A spoken “um” or “uh” is audio, not silence. AudioCut detects quiet sections by level and duration; it does not transcribe or identify filler words. Use manual trimming when you can choose the specific passage yourself.
Why the detection rule matters
Silence detection looks for runs below a level threshold. A spoken filler may be as loud as the words beside it, while a useful word ending may be very quiet. Raising the threshold to catch the filler can therefore remove material you wanted to keep.
Some pauses and fillers serve a purpose. In an interview, a hesitation can communicate uncertainty. Decide whether the edit is meant to tighten a prepared narration or preserve the character of a conversation.
Remove one filler by hand
Open Audio Trimmer and listen to the passage. Select the filler only after you have identified its boundaries. Play the audio before and after the selection, delete it and listen across the join in Processed mode.
If the words now collide or the cut sounds abrupt, restore the selection and retain more breathing room. Sometimes keeping a short filler sounds more natural than removing it. Save a new export only after reviewing the surrounding sentence.
Use silence detection for the waits between takes
Long quiet waits between narration takes are a different task. Silence Remover can mark those gaps so you can review them more quickly. It will not decide which take is correct, remove repetitions by meaning or identify the words spoken.
For a long recording, use an edit decision log to note repeated phrases and their source timecodes. Keep the original so you can revisit a decision later instead of working only from a shortened export.
| Material | Useful method | Review requirement |
|---|---|---|
| Long quiet pause | Silence detection | Keep meaningful pauses and soft speech |
| Spoken filler or repeated phrase | Manual selection | Listen to the complete sentence |
| Many words to classify automatically | A suitable transcription-based workflow | AudioCut does not provide word recognition |
Questions about this task
Can I remove all filler words with one threshold?
No. Loudness is not a reliable way to identify particular words or sounds.
Will silence removal change the words that remain?
It joins retained audio in order, but an overly broad cut can remove part of a word. Review the edges before exporting.