What "speaking time" actually measures
Speaking time is the spoken-word duration of a piece of text. Take the word count, divide by an assumed speaking pace in words per minute, and report the result. Our counter uses 130 words per minute (wpm), the median pace for clear, audience-friendly delivery — the rate of a TED talk, a measured presentation, or a polished podcast monologue.
The widget formats output as "Xs" for under-a-minute, "X min" up to an hour, "Xh Ym" beyond an hour. A 130-word paragraph reads aloud in about 60 seconds. A 1,000-word speech runs roughly 7 minutes 40 seconds. A 5,000-word keynote runs about 38 minutes. These are the orders of magnitude for planning.
Speaking pace is a deliberate choice that makes a measurable difference in audience comprehension. Faster than 160 wpm, listeners start losing detail; slower than 100 wpm, they get bored or distracted. The 130 wpm midpoint trades clarity for engagement and is the rate most public-speaking coaches recommend for content that the audience needs to remember.
Why 130 words per minute and not faster
Different speaking contexts have different ideal paces:
- Conversational speech (one-on-one, informal): 150-180 wpm. Native English conversation is brisk; you cover ground fast because the listener is right there and can ask for clarification.
- Presentation / TED talk pace: 130 wpm (our default). Slower than conversation, faster than narration. The audience is large, the content is meant to be remembered, and you need pause time for slides or beats.
- Audiobook narration: 150-160 wpm for fiction, 130-150 for non-fiction, 110-130 for technical reads. ACX (Amazon's audiobook platform) flags reads outside 9,300 words/hour as too fast or too slow.
- News broadcasting: 150-170 wpm. News anchors are trained to deliver dense information clearly without dragging.
- Auctioneer speech: 250+ wpm. The opposite end of the scale — pure auditory shorthand for in-the-room participants.
Our 130 wpm sits at the rate where spoken English is most easily comprehended by an audience. Faster paces require more skill from the speaker and more attention from the listener; slower paces feel laboured and lose engagement.
What speaking-time estimates are useful for
Three categories of question.
Talk planning. "I have a 20-minute conference slot. How long should the script be?" 20 × 130 = 2,600 words. That includes the introduction, the main content, the conclusion. Pad your draft to 2,400-2,500 words to leave 30-60 seconds of buffer for unscripted comments, applause, or audience reaction.
Podcast scripting. A 30-minute podcast episode of pure monologue (no interview, no music) is 30 × 130 = 3,900 words of script. With interview segments or music breaks, plan less. Most polished podcast scripts come in 20-30% shorter than the time slot to allow for natural pauses, ad reads, and pacing.
Video voice-over duration. A 90-second product video runs 195 words of voice-over. A 5-minute YouTube explainer runs 650 words. These are tight constraints — most scripts overrun on the first draft and need to be cut to fit time.
Audiobook estimation. A 60,000-word novel runs 60,000 / 130 = 8 hours of audio at presentation pace. Audiobook narration is usually faster (150 wpm), so the actual duration is closer to 6.5 hours. For ACX's 9,300 words/hour benchmark, divide by that to estimate billing time.
What speaking-time estimates are NOT useful for
Several speaking contexts diverge meaningfully from the 130 wpm baseline.
Q&A and unscripted talks. When a speaker fields audience questions, restates points, or improvises, the actual speaking pace is irrelevant — the limiting factor is question density and the speaker's response time. For panels, interviews, fireside chats, our estimate undercounts time because there's substantial back-and-forth not captured in the script.
Multi-speaker scripts. A two-person dialogue includes turn-taking pauses (200-500 ms per handoff). A 1,000-word two-person script runs slightly longer than the 7:40 single-speaker estimate.
Heavy-pause material. Some speeches and meditative talks intentionally use long pauses (3-10 seconds) for emphasis. The word count under-represents duration significantly.
Languages with different syllabic density. English is fast and compact; Japanese has shorter syllables and runs faster (around 7.84 syllables/second vs English's 6.19). Spanish is even faster syllabically. Our 130 wpm estimate is calibrated for English; for other languages, the duration may be 10-20% off.
Reading vs speaking. A speaker reading from a script speaks slower than someone delivering memorised content (delivery is more natural with memorisation). Read-aloud-from-paper pace can drop to 110 wpm.
How the calculation works under the hood
The widget runs the same word-counting function as the reading-time tool, then divides by 130 instead of 200:
- Trim whitespace from input.
- Split on
/\s+/to count words. - Divide by 130 — that's the speaking time in minutes.
- Format: under 1 minute → "Xs"; under 60 minutes → "X min"; 60+ minutes → "Xh Ym".
The split-on-whitespace word-count is the universal definition for English-script text. Hyphenated words count as one. Numbers count as words. Markdown markup, if pasted, counts as part of the words it surrounds.
Common gotchas
Filler words and ums. A polished speech has very few "um", "uh", "like", "you know" — and a draft script written for delivery typically also omits them. But spoken delivery from a memorised script will introduce some fillers anyway. Real spoken duration is usually 5-10% longer than the scripted-word estimate suggests.
Slides change pace. A presentation with frequent slide transitions has built-in pauses (1-3 seconds per slide change). For a 20-minute talk with 30 slides, expect 30-60 seconds of slide-transition time on top of the speech itself. Plan for it.
Q&A is unscripted. If your 60-minute keynote includes a 15-minute Q&A, the script length should be only 45 × 130 = 5,850 words. Don't pad to 60 minutes of script and expect Q&A to fit anyway.
Music or sound effects. A podcast script with music breaks needs to subtract the music duration before applying the 130 wpm calculation. A 3-minute podcast intro music + 25 minutes of speech = 25 × 130 = 3,250 spoken words.
Reading-aloud tests overestimate by 10-20%. When testing your script, you'll naturally read faster the second time through. The first read is the realistic delivery time; subsequent reads under-represent it.
When a different tool fits better
For audiobook duration estimation specifically, use ACX's "9,300 words per hour" rule. Take your manuscript word count, divide by 9,300 to get hours of finished audio. Our counter uses presentation pace (130 wpm = 7,800 words/hour), which under-estimates audiobook duration.
For video script timing where every second matters, use a video-editing-aware tool like Descript that combines text and timeline. Speaking-time estimates are starting points; video editing tightens them.
For voiceover quoting (paid voice talent typically charges per finished minute), use the actual measured pace of the talent. 130 wpm is an estimate; experienced voice actors run anywhere from 150 to 200 wpm depending on style.
For language-specific calculations (Japanese, Spanish, German), use a wpm baseline calibrated for that language. English at 130 wpm is not a 1:1 translation to other languages.
A reference grid for speaking-time conversions
At 130 wpm:
- 130 words → 1 minute (a typical 30-second commercial reads 65 words; a 1-minute slot reads 130)
- 250 words → ~2 minutes (a short email script)
- 500 words → ~4 minutes (a wedding toast or short remarks)
- 1,000 words → ~8 minutes (a TED-style 8-minute lightning talk)
- 2,500 words → ~19 minutes (a typical conference talk slot)
- 5,000 words → ~38 minutes (a 40-minute keynote, including pauses for applause)
- 10,000 words → ~77 minutes (a long lecture or webinar)
- 60,000 words → ~7.7 hours (a typical novel, at presentation pace; faster in audiobook narration)
Drop your script into the counter for the exact number; use this grid for quick mental ballparking.
A practical workflow for speech / podcast scripting
For preparing a talk or episode script with a target duration:
- Determine the target time. A 25-minute talk slot, a 30-minute podcast episode, a 90-second video script.
- Multiply by 130 to get the target word count. Target time × 130 = upper bound on words.
- Subtract 10-15% for pauses, slide transitions, audience reaction. A 25-minute talk targets 25 × 130 = 3,250 words; subtract 10% to get 2,925 words as a working ceiling.
- Write the first draft to ceiling. Don't fight to fit on first pass.
- Cut the draft to ~80% of the ceiling. Cutting always helps; you'll find the script tighter and more memorable.
- Read the draft aloud at presentation pace. Time it with a stopwatch. Compare to the target.
- Trim what doesn't earn its place. The first cut is always too gentle; the second cut is what matters.
The counter's role in this workflow is simple: tell you the target word count and let you watch it as you cut.
A final practical note: the 130 wpm baseline assumes a polished delivery. First-draft material spoken cold runs noticeably faster (closer to 150 wpm) because untrained readers rush. If you're recording without rehearsal, expect to come in 10-15% under the estimated time; if you're delivering after multiple rehearsals with deliberate pacing, expect to be very close to the estimate. Build the buffer in once, calibrate after the first real recording, and the estimate becomes reliable for every subsequent script of similar style.