Skip to content

How to stop saying 'um' on camera

Why fillers creep in when you talk to a lens, and how to swap them for pauses, from the way you write the script to how you rehearse, record and edit.

By Ebenezer Don6 min read

Everyone says "um". In conversation nobody notices, because the other person is busy listening. On camera it is different: there is no one nodding along, the recording plays every filler back at the same volume as your best line, and a sentence with three of them starts to sound unsure even when you are not.

The good news is that fillers are not a personality trait. They are a habit with causes you can find, and most of the fix happens before you press record.

What an "um" is doing

Fillers have a job. In a well-known paper on using "uh" and "um" in spontaneous speaking, the psychologists Herbert Clark and Jean Fox Tree argued that they work like words: speakers use them to signal that a pause is coming while they plan what to say next, with "um" usually announcing a longer one.

That is useful in a conversation, where it tells the other person not to jump in. To a lens, it is just a sound in the gap. So the goal is not to stop pausing. It is to pause in silence.

Why you say it more on camera

Most people use more fillers in front of a camera than at a dinner table. The usual reasons:

  • You are searching for the next line. Trying to remember a script, or working out the next point live, leaves gaps your mouth fills.
  • Silence feels longer than it is. A beat of quiet feels huge when you are the only one talking. On playback it usually sounds deliberate.
  • The sentences are not yours. Lines written to be read, with long clauses and formal words, are hard to say, and you stall halfway through.
  • You start before you are ready. The first "um" of many takes is the sound of deciding to begin.

Each one has a fix, and they work best together.

Find out where yours come from

Before changing anything, record two minutes of yourself explaining something you know well, then listen back with the script or a notepad in front of you. Mark every filler. Then look at where they sit:

  • At the start of sentences: you are planning on the fly. A clearer outline helps.
  • At transitions between points: the script does not tell you how to get from one idea to the next.
  • Before names, numbers and technical terms: you are not sure how to say them.
  • Spread evenly everywhere: it is a speaking habit, and practice is the fix.

The pattern tells you which of the next steps matters most for you.

Replace the um with a closed mouth

The most effective drill is simple. When you feel a filler coming, close your mouth and breathe in through your nose. You cannot say "um" with your mouth shut, and the breath gives your brain the moment it was asking for.

Practise it away from the camera first:

  1. Pick a topic and talk about it out loud for one minute.
  2. Every time you would say "um", stop and close your lips instead. Tap the table if it helps you notice.
  3. Do it again tomorrow with a new topic.

It feels slow at first. Listen back and you will hear something different: short pauses that sound like someone thinking, not someone lost.

Fix the script where you stall

A lot of fillers are writing problems. Read the script out loud and rewrite anything that trips you:

  • Write the transitions out. "So that's the setup. Here's where it goes wrong." Without a bridge, you improvise one, and improvised bridges are where fillers cluster.
  • Shorten long sentences. One idea each. If you need a breath in the middle, it is two sentences.
  • Spell out numbers and names the way you say them. "About two and a half thousand", not "2,487". Write a pronunciation hint next to any name you are unsure of.
  • Use your own words. If you would never say a phrase to a friend, you will stumble over it on camera.

The guide on writing a YouTube script covers writing for the ear in more detail.

Know the next line without memorising it

Memorising a script trades one filler trigger for another: the moment you lose a word, you search, and the search makes a sound. Reading from a teleprompter, or from short bullet prompts near the lens, means the next phrase is always in front of you.

How you read matters too. A fixed scroll speed pushes you to keep up, and rushing leaves no room to think, so the gaps fill with sound. A prompter you control, or one that follows your voice, lets you pause without falling behind. The guide on reading a teleprompter without looking like you are reading covers setup and phrasing.

Slow down slightly

Fillers often come from racing. When you speak faster than you can plan, your mouth gets ahead of your thoughts and has to wait for them. Slowing down a little, and taking a breath at every full stop, gives the next sentence time to arrive. You can check your natural pace with the free speaking speed test.

Rehearse out loud, then listen

One full spoken run-through before recording is the biggest single improvement for most people. It gives your mouth the shape of the words, so the take sounds familiar rather than new. Record the rehearsal, then listen for:

  • Lines where you restarted. They usually need rewriting.
  • Places where fillers bunch together. That section is often the one you understand least, so tighten the idea, not just the delivery.
  • Words you changed without meaning to. Your version is often better; update the script.

CopyMe, the AI studio for talking videos, has a rehearsal coach that does this listening for you. Choose Rehearse and read the script on the prompter; the run is recorded from the microphone and saved on your device. Afterwards the coach goes through the run line by line and writes up to four notes on pace, restarts, skipped lines and fillers, each with a fix.

Start and end each take clean

Two small habits make a big difference in the edit:

  • Start with a beat of silence. Look at the lens, breathe, then speak. That removes the "um, so" that opens many takes.
  • Restart the whole sentence after a bad stumble. Pause, then go back to the start of the sentence, not the middle of it. A clean pause before a full restart gives you an easy place to cut.

Cut what is left in the edit

Even with all of this, some fillers will make it into the take. That is fine. In most editors you can find them in the waveform: a short blob of sound before a sentence, often after a gap. Cut the filler and the pause around it, leaving a little air so the edit does not sound clipped.

Leave the odd one in. A video with every breath and filler removed can sound robotic, and viewers care far more about whether you are clear than whether you are perfect.

Do not chase zero

Professional presenters say "um" too. The target is a video where fillers never distract from what you are saying, not one where they never happen. If you go from a dozen a minute to one or two, viewers will hear someone confident.

The short version

  • Learn where your fillers come from: starts, transitions, names or everywhere.
  • Close your mouth and breathe instead of filling the gap.
  • Rewrite the lines you stall on, especially transitions.
  • Read from a prompter or prompts instead of memorising.
  • Slow down a little and breathe at full stops.
  • Rehearse once out loud and listen back.
  • Start each take with silence, restart whole sentences, and cut the rest.

CopyMe, the AI studio for talking videos

  • On camera

    How to be confident on camera when the red light makes you freeze

  • On camera

    How to look natural on camera when you are not a natural