19 October 2026

How to Record Audio for Video: Clean Sound in Any Room (2026)

Your camera has a microphone. It is bolted to the camera, which means it hears the room from wherever the tripod stands. That is why footage can look expensive and sound like a speakerphone call from 2009.

Learning to record audio for video is mostly about three things: distance, gain, and what sits behind the microphone. The gear matters, but less than the order you do things in.

How to Record Audio for Video: The Direct Answer

Put the microphone 15 to 20 cm from the speaker’s mouth, set gain so the loudest moment peaks near -12 dBFS, and record at 48 kHz and 24-bit. That combination fixes most bad sound before you open an editor.

If the mic sits across the room, no plugin will save the take. You are capturing voice and room at nearly the same level, and separating them afterwards is guesswork dressed up as repair.

What You Actually Need

The setup to record audio for video is smaller than most people expect. One microphone you can move independently of the camera, one pair of closed-back headphones, and a soft surface to put behind the mic covers the majority of shoots.

Microphone type	Where it wins	Where it fails
Dynamic handheld	Loud or echoey rooms, one speaker	Needs more gain, less fine detail
Condenser	Quiet treated rooms, voiceover	Hears the fridge, the street, the neighbour
Lavalier	Movement, interviews, walking shots	Clothing rustle, cable noise
Shotgun	Outdoors, or soft-surfaced interiors	Hollow indoors near hard walls
USB microphone	Solo recording at a desk	Usually one input, no XLR upgrade path

Two inputs cost more than one, and that is the main reason to buy an interface. A single USB microphone is enough for one person at a desk. A phone is enough for a quick interview, provided you get it close.

Method 1: Record With a Separate Microphone

This is the setup for interviews, tutorials and anything where the voice carries the video. It takes about ten minutes to arrange. Those ten minutes decide whether the result sounds considered or accidental.

1. Place the mic first, the camera second. Angle the capsule 15 to 30 degrees off the mouth axis so air from plosives misses it. Aim at the chin rather than the nose, and keep the distance consistent between takes.

2. Set the pickup pattern and the filter. Cardioid rejects sound arriving from behind the capsule, which is why the mic’s rear should face the emptiest part of the room. If you are recording voice only, engage the low-cut filter around 80 Hz.

3. Set gain with a real performance. Ask the speaker for a full-volume sentence, raise gain until peaks land near -12 dBFS, then stop. That headroom is not wasted space; a laugh or a hard “p” will use it.

4. Record ten seconds of room tone. Everyone silent, mic live, gain unchanged. Noise reduction needs that sample to subtract the room instead of guessing what the room sounds like.

5. Monitor on closed-back headphones with direct monitoring on. If you hear yourself half a second late, mute the software return. Latency makes people slow their speech without noticing, and the take sounds tired.

6. Slate every take. One sharp hand clap in frame, close to the mic. That transient is the cleanest alignment point you will get when picture and sound arrive as separate files.

7. Leave the camera mic running. Scratch audio costs nothing and gives you a fallback plus a sync reference if the recorder stops.

Method 2: Use the Mic You Already Have

The fastest way to record audio for video is the phone in your pocket. Corridors, factory floors and five-minute quotes rarely allow a stand, a boom and a sound check.

1. Switch to airplane mode and turn off vibration. A notification buzz arriving mid-sentence is unrecoverable.

2. Hold or mount the phone 20 to 30 cm from the mouth, just below the frame line. Closer than that picks up breath; farther away picks up the corridor.

3. Record in a voice memo app at the highest quality setting, with one clap at the start. Phone camera apps often apply processing you cannot switch off.

4. Keep the phone still. Handling noise travels through the body of the device and lands in the recording as low thumps.

5. Get closer rather than louder. Shouting into a distant mic sounds like shouting into a distant mic.

If the phone is also the camera, its mic sits beside the lens: it hears the frame, not the subject. Move the whole phone closer and shoot wider. Zooming in from across the room leaves the audio exactly where it was.

What the Gear Does, and Where It Stops Helping

A dynamic microphone is insensitive on purpose. It ignores most of the room, which is how a $90 handheld beats a studio condenser in a kitchen. A condenser hears detail, and it hears everything else too.

A shotgun looks directional and behaves that way outdoors. Indoors, sound reflects off the wall behind the subject and reaches the capsule a second time, slightly late. The two copies cancel at some frequencies, which is why a shotgun in a bare room can sound hollow next to a cheap lavalier.

Lavaliers win on consistency because they stay 20 cm from the mouth no matter where the subject turns. They lose to clothing: a collar rub at that distance is louder than the voice.

USB microphones remove the interface, and usually remove the second input with it. Some inexpensive models top out at 16-bit, which gives you 96 dB of range against 24-bit’s 144 dB. Since bit depth and noise floor are linked, that difference shows up as hiss in quiet rooms.

None of it matters if the mic is far away. Placement beats price every time, and microphone placement practice has not changed much in fifty years for the simple reason that physics has not either.

Rights and Responsible Use

The recording belongs to whoever made it. Your own voice and your own footage are yours to publish. When other people are audible, their voice is still their own, so ask before publishing and get it in writing for anything commercial.

Rules on recording conversations differ by country, and within the United States by state. Some places require everyone on the call to agree. A recording made without that consent can be unusable long after the technical side is perfect.

Music is a separate permission. A track you bought, or one from a subscription library, is licensed on that licence’s terms and nothing more. It does not transfer ownership, and it rarely covers every platform you might publish to. Filming a concert or a screening adds the venue’s rules on top of the copyright holder’s.

None of this changes because capture is easy. Holding a copy of something grants no rights over it. YouTube’s copyright help page explains how claims work on the platform where most of these recordings end up.

Troubleshooting: Six Things That Go Wrong

There is a constant hum under the voice

That is mains hum at 50 or 60 Hz, and it is usually a ground loop rather than a broken mic. Run the laptop on battery to test it. Dimmer switches, fridges on the same circuit and USB bus power are the common sources; a ground loop isolator is cheaper than a re-record.

The recording is quiet and hissy

Gain was set too low and the level was raised later. Hiss gets raised with the voice, so the ratio never improves. Set gain at the interface, move the mic closer, then treat any remaining floor with tools that remove background noise.

Only the plosives distort

Air from a “p” or “b” is hitting the capsule directly. Add a pop filter, angle the mic further off axis, or back off five centimetres. If the waveform is already flat-topped, that syllable is gone and no amount of gain riding brings it back.

It sounds like a bathroom

That is reverb, and it is a room problem before it is a plugin problem. Soften the wall the speaker faces and the wall behind the mic: a duvet behind the capsule is worth more than any de-reverb setting. Turn the mic’s rear toward the emptiest corner.

Audio drifts out of sync over long takes

Camera and recorder run on separate clocks that never match exactly. The gap is invisible at thirty seconds and obvious at twenty minutes. Keep scratch audio on camera, resync at the end, or use timecode; the fixes for audio out of sync cover the manual route.

The condenser records nothing at all

Phantom power is off, or the cable is wrong. Condensers need 48 V and dynamics do not, so a dynamic mic will happily work on a channel where a condenser stays silent. Check the interface switch and confirm you are using XLR, not an instrument cable.

Frequently Asked Questions

Do I need a separate microphone to record audio for video?

No. A phone held 30 cm from the mouth beats a camera mic across the room. Buy a dedicated mic when you need two people at once, when the room is bad, or when you record often enough that setup time matters.

Is 48 kHz better than 44.1 kHz for video?

Yes, for video work, 48 kHz is the safer default. It divides evenly into common frame rates and matches what cameras, editors and delivery platforms expect. The audible difference is nil; the workflow difference is real, and sample rate for video explains where the mismatch bites.

Should I record in 16-bit or 24-bit?

Record 24-bit whenever the option exists. It gives you far more room to set conservative levels without hiss, and it costs about 50 percent more disk space for audio that is a rounding error next to video.

Why does my audio sound worse than my video looks?

Because the mic is on the camera. Picture quality depends on the sensor you point at the subject; sound quality depends on the distance between the mic and the mouth. Those two distances are set by the same tripod, and one of them is usually wrong.

Do I need an audio interface?

Only for two inputs, phantom-powered XLR mics, or cleaner gain. One person at a desk with a USB mic does not need one. Two people on separate mics, or a condenser that needs 48 V, does.

Can bad audio be fixed in editing?

Partly. Noise reduction and EQ help a close, clean take with a steady floor. They cannot rebuild a voice that was recorded at the same level as a wall, and they cannot undo clipping. Ten seconds of setup beats an hour of repair.

How do I stop the mic picking up my keyboard?

Move the mic closer to your mouth than to the keyboard, and put something soft under both. A cardioid pattern pointed away from the desk rejects most of it. If the keys are still audible, the mic is too far away, not too sensitive.

Which Method Should You Use?

Use Method 1 when the voice is the product: interviews, courses, narrated tutorials, anything a viewer will listen to with headphones. The extra ten minutes of setup is the whole quality difference.

Use Method 2 when the alternative is no audio at all, or when the subject is moving through a space you cannot control. Get the phone close, clap once, and accept the room.

Skip both when the video is music-led or silent, and the sound design will be built in the edit. Recording a clean room tone still helps in that case, because silence in a video is never actually silent.

Whichever route you take to record audio for video, the rules do not move: close mic, conservative gain, soft surfaces. If you are adding narration after the cut instead, the workflow in add a voiceover to a video picks up where this leaves off.