Tutorials Audio

Audio Recording Fundamentals: The Complete Guide

Beginner · ~35 min

Overview

Good recordings are mostly won before anyone touches a fader. Four physical facts, how far the microphone is from the source, what the room does to sound on its way there, which direction the microphone listens in, and how much gain is applied where, determine most of the final quality, and no amount of processing fully reverses a bad answer to any of them. This guide covers those four in the order they matter, explains the vocabulary that gets used loosely elsewhere, and points to the specific guides for each situation. If you learn one thing here, make it the distance rule.

What You Need

Any microphone other than the one built into your laptop or camera
A stand, boom, arm, or clip that holds the microphone in position without hand noise
Closed-back headphones for monitoring while recording
An interface or recorder, or a USB microphone, which contains both
A room with soft furnishings, or anything absorbent you can put around the microphone
Recording software: Audacity, REAPER, and most DAWs have free or low-cost tiers

Steps

Get the microphone close before anything else

Direct sound weakens rapidly with distance while reflected room sound stays roughly constant, so distance does not merely make a source quieter. It changes the ratio of source to room. Halving the distance is worth more than most equipment upgrades. For speech, a hand-span from the mouth and slightly off-axis is a reliable starting position: close enough to dominate the room, angled enough to avoid plosive blasts.

Treat the room, or choose a better one

Hard parallel surfaces create the boxy, hollow character that makes amateur recordings identifiable instantly. You do not need acoustic panels to improve this. Soft furnishings, a duvet, curtains, a bookshelf of unevenly-sized books, or recording in a wardrobe all reduce reflections meaningfully. The goal is absorption and irregularity near the microphone, not silence throughout the space.

Point the microphone using its pattern, not your intuition

A cardioid microphone is most sensitive in front and rejects what is behind it, which means you aim its rear at the problem (the air conditioner, the window, the noisy laptop) rather than only aiming its front at the source. Omnidirectional capsules hear everything equally and suit lavaliers and room ambience. Shotgun microphones narrow the front lobe for distance work but still hear behind them, which surprises people.

Set gain for headroom, not for loudness

Aim for average speech peaking around -12 dBFS. This sounds conservative because it is: digital clipping is unrecoverable, while a quiet-but-clean recording can be raised later at the cost of a little noise. Raise the preamp until it sounds strong, then back it off. If your recorder offers 32-bit float, clipping ceases to be the constraint, but sensible gain still gives you the best noise performance.

Monitor on headphones while you record

You cannot fix a problem you did not hear, and most of the worst ones: a loose connector crackling, a fridge compressor cycling on, clothing rustling against a lavalier, are obvious in headphones and invisible on a level meter. Closed-back headphones prevent the monitor signal leaking back into the microphone. Listen to the first thirty seconds of every session critically rather than trusting the setup from last time.

Capture room tone before you pack up

Record thirty seconds of the space with everyone silent and nothing moving. This gives the editor a matching bed to fill gaps between takes, and gives noise-reduction tools a clean profile of exactly what to remove. It costs half a minute and is the single most-appreciated thing you can hand to whoever edits the material.

Record safety takes and redundant sources

Where the material cannot be repeated (interviews, live events, one-off performances) record a second source. A backup recorder, a second microphone, or the camera scratch track all serve. Redundancy is the only defence against the failures that are invisible until you open the files: a dead battery, a cable that was fine until it was not, a card that will not mount.

Edit and level before you reach for processing

Remove the obvious problems structurally first, cut the noisy sections, replace the flubbed phrase, ride the level where someone leaned away. Compression, equalisation, and noise reduction are for what remains, and each is much gentler and more transparent when it is not being asked to solve a problem editing could have solved outright.

Pro Tips

Distance beats price. A budget microphone six inches away will beat an expensive one across the room, every time.
Turn things off rather than filtering them out: fridges, fans, heating, phone notifications, and anything with a fan or a compressor.
Wear headphones that fully enclose the ear. Open-backed models leak enough sound to be picked up on quiet passages.
Record a short test and actually listen back before committing to a long session. Sessions get ruined by settings nobody checked.
Handle cables and stands gently while recording, microphones transmit structure-borne vibration far better than people expect.
Name and back up files immediately after the session, while you still remember which take was which.

Knowledge Base

What You'll Learn

This is the parallel pillar to the video fundamentals guide, and it is the page the other audio tutorials on this site assume. Below: why distance dominates, what the microphone specifications actually mean, how levels and loudness differ, and where each specific guide fits.

The distance rule, and why it dominates everything

The single most useful fact in audio recording is that direct sound from a source falls off quickly as you move away, while the reflected sound bouncing off the walls, ceiling, and floor stays roughly constant throughout the room. The consequence is that microphone distance controls the ratio between the two, not just the volume.

This is why a voice recorded from across a room sounds hollow and distant even after you turn it up: raising the level raises the room reflections by exactly the same amount. The hollowness is baked into the ratio, and no plugin can separate a voice from its own reflections with real fidelity. Noise reduction and de-reverberation tools have improved substantially, but they work by estimating and subtracting, and what they subtract they take from the voice too.

The practical rule: before adjusting anything else, before buying anything else, move the microphone closer. Most recordings that people describe as sounding amateur are simply recordings made from too far away in an untreated room.

Polar patterns, condensers, and dynamics, without the marketing

Two specifications matter for most decisions. The first is the polar pattern, which describes which directions the microphone is sensitive to.

Cardioid is most sensitive in front and rejects the rear: the default for speech, podcasting, and single sources. Omnidirectional is equally sensitive in all directions, which makes it natural-sounding and resistant to plosives, and is why almost all lavaliers are omni. Shotgun narrows the forward pickup for reaching a source from further away, at the cost of an unnatural character off-axis. Figure-of-eight hears front and back while rejecting the sides, useful for two people facing each other.

The second is the transducer type. Dynamic microphones are robust, handle very loud sources, and hear less of the room, which makes them forgiving in untreated spaces. The reason broadcast studios use them. Condenser microphones are more sensitive and capture more detail and high-frequency air, which also means they capture more room. In a poor room a dynamic microphone will usually produce the better recording, which is the opposite of what the price difference suggests.

Gain, levels, and loudness are three different things

These terms get used interchangeably and they are not the same, which causes real confusion.

Gain is the amplification applied at the input, before recording. It sets how much of the microphone signal is captured relative to the noise floor of the electronics. Too little gain buries the source in hiss. Too much clips it.

Level is the instantaneous amplitude of the recorded signal, measured in dBFS, where 0 dBFS is the maximum a digital system can represent. Exceeding it does not distort gracefully the way analogue tape did: it hard-clips, and the clipped portion is simply gone.

Loudness is the perceived volume over time, measured in LUFS, and it correlates poorly with peak level. A heavily compressed track and a dynamic one can share a peak level while sounding very different in loudness. Platforms normalise to loudness targets, which is why mastering something extremely loud no longer wins anything. It just gets turned down, having sacrificed dynamics for nothing.

What noise reduction can and cannot do

Modern noise reduction is genuinely impressive and routinely oversold. It works well on stationary noise (a constant hum, hiss, fan, or air conditioner) because a steady noise can be profiled from a silent passage and subtracted with reasonable accuracy. This is exactly why capturing room tone matters.

It works poorly on non-stationary problems: a door slamming, a chair creaking, overlapping speech, traffic that rises and falls, wind gusting. These change unpredictably, so there is no stable profile to subtract, and aggressive attempts produce the watery, artefact-ridden sound that is worse than the original noise.

Reverb sits in the hardest category. Because reflections are the source itself arriving late, removing them means separating a signal from delayed copies of itself. Tools do this increasingly well, but always at some cost to the voice. The reliable fix remains getting closer and deadening the room.

Which guide to read next

For spoken word, the podcast audio and clean-audio guides cover the specific chain for voice. For location work, field recording basics covers wind, handling noise, and working in uncontrolled spaces, with the iZotope RX wind-noise guide for repair after the fact.

For the technical layer, the audio interface guide explains what the hardware between microphone and computer is doing, 32-bit float recording explains why some recorders cannot clip, and the limiters and true-peak guide covers the output end.

For delivery, the LUFS guide explains loudness targets and why your podcast should sit around a specific number rather than as loud as possible. The loudness normalizer and EQ reference tools on this site are useful alongside it.

FAQ

Q: What is the best microphone for beginners?
A: A dynamic cardioid microphone on a stand, positioned close, in a room with soft furnishings. The specific model matters far less than the placement. Most well-reviewed microphones in any given price bracket perform similarly, and all of them are ruined by being two feet away in a bare room.

Q: Why does my voice sound hollow or distant?
A: The microphone is too far away, the room is too reflective, or both. Direct sound drops off with distance while room reflections do not, so distance changes the ratio between your voice and the room. Move closer and add absorption nearby before considering any processing.

Q: Should I record in 24-bit or 32-bit float?
A: Use 24-bit as the sensible default. It provides ample dynamic range for careful gain staging. Use 32-bit float if your recorder offers it and your material is unrepeatable or unpredictable in level, because it makes clipping effectively impossible and lets you set level after the fact.

Q: How loud should my recording be?
A: While recording, aim for peaks around -12 dBFS to preserve headroom. For delivery, work to a loudness target instead: broadly -16 LUFS for podcasts and streaming speech content, and around -14 LUFS for music on most streaming platforms. Peak level and loudness are different measurements and both matter at different stages.

Q: Can I fix bad audio in post?
A: Partially. Steady background noise can be reduced convincingly, small clicks repaired, and levels balanced. What cannot be undone is clipping, heavy room reverb, and anything masked by overlapping sound. Those are permanent, which is why placement and room choice deserve more attention than processing.

Q: Do I need an audio interface if I have a USB microphone?
A: No. A USB microphone contains its own preamp and converter, which is precisely what an interface provides. An interface becomes worthwhile when you need multiple microphones, want to use XLR microphones, or need low-latency monitoring of several sources at once.