![\"A](\"https:\/\/www.thevoicerealm.com\/blog\/wp-content\/uploads\/2026\/10\/lv-wave.jpg\")
\u00a9 The Voice Realm<\/p><\/div>\n
Louder was better once. It isn’t now<\/h2>\n
The instinct to push the level up comes from a good place and an old one. On analogue tape, a hotter recording really did mean less hiss, because the tape’s own noise was fixed and you wanted the music well above it. Plenty of advice written in that era is still circulating, and it tells you to get the meters bouncing as high as they’ll go.<\/p>\n
Hugh Robjohns, technical editor at Sound On Sound, answered a reader asking exactly how much headroom to leave when recording at 24 bit. His advice was to “treat -18dBFS as the equivalent of the 0VU mark on an analogue system’s meter, and that’s where the average signal level should hover most of the time.” Peaks, he said, “can be way over that, of course, typically kicking up to around -10dBFS or so.”<\/p>\n
The reason is that a modern 24 bit converter has far more quiet range than your room will ever let you use. Robjohns is blunt about what pushing the level actually buys you: “Recording ‘hot’, therefore, won’t improve the actual noise performance at all, and will just make it harder to mix against other tracks.” The noise you can hear in a home recording is the room and the preamp, not the converter, and no gain setting changes the room.<\/p>\n
He also makes a point that matters more to a voice actor than to a band. “A sensible headroom margin is essential when tracking, to avoid the risk of clipping and allow you to concentrate on capturing a great performance without panicking about the risk of ‘overs’.” You can’t act while watching a meter. Set the level so that you don’t have to.<\/p>\n
The numbers to aim for<\/h2>\n
Here’s the whole thing in two figures. Your voice, at its normal reading level, should sit around -18 dBFS on the meter, drifting up toward -12 on emphasis. Your loudest moments should peak around -6 dBFS, and you want a little space even above that. Nothing touches 0.<\/p>\n
The Audacity manual, describing its own recording meter, says the bars “remain green until the signal reaches -12 dB then merge to yellow as the signal approaches -6 dB (which is a good maximum signal level to aim for).” It adds why: “This is a sensible level to aim for when recording, because it gives you headroom to make edits affecting volume afterwards.” Red means you’ve gone past that, and in the manual’s words, “The red color warns that the signal is becoming too ‘hot’ (approaching 0 dB) and that clipping may occur.”<\/p>\n
Nick Messitte, writing for iZotope, describes gain staging as “the process of making sure the audio is set to an optimal level for the next processor in the chain in order to minimize noise and distortion.” Record too quietly, he says, and “you’d approach the noise floor, a blanket of constant, audible hiss.” Record too loud and you lose the buffer. “You need a buffer or safety zone that can accommodate transient spikes or loud moments without causing horrible distortion.”<\/p>\n
If your recording looks small on the screen at these levels, that’s correct. A healthy voice over waveform at -18 dBFS average fills perhaps a quarter of the track’s height. It looks timid next to the brick of a finished commercial, and that’s because the commercial has been through mastering. Robjohns again: “Once the mix is finished, the now redundant headroom can be removed, and that is a standard part of the mastering process.” The loudness comes later, on purpose, when you can control it.<\/p>\n
![\"Graphic](\"https:\/\/www.thevoicerealm.com\/blog\/wp-content\/uploads\/2026\/10\/lv-panel.png\")
\u00a9 The Voice Realm<\/p><\/div>\n
Set the gain in five minutes<\/h2>\n
Most home interfaces are the same in the ways that matter here, and the panel above shows which control does what. Only one knob changes what goes into the file, which is why “it sounds loud in my headphones” tells you nothing about your level.<\/p>\n
Open your recording software, arm a track, and get in your normal position at the mic, the one you’ll actually use. Read a paragraph of real copy at the energy the script needs. Watch the peak marker on the software’s meter, not the green blink on the interface. Turn the gain up until the peaks on an ordinary line sit around -12 to -10 dBFS. Then stop and read the loudest line in the script.<\/p>\n
If that line lands around -6, you’re done. If it goes higher, back the gain off a little and read it again. Don’t creep the gain back up afterwards because the quiet lines look low. They’re meant to. The average will take care of itself once the loudest moment is safe.<\/p>\n
Make a note of where the knob sits, with a pencil mark or a photo on your phone. Interfaces get knocked, cleaners dust desks, and nothing is more irritating than a batch of files recorded 8 dB apart because the knob moved on Tuesday.<\/p>\n
![\"Black](\"https:\/\/www.thevoicerealm.com\/blog\/wp-content\/uploads\/2026\/10\/lv-loud.jpg\")
\u00a9 The Voice Realm<\/p><\/div>\n
Read the loudest line first<\/h2>\n
The single most common way a home session ends up clipped is that the actor set the gain on a calm line and then got into the script. Scripts build. A character gets angry, a tagline wants punch, a laugh arrives that nobody wrote down. Every one of those is louder than the line you tested with, sometimes by 10 dB or more.<\/p>\n
So before you set anything, scan the script for the moment that’s going to be loudest and start there. If it’s an audiobook, that’s the shouting match in chapter nine, not the opening description of the weather. If it’s a commercial, it’s the tag. Set your gain against that line, with a bit of space to spare, and the rest of the read fits underneath it.<\/p>\n
Then stop worrying about it. Once the gain is set for the loudest line, the meter becomes background. Glance at it between takes, not during them.<\/p>\n
- \n
- Where you stand.<\/strong> Halving your distance to the mic adds roughly 6 dB. A new chair, a music stand moved an inch, a different script height, and yesterday’s gain is wrong.<\/li>\n
- How you feel.<\/strong> A cold takes energy out, a good night’s sleep puts it back, and a script you love comes out louder than one you don’t.<\/li>\n
- The job.<\/strong> A soft meditation read and a monster truck promo aren’t the same input. Set gain per session, not once for life.<\/li>\n
- Software updates.<\/strong> A driver or app update can reset an input to its default. If a file looks wrong, check the gain before you blame the mic.<\/li>\n<\/ol>\n<\/div>\n
When it clips anyway<\/h2>\n
It will happen. Something makes you laugh, or the dog barks and you shout, and the meter goes red. The question is what to do with the take.<\/p>\n
The honest answer is to record it again. Clipping cuts information off the top of the wave, and nothing puts it back. There are repair tools that redraw the missing peaks and they do a decent job on a single hit, but a clipped laugh across half a second never sounds right afterwards, and a client who listens on headphones will hear it. Messitte’s warning is worth keeping in your head: “A file exported at 24-bit that exceeds 0 dBFS will fossilize, petrify and otherwise immortalize that distortion going forward.”<\/p>\n
![\"Graphic](\"https:\/\/www.thevoicerealm.com\/blog\/wp-content\/uploads\/2026\/10\/lv-takelog.png\")
\u00a9 The Voice Realm<\/p><\/div>\n
Audacity, like most editors, keeps a record of the damage. Its clipping indicator, the manual says, “appears as soon as there are four or more consecutive samples in that channel of audio exceeding that maximum,” and once lit it stays lit for the rest of the session so you can’t miss it. Use that. If the indicator is on at the end of a take, find the spot and redo the line while you’re still warm and still in the same position. Five seconds now saves an email later.<\/p>\n
One more thing about the red light on the interface itself. It tells you the preamp has clipped, which is the worst kind, because it happened before the signal was even turned into numbers. Turning the track down in the software afterwards doesn’t help. The distortion is already in the file, just quieter.<\/p>\n
Too quiet has a cost too<\/h2>\n
Having spent most of this guide telling you to record lower than your instinct, a word in the other direction. Minus eighteen is a target, not a floor. A voice that averages -30 dBFS because somebody panicked about clipping is going to need 12 dB of gain in the edit, and every bit of room noise, fan noise and preamp hiss comes up with it.<\/p>\n
The fix for that isn’t more gain. It’s less distance and less noise. Sit a hand’s width from the mic, as we covered in the piece on finding the right spot in front of the microphone<\/a>, and your voice arrives louder for the same gain while the room stays where it was. Turn the computer fan away from you, shut the window, put the phone in another room. If you’re not sure how much of your recording is room, Voice Edit Suite’s free Home Studio Analysis<\/a> will measure the noise floor from a short sample, and that number tells you whether the problem is your level or your space.<\/p>\n
The sixteen bit figure from the textbook is useful here. “A 16-bit system provides a dynamic range of approximately 96 dB,” which is already far more than any home room offers, and you’re recording in 24. The maths is on your side. A voice averaging -18 dBFS over a room at -60 dBFS has a clean 42 dB between the two, and that’s plenty.<\/p>\n
Levels for delivery are a different job<\/h2>\n
Everything above is about recording. What you send to a client is measured differently, and it’s worth knowing the two never need to be the same number.<\/p>\n
Audible’s ACX platform publishes the clearest example. Its submission requirements say an audiobook file’s “volume is between -23dB and -18dB RMS”, that “peak levels are less than -3dB”, and that the “noise floor is less than -60dB RMS”. RMS is an average, close to what you’d call how loud it feels, where a peak is a single instant. A raw recording made at the levels in this guide will usually sit a touch quiet against that spec and be a doddle to bring up. A recording that clipped can’t be rescued to meet it at all.<\/p>\n
So the order is: record with headroom, edit, then normalise or limit to whatever the job asks for. That’s the last step in a sensible editing pass order<\/a>, not the first. If a client hasn’t told you a loudness target, the figures in our guide to the specs behind every voice over file<\/a> are a safe default. And if they have, hit theirs exactly, because a file that’s 2 dB louder than the one it sits next to in an edit gets noticed.<\/p>\n
![\"A](\"https:\/\/www.thevoicerealm.com\/blog\/wp-content\/uploads\/2026\/10\/lv-note.jpg\")
\u00a9 The Voice Realm<\/p><\/div>\n
A routine you can keep<\/h2>\n
None of this needs to take long once it’s a habit. Sit down, put the headphones on, check the gain knob is where your note says it is. Read the loudest line in the script and watch the peak. Adjust if the session calls for it. Record. Glance at the clip indicator at the end of each section. Fix anything red before you move on. Deliver at the level the client asked for, not the level you recorded at.<\/p>\n
That’s it. Minus eighteen on the average, minus six on the peaks, and the knob left alone until the next session. The meter stops being something you watch and goes back to being something you check, and the attention you were spending on it goes where it belongs, which is the read.<\/p>\n
Sources: Hugh Robjohns, Sound On Sound, “Q. How much headroom should I leave with 24-bit recording?”<\/a>; Jennifer Burg, Jason Romney and Eric Schwartz, Digital Sound & Music, “5.2.2 Input Levels, Output Levels, and Dynamic Range”<\/a>; Audacity Manual, “Meter Toolbars”<\/a>; Nick Messitte, iZotope, “Gain staging: what it is and how to do it”<\/a>; ACX, “What are the ACX Audio Submission Requirements?”<\/a>.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"
Where your gain knob should sit, what the meter colours mean, why peaking at minus six beats chasing zero, and what to do with a take that clipped.<\/p>\n","protected":false},"author":5,"featured_media":6536,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[95,504],"tags":[686,687,685,688,334,684,202],"class_list":["post-6535","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-voice-over-studio","category-voice-over-tips","tag-audio-interface","tag-clipping","tag-gain-staging","tag-headroom","tag-home-studio","tag-recording-levels","tag-voice-over-recording"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t
- How you feel.<\/strong> A cold takes energy out, a good night’s sleep puts it back, and a script you love comes out louder than one you don’t.<\/li>\n