---
title: "Comment by Ahmet Özel on \"A 30-second TTS demo can't tell you if a voice can narrate a book\" | Hashnode"
description: "Block-level regeneration controls long-run drift, but it moves part of the problem to the seams between blocks. I would carry a compact prosody state - pace, pi"
url: "https://hashnode.com/posts/a-30-second-tts-demo-can-t-tell-you-if-a-voice-can-narrate-a-book/6a64d3b2846eda1e01c6a233/comment/6a6539a95dee6e0fdbefc8c3"
---

[Discussion](https://hashnode.com/posts/a-30-second-tts-demo-can-t-tell-you-if-a-voice-can-narrate-a-book/6a64d3b2846eda1e01c6a233)

Comment [A 30-second TTS demo can't tell you if a voice can narrate a book](https://hashnode.com/posts/a-30-second-tts-demo-can-t-tell-you-if-a-voice-can-narrate-a-book/6a64d3b2846eda1e01c6a233)

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Jul 25](https://hashnode.com/posts/a-30-second-tts-demo-can-t-tell-you-if-a-voice-can-narrate-a-book/6a64d3b2846eda1e01c6a233/comment/6a6539a95dee6e0fdbefc8c3)

Block-level regeneration controls long-run drift, but it moves part of the problem to the seams between blocks. I would carry a compact prosody state - pace, pitch range, energy, and pause profile - into the next segment, then use short acoustic overlaps or crossfades where the engine allows it. Evaluating boundary discontinuity alongside the thirty-minute drift test would catch a pipeline that is locally stable but audibly stitched together.
