How to Improve Audio in Talking Head Videos

The fastest way to improve audio in a talking head video is to fix it before you hit record: put a microphone close to your mouth, record in a room with soft surfaces, and turn off anything that hums. Then, in the edit, apply light noise reduction, even out the volume, add a little EQ if your voice sounds muddy or harsh, and bring the final loudness to a consistent level. Recording fixes do most of the work; editing polishes.

Why audio matters more than picture

Viewers will watch a slightly soft or poorly lit video if they can hear you clearly. They'll leave a crisp 4K video if the voice is echoey, buzzing or too quiet. Audio problems are tiring to listen to, and people don't always know why they're leaving; it just feels like effort.

Talking head content is almost entirely speech, so your voice is the content. If you have a limited budget, a microphone is usually a better first upgrade than a camera.

Recording fixes (do these first)

Every problem below is easier to prevent than to fix later.

Get the mic close

Distance is the biggest factor in voice quality. The farther the mic is from your mouth, the more room echo and background noise it picks up relative to your voice. Options:

• Lavalier mic clipped to your chest, about a hand-width below your chin. Cheap, consistent, invisible-ish.
• Shotgun or small condenser mic just out of frame, above or below, pointed at your mouth.
• USB mic on a desk arm for sit-down videos.
• Your phone as an audio recorder placed close to you, out of shot, if you have nothing else.

A camera's built-in mic a few feet away is the setup most likely to sound bad.

Tame the room

Echo (reverb) comes from sound bouncing off hard, flat surfaces. Bare walls, hard floors and empty rooms are the enemy. Soft surfaces absorb sound: rugs, curtains, sofas, bookshelves, a bed, a closet full of clothes. Record in the most furnished room you have, and avoid facing a bare wall closely.

Quick test: clap once. If you hear a ringing tail, the room will sound echoey on video.

Kill the noise sources

• Turn off fans, air conditioning and noisy appliances while recording.
• Put your phone on do-not-disturb.
• Close windows and doors.
• Keep laptops with loud fans away from the mic.

Set the level properly

Record a test sentence at your loudest natural volume, like when you laugh or emphasize a point. Adjust the input level so those peaks land well below the maximum, roughly between -12 and -6 dBFS on the meter. If audio hits the top (clipping), it distorts, and distortion can't be undone in the edit. Slightly too quiet is fixable; too loud is not.

Our guide on recording talking head videos for easy editing covers the rest of a good recording setup.

Noise reduction in the edit

Most editors and audio apps include noise reduction or a "voice enhance" feature. They work best on steady, consistent noise: fan hum, air conditioning, computer whine, electrical buzz.

• Start gentle. Apply a small amount, listen, then increase. Heavy reduction creates a watery, metallic or "underwater" voice that sounds worse than mild noise.
• Listen on headphones and on a phone speaker. Artifacts that are obvious on headphones can disappear on a phone, and the reverse.
• Don't expect miracles on random noise. Barking dogs, traffic and voices in the background are hard to remove cleanly. Re-record or cut those moments.
• Be careful with AI voice enhancers. They can do impressive cleanup, but sometimes make voices sound slightly synthetic or change their character. Compare with the original before committing.

Levels and loudness

Two different things matter here: consistency within the video and overall loudness.

Consistency

Speech naturally varies: you get quieter at the end of sentences and louder when excited. A compressor reduces the gap between quiet and loud parts so viewers don't have to keep adjusting the volume. Use gentle settings; too much compression sounds flat and squashed and brings up background noise.

If you've cut together takes recorded at different times, check that each segment sits at the same level before compressing.

Overall loudness

Loudness is measured in LUFS (loudness units relative to full scale), which approximates how loud audio sounds to people rather than just its peaks. Many online platforms normalize playback volume, so extremely loud masters gain nothing. A commonly used target for online video is around -14 LUFS integrated, with true peaks kept below roughly -1 dB to avoid distortion after platform processing. Your editor may have a loudness meter or a "normalize loudness" option.

The practical goal: your videos should sound about as loud as the ones before and after them in a feed, and as loud as each other.

EQ basics

EQ (equalization) turns specific frequency ranges up or down. You don't need to be an audio engineer to use it for voice. Common fixes:

• Rumble and boominess: a high-pass (low-cut) filter around 80-100 Hz removes low rumble from traffic, desks and handling noise without affecting most voices.
• Muddy or boxy voice: a small cut somewhere in the 200-500 Hz range often clears it up. Sweep slowly and listen.
• Harsh or piercing voice: a gentle cut in the 2-5 kHz range softens it.
• Sharp "s" sounds: a de-esser reduces hissing sibilance, which gets worse on phone speakers.
• Dull voice: a small, broad boost in the upper range can add clarity, but go easy; it also brings up noise.

Small moves. If you're boosting or cutting more than a few decibels, the problem is probably in the recording.

Music under voice

Background music can make a talking video feel produced, but it's the most common reason voices become hard to understand.

• Keep music well below the voice, so far that you'd barely notice it if it stopped. A common starting point is music sitting roughly 15-25 dB below the voice level, then adjusting by ear.
• Pick tracks without vocals, and avoid busy melodies in the same range as your voice.
• Use ducking (automatically lowering music when you speak) if your editor supports it.
• Check on a phone speaker, where music often seems louder relative to the voice than on headphones.

For licensing and choosing tracks, see background music for YouTube videos.

A five-minute audio check before publishing

1. Listen to the first 30 seconds on headphones. Any hum, hiss or echo?
2. Listen to a random middle section on your phone speaker. Is every word clear?
3. Jump between sections. Does the volume stay consistent?
4. Check the music: can you understand the voice without effort?
5. Check the end: no abrupt cut-off or trailing noise.

Trimming pauses also helps audio feel tighter, because long gaps expose room noise. ShotFlick's Auto Trim removes pauses and dead air from talking-head clips as part of the edit, but it can't rescue a recording that's distorted or drowned in echo, so the recording fixes above still matter most. For a full pre-publish list, use our video editing checklist.