TTS reading speed calculator
How long will your script be once it is read aloud? Whether you are filling a podcast slot, a video voiceover, or an audiobook chapter, audio duration comes down to word count, speaking rate, and pauses. This calculator converts your words into an estimated duration at slow (≈120 wpm), normal (≈150 wpm), or fast (≈190 wpm) rates, with a pause adjustment.
How it works
The base estimate is simple, then pauses are layered on:
base_minutes = words / words_per_minute
pause_minutes = sentences × pause_seconds / 60
duration = base_minutes + pause_minutes
A “normal” TTS read sits around 150 words per minute, matching a typical human narrator. Bumping a TTS engine’s speed parameter to 1.2 is roughly the same as choosing the fast rate here, while 0.8 maps to slow.
Tips for hitting your target length
- Plan backwards from the slot. A 60-second ad at normal pace is about 150 words — write to that, then trim.
- Pauses add up fast. A one-second beat after every sentence in a 30-sentence script adds half a minute on its own.
- Slow down for instructions. Step-by-step content reads clearer at 120–130 wpm; rushing it hurts comprehension.
- Verify the final render. Numbers, acronyms, and SSML breaks shift real duration — confirm against an actual generation for tight broadcast windows.
Common use cases and target word counts
Different audio formats have different sweet spots for duration and speaking rate:
60-second ad spot: Roughly 130–150 words at normal pace. Most broadcast standards require tight fit to the slot, so knowing your word count target before writing prevents repeated re-edits.
Podcast intro / outro: 30–90 seconds is common. At 150 wpm a 30-second intro needs about 75 words; a 90-second version needs around 225. Hosts often read slightly faster than a polished narrator, closer to 160–170 wpm.
Explainer video voiceover: Typical duration is 90 seconds to 3 minutes. At normal pace, 3 minutes requires around 450 words. Allow extra time for pauses between segments and for visual-led sections where the narrator is silent.
Audiobook chapter: Audiobook narration averages 150–160 wpm in professional recordings. A 3,000-word chapter runs approximately 18–20 minutes. Many producers target 25–45 minute chapters, so chapters of 4,000–7,000 words are common.
E-learning module: Instructional content benefits from a slightly slower pace, around 120–130 wpm, to allow learner processing time. A 10-minute module therefore needs roughly 1,200–1,300 words of narration script.
Why TTS engines behave differently from the estimate
TTS engines introduce variability beyond words-per-minute:
- Numbers and dates are expanded and read slowly (
2026becomestwenty twenty-six, adding syllables). - Abbreviations may be spelled out letter by letter (
TTSasT T S) or read as a word depending on the engine. - SSML pause tags (“) add explicit silence that is not in the word count.
- Voice character affects pacing — some voices naturally run fast or slow within the same nominal speed setting.
For anything time-critical (broadcast, exam narration, locked-picture video), always render a full-length test audio before finalising the script.