LUFS : Why you should check the rules for every platform. : Magroove Blog

LUFS : Why you should check the rules for every platform.

October 18, 2019 • 11 min read

I remember mastering was one of the most hardest things when I started producing electronic music. I would spend hours analyzing waveforms of famous songs trying to reach that same quality and volume. If you sympathize with this situation, you’ve probably also spent many hours using other music as a reference. But did you know that not all platforms work with the same volume and compression rate? Did you know that, ideally, each platform requires a different type of mastering? They do, and the standard to measure the ideal volume for each platform is called LUFS. Ultra-compressed waveforms don’t work in every platform and today we’ll understand why.

LUFS, mastering according to Spotify

LUFS level for each platform

Each platform works with a different value and it’s important to be aware of that while you’re mastering. Although some of them work with the same (or similar) numbers, make sure to check them before you finish mastering. Below is a list of a few platforms and the LUFS level each one recommends.

  • Apple Music: -16 LUFS
  • iTunes: -16 LUFS
  • YouTube: -13 LUFS
  • Spotify: -14 LUFS
  • Tidal: -14 LUFS

If send the same work to all the platforms, they may alter your music and you won’t even know it. Let’s take a deeper look into the some of the fundamentals before we continue.

What is LUFS

TV and radio platforms implemented LUFS around 2013 as a measurement to standardize the volume in audio tracks. LUFS stands for Loudness Units Relative to Full Scale. Most stores and online streaming services use it nowadays. In other words, LUFS is the main reference in the digital world.

At this point, you must be asking yourself what it is that LUFS measures exactly, right? The answer is simple: it measures volume using an algorithm resembling own hearing function, similar to the Fletcher Munson curves. This means that LUFS highlights mid and high frequencies above 2000 Hz, the most sensitive volume range for human ears.

The word “loudness” itself (the one used in the abbreviation) is quite vague, even though it refers to how loud the music or one of its components is. That’s exactly the reason for the creation of many objective ways to refer to loudness. One of those is the LUFS measurement, that follows the dBFS scale.

What is dBFS, RMS and Peak Volume

To understand LUFS and its use better, you should know the terms below. That’s why, in short, dBFS represents the scale, while we use LUFS as a measurement for audio. So LUFS simply represents the RMS level of a certain audio/song on the dBFS scale. As a quick example, let’s say a song is -10 LUFS; that means that the maximum RMS volume range that the song reached was -10dBFS.

dBFS

It stands for Decibels Relative to Full Scale. All in all, dBFS is a unit of measurement for amplitude levels in digital systems, such as Pulse-Code Modulation (PCM), which has a predetermined peak level.

RMS

RMS is a measurement used to define the average volume of the tracks. Root Mean Square (RMS) is usually used to define the magnitude of the average power emitted by a sound device. We won’t dive deeper into this concept during this article, but if you want to know more about it, visit this Wikipedia page.

Peak volume

Peak volume is different from loudness. For example, you can have high peak volume even if you don’t have a high RMS level. This is because peak volume uses whatever component of the mix reaches higher on the peak meter, while RMS calculates average power. For example, say you have drum sounds in your mix, and those sounds reach high peaks; even if there is this high peak volume, the loudness (or RMS) of your track can still be low.

The meaning of normalization and the measurement of LUFS

Moving on with our explanation, let’s now understand what normalization is and its role in all of this.

Generally, “normalizing” means turning the audio of a whole track up or down until its measurements (it can be the RMS average or peak) reach the desired level. For example, it’s common for vinyl or old CD recordings to not reach 0dBFS once they are transferred to a computer. That’s why the recording is normalized, so the highest peak reaches 0dBFS. Once this is done, the song doesn’t get compressed at all (no peak is compressed or cut out, which means the track is identical to what it was but louder).

Normalization can also be achieved through RMS, and this is where LUFS comes in. I analyze and determine what the RMS average in my whole track is. Usually, this average tends to be lower during calm moments and higher in the chorus. When you normalize through RMS, you see the moment this RMS average was the highest in the song nad normalize it right then. For example, if the average of the highest moment was -12dBFS, consider that song to be -12 LUFS.

The difference between LUFS, LKFS, and LU

A deeper understanding of LUFS is essential for music professionals. Nowadays, many platforms work with standards that must be followed. Otherwise, you’re at risk of having the quality of your material compromised on each platform, as mentioned earlier. But only understanding LUFS isn’t enough. Now that you know the basics, let’s understand the differences and in which way LUFS relates to LKFS and LU.

LKFS

It stands for Loudness, K-weighted, relative to full scale. The term itself seems complicated but its definition is quite simple. LKFS is a standard of volume created to allow the normalization of sound levels for open television broadcast and other types of videos.

LKFS – a synonym to LUFS – was introduced during EBU R128.  EBU R128 is a set of guidelines established to normalize the volume and maximum signal levels allowed in audios. The European Broadcasting Union initially recommended it in 2010; its revision only came in 2014. In spite of that, EBU R128 uses the LUFS measurement, compatible with international nomenclature conventions.

LKFS measurement is used by ITU BS.1770 and ATSC A/85, which, like EBU R128, are sets of guidelines for audio standardization and normalization. While ITU and ATSC recommend to level the tracks to -24LKFS, EBU changes that to -23.

We won’t be explaining in detail what each of these guidelines say. Otherwise, this article would be enormous. But we’re suggesting a few links for you guys to check out if you want (or need) to know more about this.

LU

LU stands for Loudness Unit. First, you have to understand that the LU is different from dB (decibels). Decibels measure the level of air pressure created by the sound. On the other hand, LU is a unit established for audio work, allowing the producer to control the output volume of the audio. In more practical terms, the LU is like the decibel for sound pressure created by the audio.

True Peaks and tools that help care for and avoid it

There are many plugins available on the market to help you identify and take care of problems such as true peaking. The image shows one such feature at work.

What “True Peak” is and how it interferes in audio finalization

Generally, the music industry doesn’t really have a widely accepted practice to standardize audio peak levels (True Peak). On the other hand, as we saw earlier, the TV and movie industries have very strict guidelines in regards to this. But what is True Peak and what does it hinder? I’ll try to explain it very simply.

First, we have to understand the origins of True Peak. For that, let’s take a look at the scenario below.

After finalizing the audio track in the DAW of your choice, you start the process of exporting it. In this process, your DAW converts all the archives into one single digital archive (usually wav). To make this archive playable in sound systems, you need to convert it analog form. Before conversion, the audio signal is similar to the drawing of a flight of stairs. (The digital recording doesn’t have infinite resolution, so the curve isn’t perfect; we’ll be unavoidably working with a “bit depth” and a “sample rate”).

A reconstruction filter is also applied during this process in order to “round the edges” of the audio signal. The goal for doing this is to get a smoother listening experience. However, these filters may only cause slight variations on audio levels. This can become a problem especially for sounds that are close to the 0dBFS threshold – they could try to surpass this benchmark and get distorted.

Some sound appliances specifically created for studios would neutralize this. However, “domestic” sound appliances can cause the sound to come out distorted. That’s why an audio track may sound good in a studio but distorted in other common devices.

What you should be careful about to avoid True Peaking

There are many ways to prevent true peaks. The peak meter native to most DAWs isn’t usually ideal for this. However, some tools can help monitor and prevent true peaks in your audios. A simple Google search can show you a few good results. My tip is to always research and test (if feasible) before acquiring any tool.

Excessive compression – why you should avoid this

Compressing an audio track means reducing its peaks and raising its volume as a whole while at it. Limiters, plugins that prevent the audio from surpassing a certain volume level (0 dBFS, for example), usually conduct this process. This occurs in order to raise the volume of the tracks since the volume connects to the RMS and not to its peaks. Unfortunately, the digital audio industry abuses compressors, rasing the volume of the whole song only to smash it in the 0 dB threshold. This process raises the volume as a whole, but maintains a fixed limit, consequently raising the average volume level.

But be careful about over-compressing. This can create a series of problems for your audio tracks, such as:

  • Lack of dynamic – The common listener may not realize it, but the song itself will be less powerful. So, listeners probably won’t enjoy the song and they won’t even know why!
  • Inertia or lifelessness – Relate to the issue above. Too much compression leads to a lack of dynamic. As a result, you get the impression that the song is “lifeless”. If you work with music, I imagine you know what I’m talking about.
  • Constant compression – Although exceptions exist, always make sure your gains come back to zero.
LUFS levels in different audio tracks

The screenshot shows the waveform of two audio tracks. How much you compress defines the LUFS level of each of your tracks.

Lufs, shuffle e a Loudness War

Many of you must recognize the word “shuffle”, right? Well, “shuffling” is only possible now thanks to the relocation of songs to the digital medium. If you belong to the age of vinyl, K7s, or even CDs, you’ll remember this function didn’t exist. With the arrival of the shuffle feature, you can listen to different songs from different artists with different finalizations (LUFS/RMS volume) one right after the other.

When comparing these two recordings, due to a psychoacoustic phenomenon, we perceive the recording with the higher RMS to be better than the one with the lower. In this struggle to have more memorable song than the next artist, producers started to compress their music more and more. This basically unleashed what we call Loudness War. A race to make your song the highest/most compressed it can be so it would sound “better”.

The light at the end of the tunnel

In an attempt to save the artistic side of music  lost with ultra compression, famous names showed their support for the lighter use of compressors during audio finalization. The Bobcats, creators of the “K” pattern (K-8/K-10/K-14), were among them. Be careful not to mix up this “K” with “K-weightlifting” (LKFS). The Bobcats’ “K” before a number represents average RMS levels of finalization. In a simplified way, the Bobcats scale pretends that 0dBFS is its reference value. For example, K-8 pretends that -8dBFS is 0dB. And, of course, you must work with all the restrictions of using 0dB pretending to be -8dBFS. This works to keep the mixes more dynamic.

To end the dispute, anti-loudness movements started to emerge – they even have a bit of leverage in the industry now. Nowadays, many platforms already have limits and loudness self-regulation for songs that go over a certain RMS average. Thus, the difference won’t be critical when shuffling the songs, the production won’t have to sacrifice valuable information just to raise the volume and the user’s experience won’t be impaired.

In conclusion

Infelizmente, a maioria dos produtores ainda trabalha sob essa concepção de que o loudness é mais importante que dinâmica. De forma que, quem acaba perdendo com essa “guerra” é o consumidor final do produto. Que este artigo sirva também como aviso e apelo aos produtores. A nossa visão sobre o que é música de qualidade está tão distorcida quanto a música em si, então é hora de darmos um basta. Esperamos que toda a informação contida nesse artigo ajude você a preparar suas músicas para as diferentes mídias. Mas mais importante, que você saia desta página focado em produzir músicas de qualidade, e não apenas com o volume alto.

Unfortunately, most producers still work thinking that loudness is more important than dynamic. The one who ends up losing in this “war” is the end consumer of the product. May this article also work as a warning and a plea to producers. Our vision of what quality music is is as distorted as the music itself, so it’s about time to say “enough”. We hope all information in this article helps you prepare your music for different media. What’s more important is that you leave this page focused on producing high-quality music, not just loud music.