Learn Music and Audio Production | iZotope Tips and Tutorials
Mastering Pros: How Loud Should My Master Be?
“How loud should my master be?”—this is the age-old question that has dominated every mastering discussion through the decades, and one that we’ve covered in great detail over the years here at iZotope. It’s a deceptively simple concept that can be as complex as music itself. To simply answer this question, there’s no ONE recipe for making all masters loud. It really depends on the music. But the good news is, you can constantly achieve optimal loudness through a healthy combination of critical listening, practice, and an understanding of the key factors that directly influence loudness in mastering.
In this article, we are joined by iZotope’s Director of Education Jonathan Wyner as he shares his insight on every facet of loudness in mastering. We will also be linking to previous iZotope mastering articles that allow you to explore every topic we cover in greater detail.
In this piece you'll learn:
Want to try out these concepts in your DAW?
Start your free trial of iZotope Music Production Suite Pro to access Ozone, Insight, and more mastering tools that can help you achieve your desired loudness and get the finished sound you’re looking for.
What is loudness in mastering?
Loudness is a concept that’s often misconstrued as a purely numerical value. When we think of loud sounds, we often associate it with high sound pressure levels—an aircraft taking off roughly measures 110 dB SPL, for example. But in reality, loudness in music has more to do with how sound is received by our ears and brain. It occurs in our heads based on the perception of how music hits us. Ultimately, loudness as we seek to assess it is not a concept that exists purely in a vacuum as a simple numerical value.
Especially in mastering, understanding loudness in this context of human perception can be powerful. It opens a lot more doors for the mastering engineer to employ creative strategies for making impactful, loud-sounding records that go beyond simply pushing the limiter. Let’s go explore several factors that can directly affect loudness at the mastering stage.
Learn more about defining loudness:
Factors that directly influence loudness in mastering
Monitoring & metering
Accurate monitoring and metering are fundamental to the mastering process. We rely on the accuracy of our listening environment and metering tools for us to make conscious, effective mastering decisions that best serve the music. This also ensures that the loudness we’re aiming for in mastering translates effectively into the marketplace and to the listeners’ ears.
Metering tools such as Insight 2 & Tonal Balance Control equip us with a better picture of the overall sound of our masters (frequency response, tonal balance, stereo imaging, loudness, etc.) Although metering is absolutely crucial, it can only be effective insofar as we allow our ears to make accurate, informed decisions through an optimal listening environment.
We’ve also discussed the importance of setting a fixed, calibrated playback for our monitors when mastering. Jonathan Wyner tells us more below:
“The whole point of accurate monitoring & metering is to be able to anticipate how a record is going to translate when it does get out into the world. So we have to set some kind of standard for ourselves and our listening environment. Another way of saying this is, if we, in our mastering studios, are constantly turning our monitor volume up and down, we lose our anchor and point of reference. If we set our monitor controller fixed to a singular/calibrated level, now we're understanding the sound and the way that it will translate to the end user.”
Learn more about monitoring & metering:
Timbre or distribution of energy across the frequency spectrum
Another crucial factor to the perception of loudness involves timbre, specifically the ears’ sensitivity to certain tone colors over others. Human ears are naturally more sensitive to mid-range frequencies compared to the low-end, as explained by the classic Fletcher-Munson curves study which eventually evolved to today's ISO-226 equal loudness contour standard. Understanding this phenomenon can help you manage more loudness out of your masters, but be careful about compromising the music by sacrificing too much of the low end. Critical listening & thoughtfulness are always key.
“This rolls back to perception. One of the things we learn very early in audio engineering practice is that we can't hear bass as well as we can hear mid-range. And by bass, I mean signals (grossly-speaking) below 150 Hz or 100 Hz compared to mid-range signals between 500 Hz and 3 kHz.
But if you look at a standard VU meter, the levels don't reveal anything about perception. Anybody can do this experiment at home. Play two tones at the same level / meter reading—one at 50 Hz, and another at 1000 Hz. Between the two tones, you can hear the 1000 Hz much more clearly. It sounds louder (this points to one of the valuable aspects of LUFS metering which DOES incorporate perception into the assessment of level).
This tells us something about how we experience sound and loudness. If you produce records that have a disproportionate amount of bass energy, by definition they probably won't sound as loud to the listener. It doesn't mean we don't love 808 kick drums. We want to make sure we feel the groove and get them in good proportion. But it says something about making sure that we are paying attention to the tonal balance, the relationship between the signals that we’re more sensitive to compared to those that we’re not.”
Learn more about timbre in mastering:
Dynamic contrast and its impact on loudness
Dynamic contrast tends to be underrated when it comes to aiming for loudness. Typically, there’s an inclination to reach for the compressor to tighten the dynamics and lift the whole production up. But there's something to be said about the impact one can experience from a sudden change in dynamics. We also wouldn’t want to compromise the natural build-up and climax in an epic production by squeezing it through a compressor. In many cases, gain-riding or volume automation may be more effective in preserving these nuanced dynamic contrasts in the mastering stage. Read more on what Jonathan Wyner has to say about the power of dynamic contrasts in making loud-sounding records.
“Here is where we really get into the meaningful definition of loudness that conveys emotion. The example I always give is, when you're a kid and your mother yells at you to get out of bed, she gets your attention because she is loud. When your mother whispers, “I love you,” it's tender and makes you feel something different. If she whispers “get out of bed,” you might not be motivated in quite the same way. But if she goes from that quiet sentence, and then she yells “get out of bed!”, it conveys a heightened change in emotional state.
Back to our master. So you go from a verse to a chorus. The chorus gets hotter and it makes you feel something different, right? It's that change in sound pressure level, that change in loudness that conveys emotion.
Part of the art of making really great records is figuring out how to incorporate that dynamic contrast into your work without having there be too much. There is such a thing as too much dynamic range where you can really disturb the listener. If they have to reach for a level control all of sudden when the chorus hits, that's a problem. But the opposite can be true. If you don't build in some sense of dynamic change, you lose an opportunity to really land the message. To really enhance the experience and realize the artistic vision on behalf of the artist.”
Learn more about dynamic contrast in mastering:
Peak level vs. RMS level / Crest factor
Speaking of having too much dynamics in your mix, the “crest factor” refers to the difference between your mix’s peak level versus average / RMS level. A higher crest factor is a good indication that your mix transients might be too hot compared to the rest of the sustained elements in the production. This is a good opportunity to ask yourself if the mix may have too much dynamic range and could benefit from some thoughtful transient control. Furthermore, just like the concept of loudness, “crest factor” is also very much context-based and shouldn’t be interpreted independently from the program material. As you’ll see in the next section, genre also plays an important role in interpreting this information.
“Crest factor is a term that expresses the difference between peak and average levels. A higher crest factor typically refers to a larger observable difference between peak and average level. For example, snare drum hits may peak at zero, but when the snare drum stops, the rest of the instruments could be playing at around -16 dB RMS. The difference between these levels is your peak-to-average ratio, your crest factor. The numbers do change a little bit, depending on whether you're looking at a single bar of a song, or if you're looking across an entire chorus. But that's the basic idea.”
Learn more about Crest Factor:
The perception of loudness is heavily dependent on the genre / program material. When you think about how loud a typical heavy metal album is versus, let’s say, a bluegrass record, you get a better sense of why loudness can’t simply be determined by one singular target value. However, every genre does have defining characteristics that serve as helpful reference points when we aim for loudness (crest factor, dynamic contrast, tonal balance, etc.). Jonathan Wyner tells us more below.
“If you think about different genres, the experience of dynamics varies from one genre to the next. In classical music, the difference between pianissimo, forte and fortissimo requires that if you're sitting in front of a speaker, there would be a significant change in SPL. Whereas in a pop record, if you're going from a verse to a chorus, the difference in dynamics might be conveyed by having more background vocals, an extra synth, and a bass drop or something similar. The way that dynamics gets expressed in one genre compared to another is very different. So we have to incorporate that understanding into our production and mastering decisions.
Let's take the idea of a classic rock record. The difference between its peak and average will typically be somewhere between 8-12 dB. I'm not talking about the intro of the track. I'm talking about when the track is really cooking and full-on. I think that peak-to-average relationship is a helpful guideline. From there, you can start to expand it into other genres. With a pop record, (this would be a gross overgeneralization) you’re probably looking at 2 dB less difference between peak and average.
Moving on to your typical RMS levels by genre, if you're looking at an acoustic jazz record, the levels will probably be sitting at around -14 dB RMS, with +-2db difference between peak and average. And if you look at classical music, your levels might be in the 18-20 dB RMS range. Again, these are gross generalizations but they're all helpful starting points. These are some of the ways that can help you evaluate where the RMS is going to sit when the record’s set to full scale.”
Learn more about genre in mastering:
Managing loudness before mastering
Finally, the burden of making loud records doesn’t simply fall on the mastering engineer’s shoulders. This is especially helpful to keep in mind if you are a mixer or producer who’s mastering your own sessions. There’s a lot that can be done prior to mastering that can set your music for success when it comes to loudness—from the songwriting and arranging down to the mixing stage. For example, you can take what you’ve learned about dynamic contrasts and incorporate it into how the song is written or arranged. Crest factor is something that can be controlled to greater effect while at the mixing stage rather than mastering. There are many other elements that can enhance your music’s overall loudness before moving on to the mastering stage.
Jonathan Wyner tells us more below about how you can better manage loudness while mixing.
“When you start mixing a record, it's easy to fall into the trap of building what we call a “creeping” mix. For example, you start by setting up the lead vocal and a single instrument in the intro. Another instrument eventually comes in and so you push the vocal up so it doesn’t get lost. Once you get past the intro and the first verse, the second verse starts to get a little hotter. Eventually, the vocals come up a little bit more and if you look at the vocal levels going from the very beginning until the third chorus, you start to see a difference of 10 dB if you solo that track. That's a recipe for a real challenge in mastering.
If you haven't managed the journey of the primary elements—whether it's the drums and the bass relative to your lead instrument, whatever it may be— you create problems that are very difficult to deal with effectively in mastering. Make sure that you’re managing your gainstaging as you mix. That way, when the time comes for mastering, the intended journey of the dynamics has already been set up pretty well. It needs to be built into the mix in order for it to be realized effectively in a sufficiently ‘loud’ master.”
Learn more about managing loudness before mastering:
Tips and best practices for louder, great-sounding masters
Now that we’ve explored some of the common factors that influence the perception of loudness in our masters, here are a few additional guidelines (along with Jonathan Wyner’s nuggets of audio wisdom) to help you execute a loud and productive mastering session.
Gain-matched A/B comparison between the mastered & unmastered audio
“The point of gain-matched A/B comparisons is to avoid loudness bias—to make sure that you are not being fooled in your decision-making purely by differences in sound pressure level coming out of the speakers. We want to accurately judge whether the added level is helping or hurting a track.
This is especially meaningful when we're comparing our starting point, the mix file, compared to the mastered file. If you got 6dB of gain in your limiter, then it's gonna be really hard to assess whether your EQ settings are working or not. That's why you set up your monitoring environment so that you can easily match the before and the after. At the end of the day, you have to make sure that the level going out is where you need and want it to be.”
Stay informed about streaming & loudness normalization standards
“I recommend people acquaint themselves with loudness normalization and streaming for two reasons:
1) NOT so you can use it as a target, but rather so you understand what different kinds of presentations exist in the world when music gets played to the consumer. And because loudness normalization is becoming more and more prevalent for the consumer, it opens up the possibilities of retaining more dynamic contrast.
2) Understand that streaming platforms’ various loudness normalization standards will turn the level down for every dB it goes above their standardized level. If we’re turning our level up at the mastering stage and everything sounds great, awesome. But if we start turning it up and it doesn't sound any better (or it sounds worse), then why are we doing it?
Furthermore, when people talk about targets, the only arena where that's helpful is in understanding what the consumer is going to get. Decouple the idea of integrated loudness from everything else that we actually want to be thinking about in music production. If you can separate those two ideas, you can do a better job mastering music to satisfy the goals of the artist and the needs of the audience.“
Understand that the whole playback experience is constantly evolving
“We need to keep paying attention to the constantly changing playback experience, because as it evolves, we have to change our practice. I certainly had to change what I do in response to playback technology and mastering tools over the last 35 or 40 years. I think anybody today, 5 years from now, is probably going to be confronted with a slightly different version of the same kind of challenges. So pay attention.
AES recently published new loudness recommendations. It's about the audio metadata and recommendations specifically for playback level for streaming, spoken word versus music, and also in the context of an album. It is worth acquainting yourself with, again so you understand something about the presentation to the audience but not necessarily to define your level setting practice in production.
But there's another piece to this. One of the reasons that the standards are the way they are now is because the manufacturers of some playback devices are not providing adequate gain to get the level to what consumers need. The tendency then is to want to push the level higher so that people can get louder playback.
Over time, the possibility is that playback devices will evolve so that it will allow listeners to play back to whatever level they want. And if we get to that point where streaming services are universally adopting loudness normalization standards and device manufacturers gainstage their outputs to accommodate the recommended level, then we might see a scenario where everything, music for picture and music only presentation is being calibrated down to --23 or -24 dB LKFS which is the Broadcast Loudness Standard for picture.
Things will change, things do change, and the aesthetics of music change in response. That's what's also so cool about it. The records people were making 30 years ago sound so different from the records people are making now partly because of the difference in playback technology.”
Loudness in the service of the music, not the other way around
The hard reality is, there’s not really one simple answer for making all masters sound loud. Music is more beautifully complex than one singular target number. But the good news is, you have many reference points that can effectively guide you towards creative choices that result in constantly loud, great-sounding masters that have impact.
Loudness also doesn’t just depend on the listener’s ears, but also on the ever-changing nature of the playback experience—from the playback devices and audio formats, to the streaming platform’s evolving normalization standards. It might seem overwhelming at first, but at the end of the day, it helps to remind ourselves that every decision we make as sound engineers is meant to service the music and artist’s vision, first and foremost. Everything else comes secondary.
“You don't only want your record to be loud. You want your record to be good and loud, or as good AND loud as you can make it. If the artist is happy and is represented well to the audience, that's what matters most. Do what you believe is best for the music and for the art.”