Loudness in Audio Production: Best Practices

1. Introduction: Hi, my name is Will Edwards and welcome to this course all about loudness in audio production. So I have experience as a performer doing singer-songwriter stuff, going into studios recording. But I also have it on the other side where for many years I ran a professional commercial recording studio. I did a lot of engineering for bands, for soloists, even instrumental cues and that sort of thing. I've done music production for licensing. And then I've also done in production for artists where I'm looking at doing arrangement orchestration for them, things like that. So in this course, I'm going to talk about loudness. Specifically, this is kind of a nuanced area that you really need to know about if you're producing music, whether you're going electronic musician, you recruit, we are producing tracks to put on SoundCloud or Spotify. If you're a commercial producer or somebody just working in a home studio and you're wondering how loudness figures into your production cycle, right? Maybe you're curious about the difference between peak and true peak. What is RMS? What's dBFS? Why is my audio clipping, that sort of stuff? Understanding how to produce music, final product that's loud enough to hear, but also professionally balanced, right? Professionally normalized, so that it comes out of these different platforms sounding like professional recording. That's what we're going to talk about here. We're gonna talk about things like RMS, peak and peak. We're going to talk about crest factor, which is really important sort of way to measure and understand the loudest you record it. We talked about lumps and the loudness normalization standards that you'll find occurring in different platforms like Spotify or international broadcasts. This course is going to be perfect for anybody who is at home as a studio engineer, somebody who is a music producer, producing tracks and they need to to kinda do the mastering themselves. This course is going to cover everything you need to understand the principles and you're going to find a product, whatever software you're using, you have the tools that are required. You don't have to necessarily have the exact tools I use in these lessons. These are universal tools that come on most software packages, things like compressors, EQ, multiband compressors, some kind of metering, right? Limiting, that's sort of things, maximizers. So these are tools that a lot of people have. You probably already have them. So we're gonna get started here with some lessons that are going to introduce the principles. And if throughout the course, at any point you have questions, you can post them in the comments, or you can post to the discussion board forum message me directly. So let's get started. 2. Absolute vs. Perceived: So let's talk about absolute loudness. All right, this is an important concept to be aware of because loudness is a very subjective thing, and that's part of what makes it so tricky for audio producers to figure it out. So I want to do just a quick example here. I've got a couple of tracks. And they basically are going to demonstrate that although they have different, very, very, very different sort of levels in terms of dB, they sound very much like they are as loud as one another. So if we start with this example here, this is just a very low pitch. And if we kind of analyze it a little bit, we can see that this is coming in around 64 hertz. Okay, so c1, 64 hertz, very low pitch. And then if we were to look at the level on this channel, we can see that it's coming in at minus 6.9, so minus 70 dB. Okay, So that's relatively loud in terms of the metering, minus 70 dB is pretty close to 0 dBFS, which is pretty loud. Now, if we compare the loudness of this signal with this higher pitch here. Alright, and we just listened to this pitch here. We see that this one is actually coming in at minus 41 dB. So there's a difference here between minus seven and minus 41. So there's a huge, huge difference in the absolute loudness. And the difference is that we, as human beings, we don't hear low frequencies quite as effectively, efficiently as we hear higher frequencies. So the higher frequency here, which is, let's see what that comes in and as we play it. So this is coming in around 1k. So the sound of a one k pitch, and this is just a sine, sine wave being produced by a synthesizer. A one kilohertz pitch at minus 41 feels kind of like it's about as loud as a 64 hertz pitch at minus 70 dB. So huge, huge difference in terms of its metered loudness, but no difference or very little difference in its perceived loudness. So we want to separate out these two ideas of absolute loudness and perceived loudness. So we want to start measuring this stuff, right? That's what we're gonna get into in the next lesson where we talk about measuring loudness. 3. Measuring Loudness: So in the last lesson, we could see in here that absolute loudness isn't really a good way to measure what we perceive as loudness, because base frequencies we don't hear as well as mid-range frequencies. So another words, if, if things are perceived to be loud, it could just be that the mid-range is loud and we're missing all the base, right? It could be that the base is super loud and it's actually distorting and ruining our speakers. But we're just not hearing it that well, all we're hearing is the mid-range. So we want to have ways to measure loudness around. Sure. What's the absolute loudness? Because that's an important measurement. But in terms of the loudness we're talking about in this section, we're really talking about how to make an audio file sound and feel loud. We want to measure it in terms of perceived loudness. And the unit that we use for that is usually called L UFS or lumps. And it stands for loudness unit, full-scale. And what we wanna do is learn how this unit of measurement relates to average level, timber, and dynamic range. So this loudest unit, full-scale, kind of measures how we perceive loudness and we'll see that play out throughout the next several lessons. But I want to touch on a few different factors that we perceive differently. Okay? So there's average level which is, okay, let's say we're listening to a full band. The sound of the snare hitting the snare drum being hit each time has a much higher peak value. Then the vocalist singing the range between the quietest vocal and the loudest vocal is less than the range between, say, the loudest snare hit and the quietest kick drum hit. But our ears as biological entities for listening the sound, they actually sort of average this stuff out and we perceive sound somewhat in an average level. We don't really perceive the peaks so much. Then there's timbre, which is the frequency range that we're hearing. Are we hearing low frequencies, mid-frequencies, high frequencies. The human ear is much more sensitive to mid-range frequencies. And that will play into our perceived loudness if the music's got a lot of mid-range frequency, instruments like human voices, guitars, synthesizers, things like that. It can be perceived to be louder than if it's a rumblings set of drums. Low frequency content, right? Then there's a third factor which is dynamic range. And this is important for our loudness because we do kind of perceive the, the proximity that we think we are to a sound is somewhat dependent upon dynamic range. And that is what's the difference between the quietest and the loudest sound that we're hearing. If we're further away from things generally quiet things are extremely quiet and loud things are loud, but not that loud. It were right up next to things. If you put your head down next to a drum kit, It's like the dynamic range between the quietest hit and the loudest hit is really hard to distinguish. So the difference between loudest and quiet as things does, that's what we call dynamic range. And that does factor into our, what we perceive as loudness. So we want to try to maintain two different factors throughout all of this punch and clarity. Clarity is something that's very much tied into a timber. And dynamic range. Does the dynamic range fit what we expect to hear? And is the timbre, mid-range or upper mid-range frequency that we hear very well. Those things have to do with how we interpret clarity. And punch as well is how quickly does a transient occur, how quickly does something get loud and how quickly does it get quiet? How sharp is a transient that punch? And in the next lesson we're gonna talk about how punch and clarity are something you want to manage. How they are, they are very desirable and some of the things that we want to consider when we're trying to cultivate and maintain and develop punch and clarity even as we bring our loudness up. That's in the next lesson. 4. Punch & Clarity: Now I'm going to play this example here. And I want to talk a little bit of how hunch clarity. So I think as I play this, you can probably hear that there's quite a lot of clarity in the drums. You can hear the kick, you can hear the snare, you can hear the symbols in the high hat clearly. All right, and there's a certain amount of punch. Now the reason for that is that there's quite a bit of dynamic range in the drums. So the quietest to the loudest is something that we're noticing subconsciously. Okay. Things are clear because there's a lot of sound in the temporal range for the frequency range that our ears are very sensitive to, which is mid-range and upper mid-range. Even the kick drum, you'll notice that what you really perceive about the kick is actually an upper, mid-range or mid-range tone in the kick. It's the sound of the beater hitting the kick drum, not, not the low-end wolf of the kick. That's not what you hear so much as the beater. So we want to maintain this punch and clarity as we bring up the volume. And we're gonna go through a number of ways that we can do that. But thinking about how much punches there are all the transients. Maybe it's, maybe it's not a drum transient. Maybe it's the transient of a plucked guitar, nylon string guitar, or something like that. That has a, has a very distinct punch to it, right? It's got a very distinct transient character. We want to maintain those transient characters. Because if we squash them too much as we'll find out later, It's very detrimental to the impact of our music. But we also want this clarity where there's the ability for the listener to separate what they're hearing. Separate different instruments. If things are squash too much with limiters and compressors. If everything's too loud all the time, then it's very hard for our ears to really distinguish between different sounds. This will happen at a loud concert. If you go to a loud concert and the volume is incredibly loud, It's very hard at that level to really differentiate between guitars, bass, and drums. I mean, you just, you kind of hear it and you kinda hear feel at all but that allowed or something is at a certain point we stopped being able to determine clarity. Now in the next lesson, I'm going to introduce you to a concept called crest factor. And this is going to be a really important way to start measuring both punch and clarity. While also starting to measure there are perceived loudness that's coming up in the next lesson. 5. Intro to Crest Factor: So I'm going to go ahead and throw a copy of ozone, which is a Mastering plug-in. Mostly what I want to use this for here is to demonstrate many of the concepts. If you don't have ozone, you can get a h1 elements version of it. I'm pretty inexpensively. If you have Ozone 9 like I do, then you got incredible loudness tools. And I want to go through this, not really as a course on how to use ozone as much as it's a great tool for demonstrating the principles that we're going to be discussing in this first discussion here is about crest factor. In crest factor goes back to two fundamental measurements. First of all, peak and what is known as RMS. So RMS, which stands for root-mean-square, really, you can think of it as average loudness while peak is the loudest thing that has ever happened. So if I, if I have this loaded up and I listen to I'm going to see over here, right here in red. I'm going to see that up here is my peaks, and this is left and peak, right? And then here I'm seeing my loudness, my lumps, loudness units full-scale. And the left is kind of giving me an average. Okay, It's, it's not exactly the same as RMS, but it's kinda like RMS. And when I play this, I can see that my peak values kinda change. You'll notice that the peak values get when a new high as achieved. Whereas the integrated lumps here, this average, it's sort of will go up and down depending on the music of this plane because it's constantly re-evaluating what that average is. Now if I want true RMS, then I can change that. I can actually change the settings up here in my, in my IO. And I can decide that I want the meter type to be telling me RMS. Now we can see RMS for the left and right. So what telling me is that in this audio that I've played since I started, since I hit Play, the loudest peak in my left and right channels are minus 12 and minus 14. And then I can see that the average, the RMS, is a little different in each of the channels, but it's roughly, well as when I stop the average goes down. So I have to keep it playing and you can see it kinda hold there for a little bit. That's the RMS. And the difference between these two numbers, we call that crest factor. So the peak is the loudest thing. And often what we really want to use peak for is to protect our equipment. If the peak is too loud, it might blow a speaker, which we certainly don't want to be making music for people to listen to that's going to blow their speakers. And then RMS is kinda giving us a sense of this perceived loudness. Not quite as well as the Integrated loves will give us, but lefts don't really come into crest factor here. We want to know what the peak is and then what our RMSE is. Eigen come into my main, main sort of DAW metering here. And I can see RMS and peak. And as I play this, I can see these numbers will hold. So RMS max is minus 14.4, peek max is so far 0. So my peak is never going over 0, but my RMS is minus 14. That means the crest factor is 14.4. It's the difference between the peak and RMS. That's what my crest factor is here. 14. So 14 is going to tell me quite a lot about my mission, my mix, and the stems that I've been imported here. And there are a few considerations that we now want to go through in the next several steps. In the next several lessons, we're going to mix a method for maximizing our loudness with things you want to consider as you go. So the simplest thing is to look at this peak max and to gain things up or down. And we're going to be doing that in the next lesson. 6. Simple Gain: So when I play this audio, might peek max is already 0. So I can't really gain this up a lot. But let's say that my peak max was a lot less than that. If I were to take my mix here and I wanted to bring down all of the different audio sources. I can actually link these. And I'm just going to bring them all down about like that. Now what you're going to see is that when I play the audio back now, my peak max is a lot quieter. It's no longer 0, it's coming at minus 11 switching. So this points to an important tip for kind of best practice when you are using metering and you're really looking at what loudness let your audio play for awhile. See how these numbers actually fall out over the long-term, right? So I've got minus 7.8, so I can definitely just getting things up right, That's the simplest kind of loudness change we can make. So if I go ahead and I bring my faders up, and I'm looking at this peak max value down here in the bottom right-hand corner. So I'm just going to bring this up until my peak max. Coming in at 0 seconds. See that? Yes, I didn't bring it up a little bit. Let's say we want to bring it up to 0. I give myself a little bit of headroom there and maybe, maybe half a dB or one dB, something like that. So that is the simplest thing we can do, simple gain. Seeing that, you know, there's, there's margin between 0 dBFS, dB full-scale, which is the max before you start getting distorted audio and, and your peak. So if your peak is coming in at minus anything, that gives you a sense that you've got some margin that you can just use the levels just an or boost gains and actually just bringing things up as high as you can. We're not doing any compression or limiting yet. But that's the simplest kind of gain that you can do now, to get more loudness out of this, we definitely want to, and we're going to need to use limiting and compression and RMS. So that's coming up in the next lesson. 7. Liming & RMS: Let's talk about limiting and RMS. Now. What I'm gonna do here is once again, bring up ozone. And I'm going to be looking at the maximizer module here. And if you have ozone advanced, you can actually use these modules individually on your tracks. This maximizer is very intuitively laid out and it's going to help us understand how limiting and RMS are related for bringing up our game. Okay, so I've got my metering here in ozone set to peak and RMS. When I play this audio, I can tell here that there's about four to six dB difference here between the peaks and the RMS. Alright? What that tells me is that my crest factor is roughly six dB, which is really actually pretty significantly squashed. Nonetheless, are RMS is a bit low, we want to try to aim at getting that RMS somewhere between say, minus 12 and minus nine. Right now it's the output is coming out. Minus 13, minus 15 sometimes, right? So what I'm gonna do is I'm actually going to manipulate my threshold over here and bring this down a little bit. And what I'm going to see is that there's going to be indicators up here of gain reduction. See it when I bring this down. Now in the lower right-hand corner here where it says RMS max. We're going to see that number go up as I bring this fader here down. So this is my limit or this my maximizer. And as I bring this threshold down, it's going to bring my RMS up. And I want to just bring this up more or less as somewhere between say, minus 12 and minus 9. And this will just, this isn't going to change the sound of it, the audio too much. But I want to bring up the RMS so that our perceived average value now that we've managed peak, the perceived value or loudness is sort of more in the range of commercial audio. Okay? So I'm going to bring this down and actually bring this over here so that we can kinda have these zoomed into the same space case. We're going to watch this RMS max number is I bring this fader down. I try and bring this somewhere like that. Minus 11s, great. Doesn't matter if it goes a little hotter than that. And I wanted to be much more than a and back it off a little bit. So setting the threshold on the master limiter to sort of target and RMS between roughly minus 12 and minus 9. Depending on your genre is a good starting point. So depending on your genre is an important caveat. If you're doing EDM, it might make sense that your RMS is more like minus six, minus seven, minus eight. Whereas if you are doing symphonic or sort of coral type stuff, well maybe then your RMS would be more like minus 12. You might even get to very extreme situations where certain kinds of EDM would have an RMS of minus 3 in the final outcome. While some kinds of very dynamic music seemed very quiet and loud music might even have an RMS of minus 14 or something like that. But generally minus 12 to minus nine is a good, a good target. Okay, so now I'm just going to do a quick bypass of this and again match. This is one of the things I love about ozone is that as I bypass this, instead of it getting quieter like that, I can have the gain match. So this is what and sounds like with the limiter. And this is what it sounds like without the lemma. And we can see that in the meters here that the output meter is definitely hotter than the input meters. So we've definitely gotten a little more loudness out of it. And we can see now that the peaks minus the RMS were getting in the zone of, you know, minus 5, minus 6 still. But we brought our RMS up. And that's kind of the first step you want to do with limiting. Using something with gain, match and bypass like that is a great way to make sure with your ears that you're not fooling yourself, you're not making any mistakes as far as distortion, you can definitely overdo this. And in the next lesson, we're going to look at what that looks like if you kinda squash your audio too much. 8. Avoid Squashing: Now we can take this mix, we can crush it. And you can hear that it starts to get distorted. Right? Of course, we're, we're running our RMS and our peak right up to the maximum the air and keep you see it on the meter. You might think. Some, somebody on the street, layman's attitude towards loudness, it would be, it would be rather intuitive to say, well, we want things to be as loud as they can be. But you can hear that when you really crush this limit. What sounds bad to us is that we are undermining the sounds that we're supposed to here. There's no longer any punch, there's no longer any clarity. It's harder to differentiate the drums from the pianos, from the voices. Things are getting muddy. And so, yeah, there's a lot of perceived loudness, but we haven't maintained that punch and clarity and that's a big problem. That crest factor in this situation is just too small, right? So in order to really sense this, it really helps to do a b or you can bypass. The difference between heavy limiting and no limiting. Citizen they're limiting. Is such a significant volume difference that you AB that and you hear the volume difference. It's hard to tell. She liked better because, you know, I mean, obviously in this situation is it's overly processed, it's squashed. But even with more subtle changes, volume is something that we like. So when we hear something louder, it just kind of automatically sound better to us. We need to be able to hear at the same level how the changes that we're making, our impact in the audio. So you want to have a tool that allows you to do this kind of gain match thing. So now if we listened to the original with limited, we can tell that the limited version that it's losing that punch, It's losing that clarity. So you can definitely overdo this. That's called squashing. And it's just simple. Put a maximizer on there and bring it down to a monstrous amount, get huge amounts of gain reduction. But you don't wanna do that. You want to be maintaining that punch and clarity while also managing your crest factor, the difference between the peak and the RMS, OK, and doing having an AB with gain match, await it to a B, your before and after without changing the volume. That is going to be key to making good judgments. So peak level is something that we've discussed earlier, but we want to go back to that and talk about best practices for peak level and give me, I want to give you a sense of some specifics that you could target when you are doing mixes. So you can, you can get a good starting point that's coming up in the next lesson. 9. Peak Level Best Practices: So some best practices will depend on your target for this particular master, this loudness project. The target here, I'm not talking about normalization target or a loudness target in terms of dB, I'm talking about a target in terms of distribution. Is this going to be streaming? Is it going to be printed to vinyl? Or is it being listened to on a computer maximized for a DJ and live, live playback. There are some different considerations that you want to think about. So when you are doing streaming, when you're producing for streaming, Spotify, apple Music, that sort of thing. This streaming algorithms will definitely reduce the quality of your audio. And if you ride right up to to the top of the scale, the 0 dB full-scale. With your mastering in your loudness, and then you send it out on streaming. The thing is that the streaming compression and their algorithms for processing and streaming that audio kind of cost about one decibel or one dB of headroom. So if you go up to 0, you can be sure that it's going to distort when it goes out on these streaming services because it's going to go one dB over 0. You don't ever want to go over 0 in a digital domain. So one of the nice things to do is to use a ceiling of about minus one. Now there are some other ceilings to consider when you are going to be dithering and other words, you're going to be taking this final master and you're going to be reducing its bit depth. So let's say you're mastering at 32 bit floating point, something like that. And you're going to be delivering this down to some other lower quality audio file, like a 16-bit file for a CD, you want to give yourself at least minus 0.03 dB in terms of this ceiling. So you could set it here to minus 0.3 if you're using ozone. Minus 6 to minus 0.6 to minus 0.8 if you are mixing for MP3s. So let's see, minus 0.6, minus 0.8. If you're, if you're mastering for like MP3 AAC type files but not streaming, these are going to be distributed and played on people's computers. For streaming, I would go with a full dB. And that is because the streaming algorithms and lossy codecs kind of require that extra dB of headroom that being set. Once you've got the ceiling, then you want to go ahead and bring down your maximising to get that target RMS. Now this is for target RMS, but if you wanted loves, which you can in ozone you can easily change this. Integrated basically means left here I've selected integrated as my metering, metering type. That implies left. Now we see laughs here. On the metering. I can see that the input lefts was minus 18.9, while my output lefts is minus 10. So that's quite a significant increase. And this is one of the nice things about left over RMS is that artists is very helpful for identifying the crest factor. And in a later lesson, we're gonna talk about using that crest factor, just sort of diagnose some loudness problems. But lefts is sort of a nice way to just get an anode, a generic number you're, you're trying to target. Minus 18 here is, is way too much dynamic range. Where as minus ten is, is much more appropriate for sort of commercial production. Now if you were producing for vinyl, you would probably want to go very light on the limiting and just build in a lot more headroom. So I might be more to back this off quite a bit. Maybe getting this extra two or three dB, then I'm getting here. I might even do no limiting at all and just mix to make sure my peaks never go over. Because in the world of printing vinyl, you want to have lots and lots of headroom. If this was something that I was going to be giving to a DJ to play it alive on a super heavy duty sound system. Then in that situation, I would probably max out the peak. Knowing that the DJs going to manage the level, right? There's going to be, there's going to be a human being who's in charge of managing this level. And you want to give them as much volume as the, as, as you can. Because that's going to be important utility for playing it live. So those are some considerations for best practices when you're setting peak levels. And also when you're using, loves to try to sort of make that final output a little hotter. Also, we talked about sealing this idea of what is the maximum level minus 0.3 would be appropriate in the case that your dithering down to a lower bit depth, you want to have maybe more than that, 0.6 to 0.8. If you're doing MP3s or a C is lossy files, but they're going to be played on a computer. And then minus one dB if it's going to streaming. And then bring this ceiling all the way up to 0. If you're, if you're doing vinyl or you're doing high, high-quality WAV files, for example, they're going to be played on a computer like in a DJ set up. Now in the next lesson, I want to introduce this concept of loudness normalization because a lot of people encounter this when they are making music for streaming services. And there are some key points to understand about loudness normalization versus loudness. 10. Loudness Normalization: So it's really helpful to the listener when they're listening to your music. If they don't have to constantly reach for the volume control as song switch in their playlist. So you know, that's a sign that your loudness control, your loudness mastering really hasn't been effective. Loudness normalization is really measured in a variety of specifications that are specifications in Europe, specifications in North America, specification for radio, specifications for TV, and there are specifications for streaming services like Spotify for example. Now, in that environment, they often talk about normalizing your loudness to minus 14 lakhs. And in ozone, they actually make it really easy down here to just set a target. Okay, I'm going to do minus 14 laughs, turn on Learn threshold and play the audio. Dynamically sets the threshold for me right? Now, just wait until that's done. And then, Oh great, That's what my threshold should be. Now I spit this out and it's ready for Spotify, right? That's sort of true, but also an oversimplification. Think of loudness normalization as specifications that people have come up with in different countries and in different industries to sort of normalize their content, the loudness of their content. But don't think of it as a target that you must reach. Because a lot of these streaming services, they're going to go ahead if your track is one or two dB quieter. And then their loudness normalization standard, they're just going to bump it up there, just gonna make it loud or write. These services are very smart about making sure that the listener doesn't have to reach for their volume control. And some of that is, is basically taken care of. You don't have to worry so much about making sure that you are meeting that loudness normalization standard. My recommendation is Use the other principles we're talking about crest factor, peak level, RMS using the integrated lumps value to sort of aim for a ballpark. But don't worry about hitting the target every time. Don't treat it like a target. Treat it like it's a ballpark number that you're using as a reference to make sure your loudness is in commercial standard range. Okay? Now RX, which is another product made by the same manufacturers is ozone. They have a really nice loudness plug-in and I'll show that to you here. I've got the loudness plug-in up here, and I'm just got the piano track. Right now. I can use a variety of different standards, which is what loudness normalization really is a bunch of industrial or international standards. I often like in the example of this video, in the series of videos that I make, I usually use the loudness standard EBU are 128, which is designed to set the integrated loudness at a value of minus 23. And RX does a great job of basically targeting that loudness standards. I can just render this. And I can see that now my integrated is exactly minus 23, whereas before my integrated value was minus 21.3. So if you're looking for some kind of quick fix to just set loudness normalization. And you're having trouble getting it in your DAW. Don't worry about it. I would say get your hands on RX and use the loudness plug-in here. It's a really great effective tool for that. But loudness normalization is about standards and it's a reference point. It's not something you have to meet a 100 percent. Now in the next lesson, we're going to revisit this concept of crest factor and talk specifically about what do we do if it's too high? What do we do if it's too low? And so you'll have some troubleshooting skills for dealing with that. 11. Strategy with Crest Factor: We talked about crest factor and introduced the concept earlier in this section. And crest factor is basically the peak level minus the RMS level, the difference between peak and RMS. And we can give that a number. So in this track here, if I zoom in, I can see the peak is coming in at 0 and my RMS is coming in, say at minus 12. And I know it's fluctuating quite a bit. That means 0 minus 12 is 12, so we have a crest factor of 12. A crest factor of 12 is pretty good. If it's more than 12, it might be an indicator that transient probably need more compression. Okay, so the most common problem in popular music, for example, with a crest factor being too high, is that you, you, your drums, the attack, the transients on your drums are so extreme. The difference between loudest and quietest moments in the drums is so extreme that it means you can't bring up the level on everything without getting distortion. If you compress the drums a bit, then you bring those, those hi transients down. You can bring the levels of everything up. And you can even do this with multi-band compression, which we will talk about in the next lesson. Now what if you're crest factor is too low, right? What I'm playing this, I was finding that my RMS was like minus 4 and my peak is 0, in which case my crest factors for, well, in that situation, there's a couple of things to consider. Basically, the premaster or stems you are given are probably way too hot there over processed, probably over compressed or over limited, or maybe they were maximized before you got to this loudness mastering stage. In which case, a re-mix may be advisable, going back to the original recordings and getting a copy of the recording where you have much higher crest factor is starting with like at least 12 or 15 to get started with, you know. So if it's too high, look at your transients and maybe compressing them. And maybe with multi-band compression, if you are needing to target specific frequency ranges, then multi-band compression allows you to compress just specific frequency ranges. If it's too low, if you're Chris factors too low, it's like 345. Then you probably want to go back to the originals and kinda re-mix it. Give yourself a lot more dynamic range so that your crest factor can go up. That's going to allow you to boost the loudness while maintaining that punch and clarity with a whole lot more control. All right, so we can actually use crest factor to make important decisions about maybe going back and doing a re-mix. Maybe where specifically we need to add compression. If you track doesn't have drums in it, look at anything with a lot of transients in it, and see if the transient range, the dynamic range in your transient tracks is more than it needs to be. In which case you can, you can look at compressing those. Another thing you can do to manipulate the crest factor is with EQ simply q. If you cut EQ in a, in a range where you know that there's a lot of difference between the peak and the average, then a cut in EQ will reduce your crest factor. A boost in EQ will increase your cress vector. Okay, so that's another tool you can use is EQ if you're trying to manipulate your crest factor so that you can bring up the loudness and maintain that ideal sort of range minus 12. I think we had talked about. Basically minus five to minus ten is the recommended range for your final crest factor, somewhere between minus nine and minus 12 for the RMS. In the next lesson, I do want to touch on strategically using a multiband compressor and talk about some of the tools for that so that you have, that you have an, an, an knowledge specifically of how multi-band compression is so precisely a tool designed for mastering and loudness that's coming up in the next lesson. 12. Strategic Multiband Compression: So in the last lesson we talked about how crest factor, which is the difference between our peak and RMS, which right here is coming up. Or 567 is sort of a way we can figure out if if our tracks got enough kind of range in it, right? And in the last lesson we talked about how if it's too high, maybe we need to compress some of our transient. If it's too low, maybe we need to go back to the original recording and kinda do a remix. In this situation here I've got a crest factor of roughly 78. Let's see if I can get that crest factor down to five just using compression. So if I load up ozone again, which has a compression module in it, I can use this compressor to manipulate my crest factor. If I tried to do this manipulation just with a general compressor, it's going to compress everything right? So if I do a lot of compression here for example, and I bring up my game. You can see now the difference between peak and RMS is smaller than tea leaves, three to five, like five to seven. Okay? The thing is, is that this compression, blanket compression over the whole mix is definitely increasing my my crest factor. Which might be a good approach for a really for, for an EDM track that is really covering the entire frequency spectrum. However, there may be elements here that I don't want to compress. Like maybe I don't want to compress the low end or I don't want to compress the high end. That's when you want to use multi-band compression. Now in a multi-band compression situation, you have different compressors for different ranges in the frequency spectrum. Now, one of the best ways to do this within ozone, which has got all kinds of awesome tools, is you can actually add in your own different markers here to separate out different actual ranges are different sort of sections of the frequency range to compress. And you can even have ozone learn it in certain places them for you and listening to music. But in a more traditional multiband compressor, you'll still get, for example, if in my DAW here, I do have a built-in multiband compressor that looks like this. And it, it's basically giving you the same thing. It's got four ranges that you can modify. You can bring the gain up, bring the gain down, and then down here you can actually, you have four different compressors for each range. So most DAWs have a built-in multiband compressor in ozone here it's, it's kinda select that it'll learn the cutoffs for you. But once I've done that, I can listen to my mix. I can select different ranges just by clicking in them, right? I'm going to select my lower mids here, which gives me all the compression controls for that section, which is independent from this section. Now when I play this back, I can see my gain reduction down here, right? So I'm going to bring the gain reduction down until that number is pretty steady, minimum of two. Then I'm going to bring my gain up by about two. I can also hit auto gain here. I'm gonna bring my low end compress that. Definitely need to now compress some at PNO, I can solo this range are solid, this range. So the piano is really coming through on this lower mid-range, right? I might not want to actually bring that and gain up and bring that piano down a little bit. But through compression, maybe you and bring the gain. And doing this, you can see that multi-band compression, compression approach to EQ, but you could also attempt this with EQ. The idea is I'm trying to, trying to rebalance this instrument. And I'm trying to give myself a little bit of margin. To bring up the RMS again. Because with my maximizer after dynamics, I can now bring my on my threshold and see that RMS value over here actually go down, which is going to reduce my crest factor. So I'm seeing more like five dB crest factor here. After doing that compression. Now, to my ear, this sounds over compressed. Definitely some nicer hair that comes if I just let that dynamics go and bring this up like so. And there in lies a valuable lesson about crest factor and multi-band compression and compression and limiting, which is that you can totally overcook these ideas. You do not want to do that. You want to make sure that you're maintaining that punch and clarity. What I wanted to discuss in this lesson was how multi-band compression and compression can actually be used to manipulate the crest factor as can EQ also be used to do that? Now, one other tool which is really handy and I just want to touch on quickly because I know probably not everybody has this tool. It's, it's another isotope tools called tonal balance because it actually has crest factor sort of meter, which is showing us on the left-hand side. This is really dynamic. There's transient. On the right-hand side. This is so compressed and so, so loud that there's almost no dynamic change at all. Unfortunately, it only give us crest factor for the low range, which has a lot of useful applications. But I wish that there was a tool that gave us crest factor for each of the ranges. But what we can see from this is that when we play the sort of see this ball moving. So this is showing us the crest factor for the low end. So it's actually like a meter that shows us this, which is, is pretty handy. That's only in tonal balance. And you can actually manipulate that in ozone using something called low-end focus. So for those of you who might have, have, have this, you can go to low and focus, which is one of the options in ozone. Just drag that behind my maximizer. And this is sorted doing what we did in the dynamic section here, which is, except that it's not allowing us to manage every aspect of the frequency spectrum. It's only allowing us to manage the low end. Thus, it's called low-end focus. And you can actually drag this contrasts. And you'll see in tonal balance that this crest factor ball moves based on this value. So if I move this all the way down, sorry, all the way up, we see that the tonal balance becomes more dynamic, right? To bring this down, we see that it gets more compressed. So this is really kind of a control of dynamics, punchy, of course, and clarity, right? So as we bring it down here, it's making it way, way, way compressed as we bring it here. It's sort of enhancing the contrast between loud and quiet, thus increasing our crest vector. All right, In the next lesson, I want to touch on this concept of true peak because that gets discussed a lot in the context of loudness. And you want to know what it is and how to use it. 13. When True Peak Matters: True peak is a measurement of how, what's going to be the peak once this digital audio gets translated into an audio signal and it's actually coming out of a speaker. In my insight plug-in here, I have true peak value right here, 70 v. And here I, in my ozone, I can actually enable this true peak option here. And then it, it basically is going to look at targeting this ceiling based on true peak. In other words, what is the level going to be when it actually gets translated back to analog and comes out an analog environment, a headphone or a speaker by enabled true peak. Then it's gonna kinda simplify that. And there might be this feature in other Mastering tools, other maximizers. But true peak is different from digital peak. Digital peak is what is registering is the peak level within the digital signal. But once that goes through your digital audio converter over the wire comes out a speaker or a headphone. Then it's in the world of analog. And in that world, what's the peak going to be there? And so you want to manage this. He didn't blow anyone's speakers, right? Basically maximising to true peak means that you're still maximizing, but you are protecting audio equipment like speakers. There are a couple other cool ozone tips and tricks which I'll share in the next lesson. And then I'm going to recommend a class project that I think will help take the last several lessons. And for any student, help them digest and internalize what they've learned. 14. Ozone Tips & Tricks: There are just a couple other ozone tips I wanted to discuss. One of course was down here to learn threshold based on a target, which is great for just quickly making sure that your audio meet some specification, some loudness normalization specification. And I use EBU, our R12 eight for my video broadcasts. I'd have Spotify recommends minus 14. But it's really great. You can just change your target to whatever you know, minus 5. Let's get crazy. And then we can learn threshold. It's going to totally over compress it, but that's what it costs if you want to get to a target of minus five. If we change this to minus 18, which is going to be at a very dynamic type of target. Doesn't even have to add limiting. So you can see if it recognizes you don't even have to add limiting to get there. So learn thresholds, really awesome for loudness normalization. Also over here is this thing called transient emphasis, which is really handy in many situations where after you've done your compression, you feel that the transients, particularly in your drums, are being compromised, right? So you can recover some of those transients using this transient emphasis. So let's take, for example, a situation where we are going to bring this down and set this at minus one for streaming, leaving true peak on. And I'm going to bring this down until my RMS up here, my hand or my left is in the right range. I'm actually going to bring it down a bit more. And let's say at this point, I'm looking at the blue trace line here and I can see that those hits are being heavily compressed and we can actually turn on transient emphasis. And I bring this up, can hear the kick actually coming out as I do this transient emphasis, it's a bit much for me. So the transient emphasis is a great tool for trying to recover some of the punch. Maybe you lost when you are doing the limiting and you're maximizing is maybe sort of obscuring or hiding or diminishing the power of your transient. It's too much. You can use this transient emphasis. And it somehow manipulates the transient before the limiting is applied. If you have any questions on anything that we've talked about so far, please please let me know you can reach out to me via direct message. You can post in the discussion board. I love hearing from students that try to follow up as soon as possible. And I would love to know if there's topics you'd like me to go into more. If you have questions about things we talked about or there's things I didn't talk about that you wish I'd covered? As long as it's constructive criticism and I can respond to it by improving the course. I'm all ears. It's great to hear from you. I would, I would love your opinions or your feedback or your questions, so please feel free to reach out to me. And in the next and final lesson in this section, just going to wrap things up and suggest a project that you can complete. 15. Wrap Up & Project: So in this course, we've covered a ton. We've talked about absolute loudness, measuring loudness with love's punch and clarity. Crest factor using simple gain as well as limiting and RMS to start setting your levels. Obviously how to avoid squashing some peak level best practices, all kinds of things oriented around getting you up to speed on understanding how the factors of loudness are perceived, as well as metered and measured within different DAWs. And I'm giving you the materials, the raw materials to do all this on your own. So I'm going to suggest, once you've finished this course and you kinda wanna dive in, you can start with your own mix if you have one, great. There are also mixes and stems you can download from the internet. If you do a Google search for free stems, you'll find some out there. I've also posed to the long with these lessons a collection of stems from the Amazing Grace sample that I'm, that I'm using in the videos. So you can also use those. And the idea is load these into your project, right? So grab the stems, bring them into your DAW. Whether you're using able to in Pro Tools Logic, Cubase, that doesn't matter. Once you have got the stems loaded in, then what you wanna do is start trying to do this loudness maximization. And for that, I definitely recommend ozone. I think ozone is a great tool, but there are lots of other great mastering plugins. And you can also do many of these things with built-in plug-ins for whatever DAW that you have. But load up a maximizing plugin and you want to go through these steps. Can you do simple gain is the peaks of your project, your, your audio allowing you to just do simple gain to bring up the volume. Then, how can you use limiting and looking at your RMS value using limiting to bring that RMS up so that your crest factor is somewhere in the five to ten dB range. Five being really hot, ten being pretty good, professional level, even 12 being, being pretty good. If it's a bit more dynamic, then you want to look at doing what kind of loudness normalization are you targeting if you're using ozone, then of course, you can use this learn threshold and target to set your loudness normalization if you wanted to do that, I just talked about how you could use isotope RX and its loudness plug-in to do that as well. Look at how you can understand the sources of your crest factor and your loudness. What did the most dynamic elements in the mix are those dynamic elements to dynamic and other words, are they to transient or their peak levels, making it impossible for you to get decent loudness, bring up the level overall. And then how can you resolve that with maybe compression, multi-band compression, even EQ or dynamic EQ. Now you know what the steps are. You can modify everything I've demonstrated here to your DAW using the tools you have. And if you have any questions, you need advice on plugins or tools that you could use in your DAW to accomplish these loudness goals. Please reach out to me. I wish you the best of luck, and thanks so much for taking this course. I look forward to seeing you in another one of my courses.