A stereo effect that tames low-frequency bursts in recorded voices: the thump of a hard "p" or "b", a handling knock or a gust of wind on the mic. It turns down only the low band, only while a burst is happening, and lets you listen to exactly what it takes away.
The idea
A plosive is mostly energy below a few hundred hertz that arrives suddenly. Plosive Repair Lab splits off that low band, watches for it to jump in level, and reduces it for as long as the burst lasts. Everything above the band passes untouched. Because the removed part is simply subtracted from your signal, you can solo it and hear whether the repair is taking only the pop or some of the voice too.
Controls
- Cutoff (80 to 350 Hz) sets how much of the low end is treated.
- Threshold (−48 to −6 dB) is the low-band level at which Auto starts reducing.
- Amount (0 to 18 dB) is the most Auto will reduce, or the fixed reduction in Manual. At zero the sound is unchanged.
- Release (20 to 250 ms) sets how quickly the low band comes back after a burst.
- Mode: Auto reacts to the signal; Manual applies a constant reduction that you automate in your host where it is needed. Threshold, Release and Detection only affect Auto.
- Detection: Protect sustained reacts to sudden rises in low energy, so a held low note is left alone; Level only reacts to level alone, which also catches longer rumbles.
- Audition: Processed is the result, Removed is only what was taken out, and Dry is the original.
A meter shows how much the low band is being reduced while audio plays.
Getting started
Put it on a vocal or spoken-word track. Start with Gentle low-band catch, or try Close-mic vocal pops, Deep voice, keep body for low voices, or Wind and handling for lavalier knocks and gusts. Switch to Removed to check that you hear only the pops, compare with Dry, then return to Processed before you export. Hear what is removed starts directly in that listening mode, and Manual region automation is a starting point for automating Amount by hand.
Limits
This is a detector, not a speech analyser. Vowel onsets, bass notes and other sudden low sounds can also trigger it, and since it does not look ahead, the very start of a burst can slip through. It does not find or edit regions for you: in Manual mode, automate Amount back to zero wherever no repair is wanted. The included demo is a synthetic test burst, not a real voice.
