Here's the thing: when you're putting audio out into the world, loudness is everything. You've probably noticed it yourself. You're listening to a podcast that sounds quiet, so you crank the volume. Next episode hits and it blasts your ears. Then you switch to an audiobook that sounds compressed and weird. And you're constantly adjusting volume like a DJ between tracks. LUFS—Loudness Units relative to Full Scale—is the technical fix for this human annoyance. It's the measurement that ensures your audio sounds consistent everywhere it goes.
Understanding LUFS (And Why It's Not as Scary as It Sounds)
Let's demystify this. LUFS measures how loud audio actually sounds to human ears, not just what some meter says. That's the key difference that makes it useful.
Traditional audio measurements (like dB peak level) tell you the maximum volume of a signal. But here's where it gets weird: two tracks can have identical peak levels and sound completely different in loudness. One might feel natural and conversational. The other might feel crushed and fatiguing. Why? Because loudness perception depends on so much more than just the peak—it's about the dynamic range (the gap between quiet and loud moments), the frequency content, how the compression works. All of that stuff.
LUFS gets at what actually matters. Think of it like this: dB peak measures the tallest wave in the ocean. LUFS measures what the ocean feels like to swim in.
LUFS works because it weighs how human ears actually perceive different frequencies. Voices aren't just one frequency—they're complex waves. Your ear hears some parts louder than others, even at the same dB level. LUFS accounts for that. It also measures loudness over time, averaging out the variations, so you get a single number that correlates with "this is how loud this sounds."
When Spotify says they normalize to -14 LUFS, what they're really saying is: "We're adjusting all audio so it feels similarly loud to listeners." Your podcast doesn't need to come in at exactly -14 LUFS when you upload—Spotify will adjust it. But if you master to their target, your audio plays back unchanged. You're in control.
Different Platforms, Different Targets
So what should you actually aim for? It depends on where your content lives.
Spotify and YouTube both want -14 LUFS. That's the sweet spot for music and podcasts on those platforms. It's bright, energetic, competitive with commercial content.
Apple Podcasts asks for -16 LUFS. Slightly quieter than Spotify. This gives podcasts a more intimate feel and protects against over-compression artifacts.
Audiobooks are quieter still—-18 to -23 LUFS, depending on the platform. ACX (which distributes to Audible) wants -18 to -23 LUFS with strict peak requirements. Google Play Books and Kobo are more flexible, but generally -16 to -20 LUFS is safe.
Here's the practical advice: if you're distributing to multiple platforms, pick the strictest target on your list and master to that. So if you're doing both podcasts and audiobooks, target -18 LUFS and you'll pass compliance everywhere.
One last thing: ACX doesn't accept AI-narrated audiobooks right now. Just know that upfront if that's your distribution plan.
Why Does Loudness Matter?
Let's talk about what happens when you ignore this.
Too quiet, and your listeners are constantly turning their device up. That's friction. They might just stop listening instead. Too loud, and platforms automatically turn you down—which can introduce distortion or make you sound weird. Neither scenario is fun.
But here's what most people miss: consistent loudness is actually about respect for the listener. When someone's got you playing while they're running, or driving, or cooking breakfast, you don't want to be a volume knob they're constantly fiddling with. You want to just play at a consistent, pleasant level.
From a technical perspective, platforms normalize audio during playback anyway. But there's a difference between "your content passes through unchanged" and "your content gets adjusted." If you master to the platform target, you're in the first category. No processing artifacts. No quality loss.
There's also the question of dynamic range—that gap between the loudest and quietest moments. Wide dynamic range sounds natural and expressive. You can hear the variation in voice, the emotional nuance. Narrow dynamic range sounds consistent, easier to listen to casually, but can feel flat if you overdo it. Different content types want different approaches. A podcast with multiple speakers benefits from wider dynamic range. An audiobook being listened to in a noisy environment might want narrower range so quiet sections stay audible.
Building Your Mastering Chain
So how do you actually get there? Most professional mastering involves several steps, applied in order.
Start with EQ. This adjusts frequency balance. For voice, you're usually doing three things: cutting low rumble (high-pass filter below 80Hz to remove room noise and mic proximity stuff), adding a subtle presence boost around 2-5kHz to make words clearer, and de-emphasizing any harsh frequencies that hurt to listen to.
Then comes compression. This is where things get interesting. Compression turns down the loud parts of your audio. It doesn't touch quiet parts. So if you have a speaker who's dynamic—naturally gets louder when excited, quieter when thoughtful—compression smooths that out without destroying the expression. You want gentle compression for voice. A 2:1 or 4:1 ratio, medium attack and release times. Not aggressive. You're not trying to crush the life out of it.
De-essing is its own special thing. It's basically a compressor that only listens to high frequencies (around 5-7kHz). It catches harsh S sounds and SH sounds and turns them down. Some voices need this more than others. Bright microphones, certain voice types, scripts with a lot of sibilant-heavy words. If someone's saying "success" or "solution" repeatedly, de-essing is your friend.
Limiting comes next. This is a compressor with an extremely high ratio—basically a brick wall. It says "nothing shall pass -1 dB." This prevents digital distortion that can happen during format conversion or playback on certain devices.
Finally, loudness normalization. After all that processing, you measure your actual loudness and adjust to hit your target LUFS. This is the last step, the final volume adjustment that gets you to platform specifications.
The Process in Practice
Here's how you actually do this.
First, check where you're starting. Generate or record your audio and measure its LUFS. Most DAWs have loudness meters built in. Note the number.
Apply your mastering chain. If you're using Vois, the export presets handle all this automatically—you pick YouTube or Spotify or Apple Podcasts and the app does the work. If you're processing manually, apply EQ, then compression, then de-essing if needed, then limiting, then finally normalize to target.
Verify the results. Check that your LUFS is within 1 dB of target, true peak doesn't exceed -1 dB, and the audio doesn't sound over-processed. "Over-processed" usually means it sounds pumpy or like someone's sitting on it.
Listen to it. This is critical. Technical measurements matter, but they're not sufficient. Play your mastered audio on good headphones. Does it sound natural? Is there any distortion or weird pumping artifacts? Play some reference content from your target platform alongside it. Do they sound similarly loud? Do you feel like yours is compressed too much?
Don't skip this step. Numbers are useful. Ears are essential.
Mistakes People Make (And How to Avoid Them)
Over-compression is the most common one. People think "I need to master this" and go nuts with compression. They're trying to make it perfect. But compression removes expression. Too much and your voice sounds flat, tired, fatiguing to listen to. The fix: use gentle compression and accept that human speech has natural variation. That variation is good.
Ignoring true peak is the second one. You hit your LUFS target and think you're done. But peaks still go through the roof. Then when the listener's device converts formats or applies its own processing, distortion happens. True peak limiting is insurance. It's boring, invisible when it works right, but essential.
Picking the wrong target for your distribution is surprisingly common. You master to Spotify spec, then decide to submit to ACX for audiobook distribution. Now you're too loud. Either master to the strictest platform on your list, or be willing to create multiple masters.
Trusting numbers alone is tempting. You hit -14 LUFS and assume you're done. But technical compliance doesn't guarantee quality. Always listen critically. Compare to professional reference content from your platform. Does yours sound as good? Better? Worse? Your ear will catch what meters miss.
Vois Makes This Automatic
Here's the good news: you don't have to do any of this manually if you don't want to.
Vois includes export presets that handle mastering automatically. Pick YouTube, Spotify, Apple Podcasts, or Audiobook, and the app applies the full mastering chain and normalizes to the right LUFS. You don't have to think about compression ratios or de-esser thresholds. The export handles it. You get consistent, professional results every time.
But understanding what's happening under the hood? That's useful. Because you'll know why your audio sounds the way it does. You'll know how to tweak it if you want something different. And you won't be mystified by technical jargon when someone talks about loudness.
The bottom line: LUFS is just a way of measuring what you already know intuitively—how loud something actually sounds. Get it right, and your audio plays consistently everywhere. Your listeners don't adjust volume. Your content sits confidently alongside professional content on any platform. That's the whole game.