What is LUFS and why should I care? - in line (2023)

The music industry's 'sound wars', with everyone wanting the HIGHEST RECORDING POSSIBLE, has resulted in a lot of collateral damage including distortion, listener fatigue and lack of dynamic range. But volume wars aren't just a part of our industry, as you'll know if you've been watching TV late at night and out of nowhere, Honest Bob's Mainly Run Used Cars commercial is ten times louder than the show that preceded it. 🇧🇷 Panicking, he grabs the remote to turn down the volume before the neighbors file a noise complaint, but then has to turn up the volume when the show resumes.

Well, the European Broadcasting Union (EBU) hated it as much as we did, so they decided to do something about it. But in doing so, they also made it easy to master a collection of songs or an album. That's how they did it.

On average, a maximum experience

There are two ways to measure audio levels: peak and RMS (root mean square). Peak measures the highest level produced by the audio (even incredibly short-duration peaks), while RMS quantizes the average level. RMS correlates most closely with our hearing. For example, a snare hit has a high peak, but a sustained power chord with its high average sounds higher subjectively, even if its peak isn't as high as the snare. When both have the same RMS level, they sound more similar in terms of loudness.

To match levels on an album, some people normalize all tracks to their maximum maximum. This is rarely effective, however, as a song will sound louder at a higher average level than the others. Others normalize to an RMS value, but they can have spikes above 0, so watch the overall levels as well. Ultimately, the best way to match levels is often subjective. One technique is to normalize all tracks to their maximum peak to take advantage of the available headroom, decide which one sounds the quietest, and then lower the levels of the loudest to match the quietest as closely as possible. This isn't a particularly scientific option, but with a lot of trial and error you can find a satisfying balance.

Enter LUFS

There is a lot of reference material on the web about the EBU R128 standard, how it was derived, the many variations and the nuts and bolts behind them, so let's focus on practical applications. Basically we now have volume meters that extend the capabilities of traditional VU or peak meters and are based on a specification called volume units. LUFS stands for Loudness Unit Full Scale, which refers to full loudness units (i.e. the maximum level a system can support).Steinberg Cubase,PreSonus Studio One Professional,Magix Samplitud,iZotope-Perspective, and other programs now include this type of measurement.

Simply put, loudness units are the unit of measurement used in quantifying the perceived loudness of a piece of music by looking at the average level over time. In theory, two pieces of music with identical LUFS values ​​should sound like they are on the same level, and in practice they sound like they are on the same level, regardless of what the peak or RMS numbers say . So we have an immediate, practical benefit: if you're dominating and want consistent levels across tracks, check your LUFS readings.

The readings are negative numbers, such as B. -5 LUFS, -10 LUFS, -13 LUFS etc. as they refer to full scale. The less negative the value, the higher the average value. But here's the beauty of the system: a broadcaster like YouTube can opt for a standard LUFS level, so people don't have to constantly fiddle with the volume and just adjust the overall audio to achieve that LUFS level. (Although some companies use proprietary methods to check levels, the results are more or less the same.) Of course you can still remove daylight from your audio and produce something that is -5 LUFS, but YouTube reduces it to -13 LUFS , to bring it to the same apparent level as everything else. Or if you have a jazz track with a wide dynamic range that registers -18 LUFS, your level will be increased to -13 LUFS. So this Belgian hardcore techno track from 1990 will at least have almost the same thingRoger thatsound like a recording of a string quartet.

For streaming, the recommended default is -23 LUFS, but that's different than mastering music for playback. We often make subjective statements about how much dynamic range we want, and we can dominate any LUFS we want. In fact, I don't think the idea of ​​starting with a goal of hitting a specific LUFS number is ideal; I prefer to master what sounds good to me without worrying about LUFS first. Sometimes it's a larger number, sometimes a smaller number; depends on the material. But once you have something that "sounds good," analyzing where you are on the LUFS scale makes it easy to fine-tune the levels of other tracks so they're consistent.

Also keep in mind that playback via a streaming service is different than a CD or download. With CDs you can master as loud as you want, but I think after -9 LUFS you've lost most of the dynamic range. For club tracks you might want to master at an even higher average level like -6 LUFS as there is no standard level for clubs like there is for broadcasters or music distribution systems like YouTube. But then again, when your track is aired, it will theoretically have the same perceived level as other music.

For what it's worth, I master my "rock meets EDM" music at around -11.4 LUFS, which seems like the sweet spot between dynamics and the ability to cut through background noise. Compared to a song cut at -6 LUFS (Fig. 1), mine sounds a lot smoother. But the combination of the -6 LUFS music with -11.4 LUFS not only makes the perceived volume the same, mine actually sounds a bit louder subjectively because it has more dynamics.

Of course, when both cuts are used to create audio for in-car playback (via USB stick or CD), the -11.4 LUFS will sound smoother. Luckily, thanks to modern technology, many car audio systems have a fantastic new feature called "volume control". This allows you to set the level as high or low as you like while preserving the dynamic range of the audio! Incredible, is not it?

a practical example

my newestProject "Joy of Life"(coming November 11th) is a rolling set, so having consistent levels from one cut to the next is very important. This is the workflow I use to do this with albums. Remember, it's not about mastering, it's about putting the album together: your music needs to sound the way you want it to.beforeYou start the assembly. I use the Waves L3 for adjusting perceived average level (it works great for that), not for mastering songs per se.

1. Mix any track into a stereo file without master bus compression or clipping.
2. Max each track to -1.5dB (before I go online and say I'm an idiot who doesn't know anything about audio for maxing all tracks, read on). If you have some headroom, you can adjust the levels later if needed.
3. Take each track to the Studio One project page to assemble the album.
4. Check the LUFS of the raw tracks. Mine usually gets around -17 to -23 LUFS, indicating that not everyone has the same perceived average.
5. insert aWaves L3 Multi-Maximizer(Fig. 2) on each track to allow consistent average level adjustment for each track.

6. Adjust the Boost value for each track so that the LUFS of the processed audio is -11.4 (again, this is a personal preference for this particular project, not a rule of thumb). Studio One Metering shows the LUFS before and after any processing, including level changes. After any kind of change, the measurement will prompt you to update the reading. Then I adjust the amount of Max, Refresh, Snap, Refresh, etc. until each track shows -11.4 LUFS.

The LUFS algorithm is more complicated than just putting a meter on something (it knows not to include a fade as part of what it is trying to measure, for example), but all of that computation is happening in the background. In the case of "Joie de Vivre", the end result after listening to the edited tracks was that they all had the same perceived average level: the tracks that required more maximum power to reach the same average level got what they needed, and those Tracks that I needed less, had less.

So we're done, right? Well no. Here's an analogy for all guitarists: Although a Plek machine can fix guitar frets and cut nuts with computerized precision, it still requires a human to review the results and make decisions about possible adjustments.

more measurement

The Studio One Professional measurement (Fig. 3) shows four other important parameters. They give you data that can prompt you to go back and make some changes to the track, levels, or max amount (if you use the same approach as me).

LRAis a dynamic range reading. The typical range is from 5 (very small dynamic range, such as a commercial or aggressive club mix) to 15 (a live acoustic recording). I am using this as a reality check of the LUFS reading. Most of my music has a dynamic range rating of between 11 and 8, which tends towards a somewhat limited dynamic range to make the music "jump" a little more. However, I did notice that one song got a 7, so I checked it out even though subjectively it seemed on par with the other tracks. It ended up being a dance mix without much inherent dynamic range, so I didn't make any changes, but it's useful to have that kind of feedback.

Ptmeans "true peak" and takes into account the distortion between the samples (i.e. the distortion that can occur when the signal is reconstructed through the smoothing filter of a DAC, even though a peak meter that measures sample levels shows no distortion) . Well, I'm not a fan of distortion, so I want the TP value to be -0.1 or less for both channels. While there were few tracks with high TP values ​​(e.g. +1.2dB or +0.6dB), the solution was to lower the track's level and readjust the boost amount to match the slightly increase the perceived level to compensate for the level.

effective valueThere's the old-school RMS gauge, which I still think is important. It's another way of measuring signal dynamics, so I'm looking for consistency across different areas. If something differs by more than a few dB, I go back to the original track to see if the difference is intentional or not.

direct currentshows the DC offset of the trace. It's usually -infinity, but sometimes I run into a small compensation, which isn't too bad. if there isesSignificant DC offset, this is important to know as it messes up track clearance and makes it difficult to properly balance levels. You can go back to the original track and remove the CC offset, which allows any clipping or maximizing to work more efficiently.

Do we now have to produce a million different professors?

Since different outputs are normalized to different standards (e.g. YouTube -13 LUFS vs. iTunes -16 LUFS), does that mean you have to create separate masters for each output? Well, decide for yourself. Increase or decrease the level of your track until you get the level you want (e.g. on my tracks -11.4 LUFS I turned the volume down so they hit -16 LUFS to see the difference heard when played in iTunes). At least to my ears, it didn't make enough of a difference to justify the effort. Also, if you decide to release your music through an aggregator, you'll have to create separate masters for CD, download, and streaming, which can be a logistical disaster. Just make the best music you can, with the dynamics you want, and let the chips fall as they reach the listener's ears. Let's hope your playback systems have this new onecontrol the volumefunction mentioned above.

Yeah that's really cool...

Purists might be annoyed that we add an algorithm to our assessments on a subjective level. but it's too muchEffectiveAlgorithm. It used to take hours to properly balance levels on an album. It now takes about 30 minutes including additional adjustments.

Granted, the album format isn't as popular as it used to be, but collections of songs are still common and the LUFS measurement can help with consistency. But perhaps most importantly, in today's single-minded world, LUFS promotes a truce in the volume wars: you can't crank up the maximizer and turn up your music on Spotify because Spotify will just turn it down again. And if you like music with dynamic range, Spotify makes it subjectively sound just as loud as anything else, just with more dynamic range. That's progress.