We use cookies to improve the services we offer you. By continuing to browse this site, you consent to keep them in accordance with our Privacy Policy.
×We use cookies to improve the services we offer you. By continuing to browse this site, you consent to keep them in accordance with our Privacy Policy.
×
2
6 min
Play back the first take of almost any talking-head video and you’ll hear it. “So, um, basically, um, what I want to show you today is, uh…” Ten seconds in, three ums. Maybe four.
You can get rid of most of them without a public speaking course. Below are five habits that cut them before you record, plus two ways to deal with the stragglers, by hand or with an automatic tool. A good picture starts with lighting and framing, but this one is about the audio.
Your brain is planning the next few words and hasn’t finished yet. That’s most of it.
Psychologists Herbert Clark and Jean Fox Tree studied recorded conversations and concluded that uh and um are real words with a job to do. They announce a delay. “Uh” warns of a short one and “um” of a longer one. That’s fine at dinner. It’s less fine 40 seconds into a YouTube intro.
Harder thinking brings more of them. In one study of disfluency rates in conversation, people stumbled more describing abstract shapes than photos of children.
Then there’s the camera. A real listener nods and says “mm-hm.” The lens does none of that, so a one-second pause feels like five and you fill it with sound. The French “euh” lands in the same spots, at the start of a new point and right after “so.”
You won’t stop every um, and you shouldn’t try to. Swap most of them for silence and plan enough that you need fewer pauses. The same habits help you stop using filler words in presentations and Zoom calls.
Record two minutes about something you know well. No script. Play it back and make a tick for every um, uh, like, and you know, then divide by two. That’s your Filler Rate, fillers per minute. Write it down and redo the test in two weeks.
Toastmasters clubs even have a role for this, the Ah-Counter, who notes every filler sound in a meeting and reports them at the end. A friend tapping the table at each um works too.
If you only take one habit from this article, take this one. A pause buys your brain the same second an um does, and on camera it comes across as calm. It’s easier to edit, too. Silence comes out in one clean cut, while an um is glued to the word in front of it.
The physical trick is to finish the sentence, close your mouth, and breathe in through your nose. It’s hard to say “um” while you’re inhaling with your lips shut. Feels odd for a day or so.
Memorized scripts cause ums. The moment someone loses a word, they stall and fill the gap while they hunt for it.
Three to five bullets, a few words each, avoid that. Do write your first sentence word for word, though, since the opening line has nothing before it to build on. And if a video needs exact wording (an ad, say), write a proper video script and read it off a teleprompter.
Stop after each point, check your bullets, breathe, then record the next one. When a sentence goes sideways halfway through, don’t rescue it with “um, what I mean is.” Just start it again from the top.
The worst spot is the seam between two ideas, when you’ve finished one thought and haven’t loaded the next. That’s where the ums tend to pile up for talking-head footage. Nervous speakers speed up right there, so you do the opposite: land the point, count one beat, then start.
“Like” and “you know” come mid-sentence, usually to soften something or check the viewer is still with you, so silence won’t fix them.
Shorter sentences help, because long ones need glue and “like” is the nearest glue to hand. Being specific helps more. “It takes, like, a while” turns into “It takes about ten minutes,” and there’s nowhere left for the like to go. For “you know,” slow down.
Some will get through anyway.
In Movavi Video Editor for Windows and macOS, each um takes two splits and a delete.
On a 20-minute video, hand-cutting every um can eat an evening. Pauses, though, can be removed automatically. Open Tools > Audio > Silence removal, and Remove Silence in the Auto mode. The program finds the pauses between phrases and joins the clips without gaps. If it cuts too much, set the pause length and the volume threshold yourself.
If you edit in Premiere Pro or DaVinci Resolve, for the ums specifically, a transcript-based tool such as SilentCut’s Speech Cleanup can cut them, along with stutters and false starts, from English speech in one pass.
If a jump cut looks jarring, cover it with B-roll or punch in slightly on the next clip. Your subtitles will read like proper sentences, too.
No. Strip out every hesitation and the person on screen sounds breathless, like an auctioneer. A couple of ums sound more natural than none.
In word-recognition experiments, Jean Fox Tree found that hearing an “um” neither helped nor hurt people’s ability to recognize the words that followed. A later study found that listeners recognized the next word faster after a delay of any kind, whether it was an “um,” a silent pause, or a tone. So an occasional um doesn’t cost you clarity.
A rough rule is to cut nearly all of them in tutorials, ads, and shorts, and leave a few in interviews and vlogs.
Most ums are your brain planning out loud. Count yours on a two-minute test, swap them for silence, talk from bullets, record one idea at a time, and slow down between points. Whatever still gets through, cut by hand or from the transcript, fillers before silences. Then run the two-minute test again.
