaps that reveal exactly which caption words drive which acoustic features. Beyond debugging, these insights could enable more controllable and expressive voice assistants, audiobook narration tools, and accessibility applications where precise vocal style matters.