A spoken take is rarely tight. You stop to think, you breathe, you look at your notes, and the recording keeps running. For a podcast intro, a voice-over or a narrated short, those pauses are what makes it feel slow, and cutting them by hand means finding each flat stretch of waveform, selecting it, deleting it and dragging the next sentence over. The Sound Editor in the osFoundry desktop app does the whole take in one pass with a command called Strip Silence, and a second command closes the gaps it leaves.
The example below is a 27.6-second test recording: five short sentences with two to three and a half seconds of quiet between them, and a little room hiss under everything.
Open the recording in the Sound Editor
In the Sound Editor, choose File → New… and click Create to start an empty session. Drag the WAV file onto the window. It lands on a new track named after the file. Use the zoom-out button in the toolbar so the whole take fits on screen.
Every flat line between the bursts is a pause. In this take they add up to more than half the running time.
Strip the silence
Click the clip and press Command+A, then choose Edit → Strip Silence (Command+U). The dialog has four numbers.
Threshold is the level below which audio counts as silence. The default of −48 dB sits above the hiss of a quiet room and far below speech, so for a clean recording it can stay. In a noisier room, raise it until the flat stretches stop counting as sound.
Min strip duration is the setting that decides how the result sounds. Only quiet stretches longer than this get removed. The default is 50 ms, short enough that the tiny gaps inside a sentence, between words and on a quick breath, count as silence too, and a sentence comes back in pieces. At 500 ms, those short gaps stay inside the clip and only the real pauses between sentences go.
The two pads extend each clip a little past the point where the sound crosses the threshold. The start of a word, especially a soft consonant like an "h" or an "f", is often quieter than −48 dB, and a clip that starts exactly at the threshold clips it off. Here the start pad is 100 ms and the end pad 300 ms, so each sentence keeps its first sound and a short tail after the last word.
Click Strip. The one long clip becomes five, one per sentence, with empty track between them.
Close the gaps
Stripping removes the silence, but it does not move anything. The session is still 27.6 seconds long, it is just quiet in different places.
Choose the Grabber tool (the hand in the toolbar), click the first clip, and Shift-click each of the others. Then choose Edit → Space Clips…. The clips slide together end to start, with no gap.
There is still a natural pause between sentences, and it comes from the pads: the 300 ms tail of one clip plus the 100 ms lead-in of the next gives a 0.4-second breath at every join. If the result feels rushed, raise the end pad and strip again rather than adding gaps by hand.
Bounce one file
Click an empty spot in the timeline so nothing is selected. Bounce Mix bounces the edit selection, or the whole session when nothing is selected, and the whole session is what you want. Choose File → Bounce Mix…, tick Normalize to −0.1 dBFS, and click Bounce.
The file that came out of this example is 13.76 seconds long, peaking at −0.1 dBFS: the same five sentences in half the time. The Sound Editor runs inside the app on your own machine, so the recording is never uploaded anywhere, which fits the local-first way the osFoundry desktop app works.
For a different recording the numbers to revisit are the threshold, if the room is noisy, and the end pad, which sets the rhythm. The desktop app is free from the osFoundry download page for macOS, Windows and Linux.
