when in doubt, just make your vocals louder

I mix loud for energy, then set vocal balance at conversation level before I call a mix done. If vocals sink when I drop the volume, they'll sink more in mastering and on phones. What I do now:

1. Mix at the volume that feels right - energy and emotional impact first
2. Check vocals at conversation level - before I consider a mix anywhere close to finished
3. Push vocals louder than my instincts say - accounting for Fletcher-Munson, limiting, and small speakers
4. Leave headroom for mastering - peaks at -3 to -6 dBFS; vocals roughly 3-6 LUFS below the full mix

I spent months convinced my vocal levels were perfect. They sat right where I wanted them in the mix, cutting through without dominating. Then I'd move into mastering and find that they're sounding... buried. Not destroyed, just pushed back in a way that made me question everything I thought I knew about mixing.

I kept second-guessing everything - my mastering approach, my monitoring setup. Maybe I was being too aggressive with the limiter. Maybe my room wasn't treated well enough and I wasn't hearing the mix accurately. It took way too long to realise the problem wasn't a gear issue. It was an ear issue - specifically, my mixing volume.

your ears are working against you

I can't judge final vocal level at my usual monitoring volume - I have to check at conversation level.

Here's what I didn't understand: human hearing doesn't respond the same way at different volumes. There's this phenomenon called the Fletcher-Munson curves that shows how our ear sensitivity changes dramatically based on playback level. At my typical mixing volumes (loud enough to feel the track but not concert-level) my ears responded fairly evenly across the frequency spectrum. Everything sounded balanced.

But at consumer playback levels - conversation-level volumes or quieter, which is where most people actually listen - our hearing becomes "bowl-shaped." Bass and treble drop off relative to the midrange. The problem for vocals is that while the presence range (2-5kHz) where vocals live stays relatively strong, all the supporting content that gives them fullness gets perceptually attenuated. The body below 500Hz and the air above 8kHz just disappear at lower volumes.

So vocals that sounded perfectly balanced at my mixing volume would sound thin and incomplete when my audience played them back quietly on their phones. I was making decisions at a volume that didn't represent how anyone would actually hear the final track.

mastering limiters prefer drums over vocals

Peak limiting reacts to drum transients, not sustained vocal energy - and that was riding my vocals down relative to the instrumental.

The second thing working against me: limiters are designed to catch transients. During mastering, when that limiter is working to maximize loudness, it's reacting primarily to drum hits and percussive attacks - fast, sharp peaks that need immediate control.

Vocals are sustained energy. They don't trigger the limiter the same way a snare hit does. What I started noticing in research was this pattern where vocals could get "ridden down" relative to the instrumental as the limiter focused on controlling drum peaks. The limiter catches a snare hit, then releases before the next transient - and vocals living in those inter-transient periods can experience uneven perceived levels.

I found a quote from mastering engineer Jack Ruston that made it click: "We need to be aware of anything that has high levels of sustained energy, as it can cause any bus compression or final limiting to suffocate the vocal." Peak limiting was literally working against vocal clarity while I was assuming mastering would just preserve whatever balance I'd created.

the data from hit songs doesn't lie

On commercial releases, vocals tend to sit in a tight, predictable range - roughly 3-6 LUFS below the full mix.

When I dug into actual commercial releases, the pattern was obvious. An analysis of the top 25 Spotify songs (by Mastering The Mix in 2023) showed that 21 of 25 tracks had vocal levels within ±1.5dB of each other. The average landed approximately 4.5 LUFS quieter than the overall track loudness - establishing a target range of 3-6 LUFS below the full mix.

That consistency isn't accidental. Professional engineers have converged on keeping vocals prominently placed because they understand these perceptual and technical factors. Andrew Scheps articulated the strategy that changed my approach: mix loud to establish emotional impact, then set vocal balance at low volumes. "You can't judge the vocal level properly when it's loud," he explains, "because the vocal will sink into the mix more when it's loud."

That aligned perfectly with what I'd learned about Fletcher-Munson. At high monitoring levels, instruments compete more effectively with vocals because the frequency response is flatter. Drop the volume, and you immediately hear whether your vocals will translate.

your listeners can't hear vocal fundamentals anyway

Most listeners aren't on full-range monitors - phone and tablet speakers roll off steeply in the low end, so vocal body doesn't translate the way it does in the room.

The final piece that convinced me to push vocals louder: most people aren't listening on studio monitors. Phone and tablet speakers roll off steeply in the low end - often somewhere around 100-200Hz, depending on the device - so the fundamental body of most vocals (roughly 100-300Hz) gets heavily reduced, not gone entirely, but enough that warmth and chest disappear. Your brain still reconstructs pitch from harmonics above that, which is why the vocal can feel present in the room and thin on a phone at the same time.

Listeners perceive vocals entirely through harmonic content in the presence range. They're not hearing the actual notes being sung - they're hearing the overtones and consonants that let their brain reconstruct the pitch. This makes vocals inherently more vulnerable to disappearing in a mix when played back on consumer systems.

I used to think saturation and harmonic enhancement were just trendy mixing tricks. Once I understood that vocal intelligibility lives almost entirely in the upper-mid and high frequencies - specifically the 1-8kHz range where consonants live - it made complete sense why adding upper-mid harmonic content through saturation helps vocals cut through on small speakers.

what i do now

I still mix at the volumes that feel right for establishing energy and emotional impact. But before I consider a mix anywhere close to finished, I drop my monitoring to conversation level and listen specifically to whether vocals maintain their presence. If they start sinking or sounding thin, I know they'll disappear even more at mastering and on consumer playback systems.

My mixing mindset now is almost 80% mastering in mind. I know all the potential problems that can arise with vocals during limiting and loudness maximization, so I'm actively solving those problems in the mix stage to save myself the headache later. It's way easier to push vocals up and add harmonic content during mixing than to try recovering buried vocals after mastering.

I also started leaving 3-6 dB of headroom with peaks at -3 to -6 dBFS going into mastering. This gives me room to work without fighting against a mix that's already been pushed to the ceiling. The more dynamic range I preserve, the less the limiter rides down those sustained vocal elements.

The shift in mindset was accepting that what sounds balanced to me at mixing volumes probably isn't balanced for how people will actually hear it. Fletcher-Munson curves, limiter behavior, and consumer playback systems all conspire against vocal prominence. The vocals need to be louder than my instincts initially tell me - not because I'm a bad mixer, but because I'm accounting for predictable physics that will work against them in the final master.

questions i still get about vocal levels and mastering

Why do my vocals sound buried after mastering when the mix felt perfect? Mixing volume, not bad gear. If vocals sink at conversation level, mastering will push them back further.

How loud should vocals be relative to the mix? Roughly 3-6 LUFS below the full mix on the top 25 hits. I check the number, but the quiet listen is what counts.

Why check vocal balance at low volume? Loud mixes lie - vocals sink into the balance. Scheps mixes loud for impact, sets vocal level quiet. Same habit.

How much headroom should I leave before mastering? Peaks at -3 to -6 dBFS, 3-6 dB overall. A slammed mix gives the limiter no room for sustained vocals.

Do phone speakers really change how vocals translate? Small speakers roll off around 100-200Hz - vocal fundamentals get reduced, harmonics carry the rest. Full on monitors, thin on a phone.

If you're finding your vocals keep getting buried after mastering, what worked for me was simple: making them louder than felt right at my mixing volume. The hit songs prove it, and honestly - it's one of those things that seems obvious once you understand what's actually happening between your studio monitors and someone's phone speaker.