Remove clicks, pops and impacts from an audio recording
When to use De-click or spectral attenuation, and how to preserve speech.
De-clip
Estimate repairs for short clipped peaks in overloaded audio.
When should you choose this tool?
Distortion on loud syllables and flat waveform tops can indicate clipping. De-clip estimates new peaks for short clipped sections. If audio is merely loud but undamaged, Gain is the simpler choice.
How it changes audio
Threshold and maximum width identify short peak defects; interpolation estimates their missing shape. Repair can create higher peaks, making output attenuation part of the setup.
What do the parameters control?
- Peak threshold
- −24–0 dBFS. Match the level of the actual flat top. The processing graph shows affected regions from current audio.
- Maximum repairable peak width
- 0.1–20 ms. Longer overloads increase uncertainty; do not widen merely to include every event.
- Peak recovery limit
- 0–18 dB. Bounds estimated peak recovery; it does not establish the original peak height.
- Output gain
- −18–0 dB. Leaves room for repaired peaks. This tool has no limiter; check final levels separately.
Try a short passage
- Zoom into a distorted syllable and find the clipped level. Match the threshold to it.
- Leave output headroom and start Preview. Raise width only to cover the actual short defect.
- Compare the distorted syllable at comparable listening levels, then check peaks after applying.
What to listen for
The harsh edge should soften without a new click or excessive peak. Judge the sound as well as the graph.
Limits of the method
An estimated peak is not proof of the lost original. Long flattened passages and analogue distortion may improve only partially.
Preview covers a short passage. After applying, review the full treated section; you can undo the step from history.
Related problems and workflows
Fix clipped, distorted audio: De-clip or peak reduction?
Identify clipped peaks, evaluate estimated repair and leave output headroom.