Broadcast and Streaming Loudness Specs for Commercials.

This is the single most common source of confusion, and it costs people real money in rejected deliveries. LKFS and LUFS are not two different scales. Both are the loudness unit defined by ITU-R BS.1770 — a K-weighted, gated average of the whole programme. ATSC, the North American broadcast standard, calls the unit LKFS. The EBU, in Europe, calls the same unit LUFS. One LU equals one dB.

So the difference between the American target of -24 and the European target of -23 is not a difference of standards or units. It is one decibel. That is small enough that people assume it does not matter and large enough that a compliance system will reject the file.

A delivery specification is a contractual requirement: hand in a file measuring this, or it is rejected. Playback normalisation is something the platform does after the fact — it measures whatever you uploaded and turns it up or down toward its own house target so the listener is not reaching for the volume between tracks.

The practical consequence is that mastering a streaming release louder than -14 LUFS buys nothing. The platform simply turns it back down, and everything sacrificed to get there — the dynamic range, the transients on the drums — is gone permanently while the loudness advantage is not. Nobody hears the extra level. They only hear the damage.

It also means the two kinds of number cannot be compared directly. Netflix asking for -27 LKFS dialogue-gated and Spotify normalising to -14 LUFS are not a contradiction, and neither is louder or quieter than the other in any meaningful sense. They are measuring different things for different purposes.

Integrated loudness is an average across the whole duration. That has a consequence people discover the hard way: a 15-second lift taken from a compliant 30-second master frequently does not measure the same, because the loud and quiet passages no longer sit in the same proportion. Cut the quiet opening off a spot and the average of what remains rises.

So every deliverable gets measured on its own — the 30, the 15, the 6, the vertical social cut, the radio edit. Measuring the master and assuming the children inherit it is the most common route to a spot being bounced back the day before air.

The same applies to a spot that is re-versioned for a different territory. A file that was correct at -23 LUFS for a European broadcaster is 1 LU hot for a US broadcaster, which is inside the plus-or-minus-2 tolerance and therefore usually passes — but stacked with any other drift it is what pushes a marginal file over.

Not all of the figures above have the same standing, and no other reference we could find says so. Two are published specifications that anyone can open and check: EBU R 128 states -23 LUFS and a -1 dBTP ceiling, and Spotify states -14 LUFS along with its -11 and -19 listener settings. ATSC publishes A/85 as the governing US standard — currently revision A/85:2026-07, approved 8 July 2026 — and -24 LKFS with a -2 dBTP ceiling is the settled reading of it across the industry.

The streaming and delivery figures for Apple Music, YouTube and Netflix are different. Apple Music and YouTube still publish no numeric loudness target: Apple's Digital Masters page states no LUFS figure and no YouTube help page states one, so the commonly quoted -16 LUFS and -14 LUFS are industry convention, reported consistently but not something you can cite to the platform. Netflix is the exception. Its sound-mix specification now sits on a public studio-partner help centre and, read on 17 October 2026, sets -27 LKFS ±2 LU, dialogue-gated, for home mixes — a delivery target, not playback normalisation.

The practical consequence is narrow but real: if a delivery is rejected on loudness, you can point a broadcaster at EBU R 128 or ATSC A/85 and settle it. For a streaming platform there is no public document to point at, so the argument is empirical — measure the file, measure what the platform served back, and compare.