Bad recordings rarely happen because of one big mistake. They happen because of five small ones — checked in the wrong order, or not at all.
This is a pre-flight checklist. Short, sequenced, and built around the mistakes that actually show up in post-production — not the ones that make for good forum arguments. Run it before you hit record, every time.
Gain is the first thing to check because it’s the hardest to fix retroactively. Get it wrong and you’re either fighting clipped distortion that no plugin removes, or boosting a weak signal until the noise floor comes up with it.
The common mistake is setting gain to “loud enough to see movement on the meter” and calling it done. That’s not the right target. The right level is about headroom — leaving enough space above your speaking level that an unexpected laugh, an emphasised point, or a moment of raised volume doesn’t push the signal into clipping.
The practical check: set your gain, then speak at the loudest level you realistically expect to hit during the session — not your conversational tone, your peak. If you tend to laugh loudly or get animated when making a point, do that during the check. Your peaks should land between -12dB and -6dB on the meter. That gap between your loudest moment and 0dB is the headroom you’ll be glad you left when the session doesn’t go exactly as rehearsed.
This is a quieter problem than the others, but it costs more time than almost anything else on this list when it goes unnoticed.
Many recording setups route monitoring audio differently from what’s actually being captured. You might be hearing the processed output — with software EQ, compression, or a monitor mix applied — while the raw signal being recorded is unprocessed. Or you might be hearing direct monitoring from your interface while your DAW is recording at a different gain stage entirely.
The risk: you spend the session adjusting based on what you hear, while the actual recorded signal has a completely different problem you can’t detect by listening. You finish, sit down to edit, and discover an issue that was invisible during the session because you weren’t actually monitoring the captured signal.
Before you record, confirm explicitly: is the audio in your headphones coming from the live input, the processed monitor mix, or the recorded track itself? Most interfaces and DAWs let you toggle this, and most people have never checked which one they’re set to. Knowing the answer saves hours of chasing problems that don’t exist — or missing ones that do.
This is the step most people leave out entirely, and it’s the one with the most consistent impact on recording quality.
Your room has a noise floor — a baseline level of ambient sound that’s always present, whether or not anyone is speaking. Air conditioning, refrigerator hum, traffic outside, the faint electrical buzz of certain lighting fixtures. Your brain filters these out automatically because it’s adapted to your environment. Your microphone does not filter anything out. It captures all of it, continuously, underneath every word you say.
The check is simple and takes 30 seconds: before you start recording your actual content, hit record and stay completely silent for 30 seconds. Don’t speak, don’t move, don’t touch anything. Then listen back — ideally on headphones, in a quiet space yourself.
What you’re listening for: any continuous hum, hiss, rumble, or background tone. If you hear something, you now know two things. First, that sound is going to be present underneath your entire recording whether you address it or not. Second, you have time to fix it — turn off the AC, unplug the fridge in the next room, close a window — before you’ve recorded anything you’d need to redo.
This single test tells you more about your actual recording environment than any plugin, meter, or piece of software ever will. A spectrum analyser shows you frequencies. A 30-second silence test shows you the truth.
This check exists because the failure mode is silent. Your DAW doesn’t always warn you clearly when something is misconfigured — it just records, and you don’t find out until you’re listening back or trying to export.
Two things to confirm before every session:
Storage location. Confirm your DAW is actually writing to the drive you think it is, and that there’s enough free space for the length of session you’re planning. An external drive that’s disconnected, a project pointed at the wrong folder, or a drive that’s nearly full are all failure points that won’t necessarily stop the recording — they’ll just produce a corrupted or incomplete file you discover later.
Sample rate. Confirm your interface, your DAW project settings, and your operating system’s audio settings are all set to the same sample rate. A mismatch here doesn’t always throw an error — sometimes it just produces audio that sounds subtly wrong (pitched incorrectly, or with artefacts) without an obvious cause. Discovering a sample rate mismatch in post-production, after a session is finished, is not a problem you can fix by re-exporting. You have to re-record.
Both of these checks take under a minute combined. Both of them prevent problems that are otherwise expensive — in time, and sometimes in content that simply can’t be recovered.
This is the final check, and it exists because it catches everything the previous four might have missed.
Record 20 seconds of yourself talking normally — read a sentence, describe your coffee order, anything. Then stop and actually listen back to it, on headphones, before you proceed to the real session.
This step catches the combination problems that individual checks miss: a cable that’s slightly loose and introducing intermittent crackle, a setting that looks right on screen but isn’t behaving correctly, an unexpected hum that only appears once actual signal is present, or a level that seemed fine on the meter but sounds wrong to the ear.
The value of this step is timing. If something’s off, you want to know now — while you’re alone, before your guest has arrived, before the interview starts, before thirty minutes of content has been recorded on a broken setup. Finding a problem after the session wraps and your guest has left is a different kind of problem entirely: one that often can’t be fixed, only apologised for.
20 seconds of test recording costs you almost nothing. Skipping it costs you the ability to catch a mistake while it’s still cheap to fix.
None of these checks are complicated. None of them require specialised knowledge or expensive tools. What they require is doing them — consistently, in order, before every session, regardless of how rushed you are or how many times you’ve recorded before.
In sequence:
The checklist exists because professionals in any field where mistakes are costly — aviation, surgery, broadcast — don’t rely on memory under pressure. They run the same sequence every time, specifically because the cost of skipping a step only becomes visible after it’s too late to fix cheaply.
There’s a limit to what a checklist can fix. If your room’s noise floor test reveals a persistent hum you can’t eliminate — a neighbour’s construction, a building’s HVAC system, traffic that doesn’t stop — no amount of gain staging or monitoring discipline solves that. The checklist tells you the problem exists. It doesn’t remove it.
A professionally treated recording space removes that variable before your checklist even starts. At Villo Studio in Canggu, Bali, the room is acoustically treated, the signal chain is configured and tested before you arrive, and the noise floor is controlled by design rather than by hoping the neighbours stay quiet.
Visit villostudio.com to book a session or request a production proposal.
