Text to speech time: how long will my text take to read aloud?
Before you voice a script you usually need to know how long it will be: a 30-second ad slot, a 60-second Short, a 10-minute podcast, a lesson that must fit a class. Word counts alone do not tell you, because every language and every voice has its own pace.
We measured the SuperClonica engine on the same story in seven languages at its default speed, so you can estimate the length before voicing anything, and then get the exact number in seconds with one line of Python.
Measured speaking rates
Measured on a calm narrative paragraph with a male and a female stock voice at the default speed; the two voices differed by less than three percent.
- English: about 190 words a minute, about 1,000 characters.
- Spanish, German and French: about 165 words a minute.
- Portuguese: about 150 words a minute.
- Russian: about 130 words a minute, about 840 characters.
- Japanese: about 320 characters a minute.
Hear the pace free online
On the page of any voice, type a line of 30 to 100 characters and press Voice it: you hear the pace of that voice in your language in a few seconds, with no account.
Quick estimates
Leave a few seconds of air: a script that fills the slot to the last second sounds rushed once music and pauses are added.
- A 30-second ad: about 95 words in English.
- A 60-second Short or Reel: about 190 words.
- A 10-minute podcast or video: about 1,900 words.
- 1,000 words: about 5 minutes 15 seconds.
The exact length of your script
Speed changes the length almost proportionally: the default is 1.05, and --speed 0.85 makes the same text about a quarter longer. For the exact length, install the free engine with python -m pip install superclonica-cpu, save the script as script.txt and run: from superclonica import Superclonica; tts = Superclonica(); audio, rate = tts.synthesize(open('script.txt', encoding='utf-8').read(), 'M1', 'en'); print(round(len(audio) / rate, 1), 'seconds'). The Docker server returns the same number in the X-Duration-Seconds header.
Why your text may run longer
Digits are spelled out, so 1,250 or 2027 becomes several spoken words. Abbreviations may be read in full. Many full stops and separate lines with silence between them add time, while long sentences without commas read faster. Measure the real script once instead of relying on the average.