Understanding LUFS Loudness Standards for Spotify, YouTube and Apple
Peak level tells you almost nothing about how loud something sounds. A sparse acoustic recording and a dense electronic master can both peak at -0.1 dBFS while one sits 10 dB louder to the ear. Every streaming platform now normalises by perceived loudness rather than by peak, which is why a track mastered as loudly as possible no longer wins anything: it is simply turned down on arrival, with its dynamics already crushed for no benefit. LUFS is the unit that measures what you actually hear, and it is defined properly rather than by feel. The measurement here follows ITU-R BS.1770-4. The signal is K-weighted first, a filter that approximates the ear's greater sensitivity to midrange than to deep bass. It is then measured in overlapping 400 ms blocks. Then it is gated twice: an absolute gate discards anything below -70 LUFS, and a relative gate discards blocks more than 10 LU below the average of what survived. That gating is the part that makes the number meaningful, because without it a track with long quiet passages would measure far quieter than it sounds. Four numbers come out. Integrated loudness is the single figure for the whole programme, and it is what platforms compare against their target. Short-term maximum is the loudest 400 ms block, useful for spotting a section that jumps. Loudness range describes the spread between quiet and loud passages: a heavily compressed master has a small range, an orchestral recording a large one. True peak is the highest level between samples, not just at them, which is the value that matters because converters and lossy encoders can overshoot where no individual sample does. The targets are given as reference points rather than rules. -14 LUFS for Spotify and YouTube, -16 for Apple Music and for podcasts under the AES recommendation, -23 for EBU R128 broadcast. Selecting one lets you export with the gain applied to hit it, with a limiter in the chain in case the shift would push the true peak too high.
How to Use the Audio369 Online Loudness Meter (Step-by-Step)
Measure the final mix rather than a section. Integrated loudness is a figure for the whole programme, and a fragment will not represent it.
Integrated for the overall figure, short-term maximum for the loudest moment, loudness range for the spread, true peak for the real ceiling.
Each target is shown with the difference from your measurement, so you can see immediately whether you are above or below it and by how much.
Choosing one applies the gain needed to reach it. The panel warns you where the resulting true peak would land, and a limiter catches the overshoot.
Technical Architecture & Audio Engine Specifications
Where the platform targets sit
| Destination | Target | What happens if you are louder |
|---|---|---|
| Spotify and YouTube | ✓ -14 LUFS, peak -1 | Turned down on playback; your dynamics stay crushed |
| Apple Music | ✓ -16 LUFS, peak -1 | Turned down, with no benefit to being loud |
| Amazon Music | ✓ -14 LUFS, peak -2 | Normalised like the rest |
| Podcasts (AES) | ✓ -16 LUFS, peak -1 | Listeners keep reaching for the volume control |
| Broadcast (EBU R128) | ✓ -23 LUFS, peak -1 | Rejected outright by many broadcasters |
Who Uses Audio369 Loudness Meter? (Practical Creative Workflows)
Checking a master before release
Knowing you are at -9 LUFS against a -14 target tells you the platform will pull you down by 5 dB, which is worth knowing before rather than after.
Delivering a podcast to specification
Networks and directories increasingly specify -16 LUFS, and an episode that misses it either gets rejected or sounds wrong beside everything else.
Matching episodes to each other
A series where each instalment measures differently makes listeners adjust the volume every week. Measuring each one fixes that.
Meeting broadcast requirements
EBU R128 at -23 LUFS is enforced rather than suggested, and a measurement is the only way to know you comply.
Understanding your own material
Loudness range tells you how compressed a mix is, which is often more informative about what you have made than any peak reading.
Frequently Asked Questions (FAQ)
What is the difference between LUFS and dBFS? +
dBFS measures the level of individual samples against digital full scale. LUFS measures perceived loudness over time, weighted for how the ear responds. Two files at the same dBFS peak can be very different in LUFS, which is why platforms use the latter.
Should I master as loud as possible? +
No. Every major platform normalises to a target, so a louder file is simply turned down on arrival. All you gain is the damage that heavy limiting did to your dynamics on the way.
What is true peak, and why is it different from my peak meter? +
True peak estimates the level of the waveform between samples, which can exceed the highest sample value. That overshoot is what clips in a converter or after lossy encoding, which is why targets specify it separately.
Why does my quiet track measure louder than I expected? +
Because of the gating. Silence and very quiet passages are excluded from the integrated figure, so the number reflects the parts people actually listen to rather than being dragged down by pauses.
Is this as accurate as a studio meter? +
It implements the same standard: K-weighting, 400 ms blocks, and both gates. A studio meter adds real-time display, more history and calibration options, but the integrated figure is computed the same way.
What integrated LUFS target should I aim for on Spotify and YouTube? +
Spotify targets -14 LUFS with a -1.0 dB True Peak ceiling. YouTube normalizes audio to -14 LUFS as well. Mastering louder than -14 LUFS causes streaming services to turn down your track automatically.
What are the loudness standards for broadcast television and radio? +
Broadcast television follows the EBU R128 standard in Europe (-23 LUFS +-0.5) and the ATSC A/85 Calm Act in the United States (-24 LUFS +-1.0) to prevent startling volume jumps between shows and commercials.
How does K-weighting filtering accurately reflect human hearing perception? +
The ITU-R BS.1770-4 standard applies pre-filtering and RLB weighting curves that de-emphasize bass below 100 Hz and boost 2 kHz to 4 kHz presence, replicating the human ear's equal-loudness sensitivity.