wiki / narration-tts-pause-emphasis-markup / edit

Suggest an edit.

Improvements are reviewed before going live, like any wiki.

Title

One-line summary

The prompt

Prepare this script for text-to-speech synthesis by adding pause and pacing control, without rewording the content.

Script: [PASTE SCRIPT]
TTS engine or app: [ENGINE NAME, OR "UNKNOWN"]
Pause syntax it supports: [e.g. SSML BREAK TAGS, A CUSTOM PAUSE TAG, OR "PUNCTUATION ONLY"]
Does it support emphasis markup: [YES/NO]

If the engine supports explicit pause markup, use three pause lengths: about 300 milliseconds between related sentences where the default gap sounds rushed, about 600 milliseconds at topic shifts, and about 1000 milliseconds before the final line or any line that should land hard. If I said punctuation only, control pacing with structure instead: paragraph breaks for long pauses, an em dash for a mid-sentence beat, a period in place of a comma to force a stop. Either way, split any sentence over 25 words — long sentences flatten TTS prosody.

Also:
- Replace characters engines mangle or read aloud: slashes, parentheses, asterisks, ampersands, plus signs. Rewrite "24/7" as "twenty-four seven" and "&" as "and".
- Write numbers, dates, times, and units as words, in the exact spoken form.
- For any name or term likely to be mispronounced, add a phonetic respelling to a "pronunciation substitutions" list after the script. Apply substitutions inline only where the respelling reads identically to a human.
- Add emphasis markup only if I answered yes above, and mark at most one word per sentence.

Return the marked-up script ready to paste into the engine, then the pronunciation substitutions list, then a short note on anything that will still sound wrong and can only be fixed in the engine's own settings.

Write it so someone can paste it straight into Claude, ChatGPT or Gemini and fill the [BRACKETS].