WORKED EXAMPLE
Example: remove a repeated sentence
Let’s begin with the first step. Let’s begin with the first step. Open your project.
Let’s begin with the first step. Open your project.
1. Record or import the audio
Record a take in the studio, or choose Import audio to add a file from your device to the current project. Create a new project first if you want to keep the imported file separate from an existing recording.
Open Edit audio and play the recording. Make sure you are working on the right project before selecting a passage. The original takes remain available while you edit.
2. Generate an editable transcript
Sign in with Google, then choose Transcribe locally. The first use requires downloading a Whisper model; the default Tiny English model is about 40 MB. Transcription runs on your device, rather than uploading your recording to a transcription service.
Watch the transcription progress and emerging text. Wait for the final transcript before making word-based edits: provisional text and final word boundaries can differ. If a name or phrase is recognized incorrectly, listen to the audio to identify the passage.
3. Select the exact passage
Click a word to select it. Shift-click another word to extend the selection into a passage, or double-click a word to select its sentence. The selection toolbar shows the text and duration, while the waveform highlights the matching time range.
Choose Play selection before you cut. In the repeated-sentence example above, select only one occurrence, not both. The better delivery may be the first or the second one; the transcript alone cannot tell you which sounds best.
4. Delete the words and their audio
Choose Delete words & audio. The selected interval is removed from the current audio edit, and the remaining transcript timing shifts with it. Use Undo if you remove too much.
Word timestamps are approximate. Listen across the join, not just to the deleted passage. If a consonant is clipped or a breath sounds abrupt, undo the cut and refine the start and end using the waveform zoom and time fields. Keeping a little natural silence can make the join sound more comfortable.
5. Replace a sentence instead of removing it
If the passage needs different wording or a better delivery, select it and choose Re-record words. The studio prepares a replacement recording for that range. Read the replacement, finish the take, and listen across both ends of the new passage.
A replacement or added take needs a fresh transcription to give the new audio accurate word selections. You can continue working with the waveform while deciding whether to transcribe again.
6. Export the result and keep an editable copy
Choose Export audio to preview and download the complete edited WAV. Save project backup keeps the script, transcript, original takes, and edit history in a .cadence file.
Deleting words from the edit does not erase the source take. This is useful for revising decisions later. Permanent deletion of a take is a separate action in the original-take recycle bin.
Make room for your next take.
Put this guide into practice with a short recording. Your scripts and audio stay on your device.
Open Voice Cadence