What a plosive actually is
Plosive consonants are made by closing the mouth or tongue completely, building up air pressure, and releasing it suddenly. In English the main ones are p, b, t, d, k and g, and p and b are usually the worst offenders because the lips release the air directly forward. In normal conversation that puff of air disperses before it reaches a listener. A microphone a few centimeters from the mouth catches it at full strength.
What the microphone records is not the consonant itself but the air hitting its diaphragm. That produces a short, powerful low-frequency thump, mostly well below the frequencies that carry speech, often lasting a few tens of milliseconds. A strong pop can be louder than anything else in the recording. It can overload the microphone or preamp, trigger a compressor so the following words duck in level, or set the peak that stops you from making the whole track louder.
Why some setups pop more than others
Distance is the biggest factor. Air from a plosive spreads and slows quickly, so moving from 3 cm to 10 cm makes a large difference, and moving slightly off the line of the mouth makes even more.
Microphone design matters too. Directional microphones, such as the cardioid models common for podcasting, respond to pressure differences between the front and back of the diaphragm, which makes them sensitive to blasts of moving air. They also show proximity effect: more bass the closer you get, which exaggerates the low-frequency thump. Omnidirectional microphones, including many lavaliers, are generally less prone to pops, though not immune. Large, sensitive condenser capsules placed close to the mouth are the classic popping combination; many broadcast dynamic microphones have a thick built-in windscreen partly for this reason. How these designs compare for spoken word is covered in choosing a microphone for speech.
Speaking style plays a part. Energetic, emphatic speakers and people who lean in when they get excited produce stronger pops than someone reading calmly.
Pop filters and foam covers
A pop filter is a screen placed between the mouth and the microphone that breaks up the air blast while letting sound pass. Common types:
- Nylon mesh screen
- One or two layers of fabric on a hoop, clamped to the stand. Inexpensive and effective; can slightly soften high frequencies
- Metal mesh screen
- Perforated metal, often angled to deflect air sideways. Durable, easy to clean, little effect on tone
- Foam windscreen
- A foam cover that fits over the microphone. Common on handheld and dynamic microphones; adds distance as well as diffusion
- Built-in grille and foam
- Many podcast and broadcast microphones have internal foam layers behind the grille
A screen works best a few centimeters in front of the microphone, with the speaker a few centimeters behind the screen. That spacing also sets a consistent minimum distance, which helps level and tone. A foam cover alone is often enough on a dynamic microphone; a sensitive condenser close to the mouth usually needs a proper screen.
Outdoor wind is a different problem that needs much heavier protection, such as furry covers and blimps, and is covered in reducing wind noise in outdoor video. A pop filter will not stop a breeze.
Microphone angle and distance
Changing where the air goes is free and often works better than any accessory:
- Place the microphone slightly to one side of the mouth, or just above or below it, and point it toward the mouth. The voice still reaches the capsule, but the stream of air passes beside it.
- Keep a consistent distance. For a typical podcast microphone, somewhere around a hand's width is a common starting point; a fist's width or more for sensitive condensers.
- Use the microphone's high-pass or low-cut switch if it has one. It reduces the low-frequency part of the thump at the source.
- Do a quick test: say a line full of p and b words, such as "Peter's puppy bounced past a big brown box", and listen on headphones before recording anything important.
- Brief guests. Asking someone to speak slightly across the microphone rather than straight into it is easier than fixing forty pops later.
An old studio trick is to fix a pencil vertically across the front of the capsule with a rubber band, so the air splits around it. It can help in a pinch, but a proper screen or a change of angle is more predictable.
Fixing pops after recording
Pops that slipped through can usually be reduced. Work on a copy of the original file and start with the least invasive method.
- Targeted high-pass filter. Select just the pop, usually the first few tens of milliseconds of the word, and apply a high-pass filter somewhere around 100–200 Hz to that region only. Because the thump is mostly low frequency, this removes it while leaving the consonant largely intact.
- Clip gain. Lower the level of the pop region by several decibels with a short fade at each edge. Simple and effective in any editor that allows region volume changes.
- Dedicated de-plosive tools. Many audio repair suites include a module that detects pops automatically and filters only those moments. Useful on long recordings with many pops; check the results, because some settings thin out the voice.
- Spectral editing. In a spectral editor, the pop shows as a bright low-frequency burst that can be attenuated directly.
What does not work well is a high-pass filter set high across the whole track. It reduces the pops but also strips body from every word. Apply your normal speech high-pass for the whole file, then treat the pops individually. The broader role of that filter is covered in EQ for voice recordings.
If a pop drove the signal into clipping, the waveform will show flattened tops and you will hear a crackle on top of the thump. Filtering removes the thump but not the distortion; repairing clipped audio explains what can and cannot be recovered.
A worked example
A narrator records a 25-minute training module with a large-diaphragm condenser about 4 cm from the mouth and no screen. Playback reveals heavy thumps on most p and b words, and the compressor ducks the next syllable each time. Rather than re-recording, the editor runs a de-plosive module across the file, then listens through and fixes six remaining pops with clip gain. For the next session, the narrator adds a nylon screen, moves back to about 12 cm, and angles the microphone slightly off to the side. The test line now records cleanly, and the editor's repair time drops to almost nothing.
Do plosives affect transcription
Occasional mild pops usually make little difference to speech recognition. The thump is brief and mostly sits below the frequencies that carry consonant detail, and modern recognizers are trained on plenty of imperfect audio.
Pops become a problem when they cause side effects: clipping that distorts the word, a compressor pumping that makes the following syllable much quieter, or a recording so dominated by thumps that the speech itself was recorded too quietly to keep peaks in check. Fixing the pops in those cases also fixes the levels, and the following words become easier to recognize.
Common mistakes and limits
- Relying on a pop filter while the speaker moves to 2 cm from it. Distance and angle still matter.
- Filtering the whole track aggressively to hide a handful of pops.
- Repairing pops after compression and limiting. The compressor has already reacted to them; repair first, then process.
- Assuming every thump is a plosive. Desk knocks, cable movement and handling noise sound similar but need different fixes; lavalier rustle in particular is covered in lavalier microphone noise.
- Expecting repair to be invisible on severe pops. A strong pop that overloaded the preamp can only be reduced, not removed cleanly.
What this means for mydubly
mydubly has no pop removal or audio repair tools, so any de-plosive work happens in your editor before you upload. When a job starts, the browser decodes the audio, mixes it to mono at 16 kHz and sends it in roughly 30-second compressed chunks for recognition. Conversion to 16 kHz keeps frequencies up to 8 kHz, which includes the low-frequency thump of a pop, so repairing severe pops beforehand is worthwhile when they distort words or cause level pumping.
If you dub the video, the translated version replaces the original voice with a synthesized one, so pops in the original voice do not carry over into it. They can still affect the recognized text that the translation and subtitles are built on. For podcast workflows, the podcast transcription page covers formats and outputs.
Next step
Before your next recording, run the test line with your usual setup and listen on headphones. If it pops, change the angle first, add a screen second, and move back third. For recordings you already have, repair the worst pops, then transcribe the file with mydubly's audio to text tool and check the passages around any remaining thumps.
Frequently asked questions
What is the difference between a pop filter and a windscreen?
A pop filter is usually a mesh screen mounted in front of the microphone to break up air from plosives indoors. A foam windscreen fits over the microphone and does a similar job with some protection against light air movement. Neither is enough for outdoor wind, which needs furry covers or full blimps.
Why do I get pops even with a pop filter?
The most common reasons are speaking too close to the screen, pointing the microphone straight at the mouth, or a screen placed directly against the microphone so it cannot diffuse the air. Leave a gap between screen and capsule, back off a little and angle the microphone slightly off-axis.
Do dynamic microphones pop less than condensers?
Often, but not always. Many dynamic microphones used for broadcasting have heavy internal windscreens and are less sensitive, which helps. A dynamic microphone used very close and straight on can still pop, especially with strong proximity effect, so technique still matters.
Can I remove plosives automatically?
Yes, many repair tools include automatic de-plosive processing that detects and filters pops. Results vary with the recording and the settings, so listen through afterwards. Manual fixes with a short high-pass filter or clip gain are slower but give more control for the pops that remain.
Will plosives make my transcript less accurate?
Mild pops rarely matter because they are short and mostly low frequency. Severe pops that clip, or that make a compressor duck the following words, can obscure parts of speech. Repairing those before transcription and checking the affected passages is the practical approach.
Are plosives a problem in languages other than English?
Yes. Most languages have stop consonants that release a burst of air, though how strongly aspirated they are varies between languages and speakers. The same fixes apply: distance, angle and a screen work regardless of the language being spoken.