What Is LUFS? Audio Loudness for Video Explained (2026)

LUFS stands for "Loudness Units relative to Full Scale" — a standardized measurement of how loud audio actually sounds to a human ear, not just how high its peaks are. Streaming platforms normalize everything to a target loudness (YouTube, Spotify, TikTok, and Instagram all land around -14 LUFS; broadcast uses -23 to -24 LUFS). If you master your audio to the right LUFS target, your video sounds consistent and the platform won't turn it down. If you don't, your carefully-loud mix gets automatically lowered — or your quiet mix stays quiet.

Audio is where creators lose viewers fastest, and loudness is the most misunderstood part of audio. People crank their master to be "loud," upload it, and the platform quietly turns it back down — undoing the work. This guide demystifies LUFS: what it is, why it replaced the old way of measuring loudness, every platform's target, and how to actually hit it. The one-line version lives in the LUFS glossary entry.

What is LUFS?

LUFS (Loudness Units relative to Full Scale) is a unit that measures perceived loudness — how loud audio seems to human ears over time. It was standardized (in the ITU-R BS.1770 and EBU R128 specifications) specifically to solve a problem that older measurements couldn't: two clips with identical peak levels can sound wildly different in loudness.

You may also see LKFS (Loudness, K-weighted, relative to Full Scale). LUFS and LKFS are the same thing — different standards bodies, identical measurement. If a spec says -24 LKFS and another says -24 LUFS, they mean the same loudness.

LUFS values are negative numbers, because they're measured below full scale (0 dBFS, the digital maximum). So -14 LUFS is louder than -23 LUFS. The closer to zero, the louder. This trips people up constantly: a smaller negative number = louder.

The key word is perceived. LUFS uses "K-weighting" — a filter that approximates how the human ear responds to different frequencies — and measures loudness over time, so it reflects what you actually hear, not just an instantaneous electrical peak.

Why not just use dB or peak levels?

For decades, people measured audio by its peak level in dBFS (decibels relative to full scale) — the highest instantaneous sample. The problem: peak level tells you nothing about how loud something sounds.

Consider two examples:

• A quiet acoustic guitar with one sharp pluck can peak at 0 dBFS but sound quiet overall.
• A heavily compressed pop master that's loud the entire time can also peak at 0 dBFS but sound much louder.

Same peak, totally different perceived loudness. Peak measures the ceiling; it says nothing about the average energy your ears integrate over time. This gap fueled the infamous "loudness war" — decades of music and ads being compressed and limited to be as loud as possible, because "louder" grabbed attention, even though it crushed dynamics and caused listener fatigue.

LUFS ended the loudness war for streaming, because it lets platforms measure and normalize perceived loudness — so mastering something artificially loud no longer wins. It just gets turned down to the same target as everything else.

LUFS vs dBFS vs RMS vs true peak

These terms all describe "level," but each measures something different. Here's how they relate:

The two you care about in practice are LUFS (how loud it sounds — your loudness target) and true peak (dBTP — your safety ceiling to prevent distortion). A typical delivery spec is something like "-14 LUFS integrated, true peak no higher than -1 dBTP."

Integrated, short-term, and momentary LUFS

LUFS is measured over different time windows, and knowing which is which prevents confusion:

• Momentary LUFS — measured over a very short window (~400ms). Reflects instantaneous loudness.
• Short-term LUFS — measured over ~3 seconds. Useful for watching loudness move through a section.
• Integrated LUFS — measured over the entire clip or program. This is the number platforms use and the one you target.

When a spec says "-14 LUFS," it almost always means integrated LUFS — the average perceived loudness across the whole video. Your loudness meter will show all three; the integrated value is your headline number.

Two related metrics you'll see on a meter:

• LRA (Loudness Range) — how much the loudness varies across the program (dynamics). A film has high LRA (whispers and explosions); a podcast wants low LRA (consistent, always intelligible).
• True Peak (dBTP) — your peak ceiling, kept at -1 dBTP (or lower) so lossy encoding doesn't introduce distortion.

Platform loudness targets (2026)

Every major platform normalizes uploaded audio to a target loudness. Master near the target and your audio plays as intended; master far from it and it gets adjusted. Here are the widely-cited targets:

A few practical takeaways:

• For most web and social video, targeting -14 LUFS is a safe, near-universal default.
• For podcasts and spoken-word, -16 LUFS is the common standard (some use -19 for mono).
• Broadcast is much quieter (-23/-24) because TV prioritizes dynamic range and consistency across channels.
• These targets shift over time and vary slightly by source — treat them as strong guidance, not gospel, and re-check current platform docs for critical deliveries.

Why LUFS matters: loudness normalization

Here's the crucial part most creators miss: platforms turn your audio up or down to hit their target. This is called loudness normalization, and it's on by default almost everywhere.

• If you master your video to -9 LUFS (very loud) and upload to YouTube (target ~-14), YouTube turns it down ~5 dB. All your loudness-maximizing was undone — and worse, you kept the squashed dynamics with none of the loudness benefit.
• If you master to -23 LUFS (quiet) and upload, some platforms won't turn it up (many only turn down), so your video plays noticeably quieter than everyone else's — a real problem when viewers scroll from a loud clip to your quiet one.

So the goal is not "as loud as possible." The goal is the right loudness — close to the platform target — with peaks controlled. That gives you three wins: consistency (your videos match each other and everyone else's), no penalty (you're not turned down), and preserved dynamics (you didn't crush the mix chasing loudness that gets normalized away anyway).

Related audio concepts: ducking and dialogue leveling

Loudness work in a video edit isn't just one master number — it's keeping the dialogue consistently intelligible. Two related techniques:

• Ducking — automatically lowering the music (or game audio, or ambience) whenever someone speaks, so the voice always sits on top. You've heard this in every podcast intro where the music drops under the host. It's often done with keyframed volume or a sidechain compressor.
• Dialogue leveling — evening out a speaker whose volume wanders (leaning in and out of the mic), so every word is at a consistent level before you set the overall LUFS.

Get dialogue leveled and ducked first, then set your integrated LUFS target. Loudness is the final stage, not the fix for a messy mix.

How to mix to a target LUFS

Here's the reliable order of operations for hitting a loudness target without crushing your audio.

Step 1: Clean and level the dialogue first

Remove noise, even out volume swings, and make speech consistently intelligible. Loudness targeting comes after the mix is balanced, never before.

Step 2: Duck music and effects under speech

Lower music and background beds whenever dialogue is present, so the voice stays on top. Aim for music several dB below the dialogue during speech.

Step 3: Add a loudness meter

Use a LUFS meter (built into most editors and DAWs) and play the whole program to read the integrated LUFS. Don't trust your ears alone — monitor level fools everyone.

Step 4: Adjust overall level toward the target

Raise or lower the master until integrated LUFS sits near your target (e.g., -14 for social, -16 for podcasts). Small gain moves, not aggressive limiting.

Step 5: Control true peak with a limiter

Put a true-peak limiter on the master set to -1 dBTP so nothing distorts after encoding, while keeping your integrated loudness on target.

Step 6: Verify and export

Re-measure integrated LUFS and true peak on the final render. Use export presets that match the platform, and check the numbers before you publish.

Common loudness mistakes

• Mastering as loud as possible. Platforms normalize it back down; you lose dynamics for nothing.
• Confusing peak with loudness. Hitting 0 dBFS peaks doesn't mean you're at the right LUFS.
• Ignoring true peak. Peaks that look fine can distort after lossy encoding without a true-peak limiter.
• Targeting loudness before the mix is balanced. Fix dialogue and ducking first; loudness is the last step.
• Using the wrong target. Broadcast (-23) audio uploaded to social sounds too quiet; social (-14) audio on broadcast is too hot.
• Quiet dialogue under loud music. The most common intelligibility killer — duck the music.
• Not metering at all. Guessing by ear leads to wildly inconsistent videos.

How AI video editors handle loudness

Loudness is a perfect candidate for automation: it's measurable, standardized, and tedious to do by hand. Modern AI video editors handle the whole chain — leveling dialogue, ducking music under speech, and normalizing the final mix to a platform-appropriate LUFS target — without you opening a meter.

Because the editor already has a transcript and speaker detection, it knows exactly when someone is speaking, which makes accurate ducking and dialogue leveling straightforward: keep the voice consistent and on top, drop the music underneath, then set the overall loudness. When you export for multiple platforms, the loudness can be targeted per destination automatically.

In Loopdesk (disclosure: Loopdesk is our product), this is directed in plain English — "level my voice against the music and make it consistent" — and applied across the whole session, which is especially valuable when stitching clips recorded at different levels. It belongs to the sound-mix stage of the edit. The concept on this page is what the AI is optimizing toward: a consistent, platform-correct integrated LUFS with controlled true peaks. Knowing the target means you can tell it exactly what you want. LUFS is one entry in the full video editing terminology guide.

Frequently Asked Questions

What does LUFS mean?

LUFS stands for "Loudness Units relative to Full Scale." It's a standardized measurement of how loud audio actually sounds to human ears over time, unlike peak level which only measures the highest instantaneous sample. LUFS values are negative, and a smaller negative number means louder.

What LUFS should I use for YouTube?

YouTube normalizes to around -14 LUFS integrated, so mastering your video near -14 LUFS with true peaks at -1 dBTP is the safe target. Master much louder and YouTube turns it down; master much quieter and it may stay quiet.

What is the difference between LUFS and dB?

dB (specifically dBFS peak) measures the highest instantaneous level, while LUFS measures perceived loudness averaged over time using ear-weighting. Two clips can have the same peak dB but very different LUFS — that's exactly why LUFS was created.

What LUFS is best for podcasts?

The common standard for podcasts is -16 LUFS integrated for stereo (some use -19 LUFS for mono), with true peaks at -1 dBTP. This keeps spoken-word content consistent and intelligible across episodes and platforms.

Why does my audio get quieter after uploading?

Because of loudness normalization: the platform measures your integrated LUFS and turns it down to hit its target (e.g., ~-14 on YouTube). If you mastered louder than the target, the platform lowers it. Mastering near the target avoids the surprise.

What is true peak (dBTP)?

True peak measures peaks that occur between digital samples (inter-sample peaks), which can exceed the sample peaks and distort after lossy encoding. Keeping true peak at -1 dBTP with a true-peak limiter prevents that distortion while you hit your LUFS target.

Is louder always better for audio?

No. Because platforms normalize loudness, mastering as loud as possible gets turned back down — so you lose dynamic range for no loudness gain. The goal is the right loudness (near the platform target) with controlled peaks, not maximum loudness.

Can AI set the right loudness automatically?

Yes. AI editors can level dialogue, duck music under speech, and normalize the final mix to a platform-appropriate LUFS target automatically — often per platform on export — because they already know when each person is speaking from the transcript and speaker detection.

---

Want consistent, platform-ready audio without touching a loudness meter? Open Loopdesk: it levels your voice, ducks music under speech, and normalizes loudness for each platform automatically — just tell it what you want in plain English.