The Loudness Wars: Why Music Isn’t What It Used to Be
Are you excited? I am! This is my first “in-depth” (i.e. horrendously long) post!
Okay. I’m about to explain why my new song, “My Creed,” is quieter than my first five uploaded songs. I initially wanted to do this rather briefly, but I think it would be fun to talk about it in some length (i.e. AN UNREASONABLY LONG ESSAY). Consequently, I’ve created (as well as I am able to with my knowledge and experience) what can really be called a full treatment for the layperson on the topic of the Loudness Wars. Also consequently, if you already have a basic understanding of this topic, you should probably go do something more worth your time, like going over to Aunt Ruthie’s with a few buddies of yours and playing a nice game of Bridge because, you know, Aunt Ruthie doesn’t really have anybody and she would just be so happy if she had a visitor once in a while. Besides, it would be really neat to learn to play Bridge, don’t you think? And it would make her feel special to teach someone something new. It would be a great bonding experience for the two of you. And then, at her inevitable funeral, you’d have something to say, like, “I’ll never forget all those wonderful evenings of Bridge. She taught us all how to play, you know.” As opposed to, “Wait, Mom, Aunt Ruthie…she was the one in the wheelchair, right?”
However, if you aren’t knowledgeable about the topic, or if you’re perfectly content harboring a lifetime of bitter family regrets, then I invite you to stay! It might be interesting for you. I’ve added some sound clips and visual aids, so it’ll be fun as a nice game of Bridge!
Also, if I’ve played my cards right (Get it? Get it?), you should leave this site with a solid understanding of:
1. What the Loudness Wars are and why music today sounds louder than it did twenty years ago.
2. How easily we can be tricked into believing that louder mixes are better.
3. How to keep your man interested.
3. How this has negatively affected music and why mixing engineers hate it.
Here we go!
Okay, first off, I’d like to make it clear that I’m no expert, so I’m going to try to refrain from asserting things I’m not sure of. That being said…
So there’s this thing that’s been going on in the music industry for the past, oh, probably almost twenty years, dubbed the Loudness Wars! (insert stock music which sounds as similar as possible to the theme from Star Wars without risking lawsuit). Basically, it means music has been mixed and mastered to progressively (or regressively) louder and louder levels in recent years. You may have noticed that when your iPod is on “shuffle” and “Baby Don’t Forget My Number” by Milli Vanilli comes on right after one of brutal death metal band Cannibal Corpse’s latest hits (say, “As Deep as the Knife Will Go” or “Followed Home Then Killed”), you suddenly have to turn the volume way up just to hear the synthesized orchestra hits, even though mere moments ago, vocalist Corpsegrinder’s guttural screams of “Silently I enter through the broken basement window now to wait” fell upon your ears with horrific clarity. Well, my friend, you are experiencing firsthand the casualties of the Loudness Wars.
1989: Before the Loudness Wars destroyed, you know, good music. Um…by the way, I wanna know how they got Michael Jackson to play keyboards for them.
Why are the Loudness Wars happening, you ask? Because if you’re going by first impressions, louder is simply better. Here, I’ll prove it to you (hopefully). Listen to the following sound clip. You’ll hear the same recording four times in a row. Each time you hear it, however, it will sound different because I’ve done some different processing for each one. As you listen, try to make a note of whether the recording you’re listening to sounds better or worse than the one directly before it. Don’t think too much. Just first impressions. Make sure not to make any volume adjustments once the recording has started, or my experiment won’t work. I’ll give you a good starting volume. You know that sensational Milli Vanilli video I gave you the pleasure of enjoying? Well, listen to it again, but set the YouTube player volume to be roughly 25%. Then, adjust your computer volume until you’re hearing Milli Vanilli at a comfortable listening level (I know. I get it: the only comfortable way to listen to Milli Vanilli is not at all. Just bear with me, for God’s sake). This should be a good volume. Anyway, here goes:
Hey, Matt! Do you sound any good talking into that microphone?
Okay, so, I haven’t tested this on anyone yet, but hopefully you had something like the following experience: you heard the first one and thought nothing. You heard the second one and thought, “Oooh, it sounds so warm and fuzzy and up close now!” Then, you heard the third one and thought, “Oh, it’s really quiet again. Lame!” And finally, upon hearing the fourth one, you thought, “Ahh, there we go. That’s more like it.” If not, I’m terribly sorry. It’s just an example.
Now, here’s the thing: the first one and the fourth one are identical, except for their volume, and the second and third are identical, except for their volume. Basically, I took one recording and equalized it two different ways, but I played the recording a total of four times at various volume levels. Here are the two equalized versions you heard at approximately the same volume:
Yeah, but do you really? Tell us, Matt! We wanna know!
As you can hear, the difference is significant, but if my hunch is correct, you had a positive reaction to both of them at higher volumes and a negative reaction to both of them at lower volumes. You may have noticed the difference anyway, but I’m sure you’ll agree that your judgment is better when loudness isn’t there, hypnotizing you with its seductive powers…
Record producers at some point began to compete with each other using sound levels. And for reasons that I don’t want to go into (nor do I fully understand), the advent of digital music production made it possible to go EVEN LOUDER, so I think people got a bit carried away. The problem is that it’s sort of like the nuclear arms race: victory is meaningless if you destroy everything you’re fighting for. You see, when you produce sound recordings, you can only make the signal so loud before it distorts. Now, digital sound production turns distortion into a sort of on/off switch. It used to be (during the analog days) that, as you got louder, the signal would gradually become more and more distorted (I think). Nowadays, the nice thing is that you can go a lot louder without distorting, but once you do, you REALLY do. It’s basically a big ceiling, and if any of your sound is loud enough to hit the ceiling, it crashes right through, and the results are, well, crackly. Here’s an example of mortifyingly awful digital distortion using a clip from my song, “My Creed”:
Yeah, for this new album, we were totally going for that “subway announcer” vibe ’cause, I mean, we figured that’s where, like, so many of our fans listen to our music on their iPods, you know?
Now, the good news is, there are tools called “limiters” (sometimes called “brick wall limiters”) which stop the sound just before it slams through the roof. The volume is instantly decreased and will not be allowed to go any louder. This is what it sounds like if you push the level just as high as in the previous clip, but use a brick-wall limiter on the signal first:
That polished, over-produced, “sellout” subway announcer sound…
As you can probably tell, it doesn’t sound nearly as bad, but it’s still distorted. Let it be known, though, that nobody in their right mind ever pushes a limiter this hard. I create these clips to illustrate an obvious effect. Here’s what the clip I used sounds like originally, before the mastering phase:
That under-produced, hack-who-seriously-thinks-he-can-produce-music-in-his-bedroom, wannabe-sellout sound
Unfortunately, even with brick-wall limiting, if pushed too far, everything eventually distorts.
Now, why didn’t the original clip sound any quieter than the previous two? It’s because for the first two, I turned down the output volume at the very end; that is, I turned it way up first, let the effects of that happen, then turned the final result back down in volume. I did this because I didn’t want you thinking that the louder clips sounded better just because of their volume (if that were even humanly possible). This way, you can really compare them.
This brings up a second point, though. I can turn the volume down after I’ve squashed the audio, but the damage is done, which means that if an album is mixed extremely loudly, you can turn down the volume on your CD player or iPod, but it will still sound like butt-poop.
Now, another effect of limiting is something called dynamic compression, often called just “compression.” In fact, a brick-wall limiter is technically an extreme version of a compressor. What is dynamic compression? Like the name suggests, it is a compression of the dynamics in a sound source. Don’t worry. I’ll explain. In music, dynamics is the element of louds and softs, so when you compress dynamics, you are decreasing the range of louds and softs. In other words, you’re making the quietest parts louder and the loudest parts quieter so that the end result is closer to being all one consistent volume. You see, recorded sound doesn’t maintain a single constant volume at every moment. There are peaks and valleys, which mixing engineers can see as a visual representation called a “waveform.” It looks like this:
See how the waveform (the lighter blue portion of the image) looks uneven and “wavy,” with little “spikes” happening every once in a while? Those spikes are snare drum and bass drum hits, and believe me, they’re already heavily compressed, but the song sounds really good. What you’re looking at is part of the waveform for the song, “Suck My Kiss” by the Red Hot Chili Peppers, released in 1991 (I don’t want to stream it because I don’t have the rights, and I would direct you to the YouTube version, but that one is more compressed than the original release, so you won’t be hearing what I hear anyway).
There are probably some “purists” out there who believe that compression is always bad and that if a musician is good enough, he should be able to play exactly how loudly or softly he intends to at any given moment. Classical musicians are often this good, and orchestral recordings are produced very dynamically. While this is appropriate for classical music, I think that compression can make pop or rock music sound really good. Yes, even heavy compression. If used skillfully, compression can create a very smooth, balanced sound so that, as you listen, you can choose which instrument you want to focus on, and you’ll always be able to hear it sitting in its snug place among the other instruments.
The trick is, in order to achieve this, you have to compress each instrument individually, which means you have to record each instrument individually. Orchestras are usually recorded in single takes, playing as a group. Rock music was once recorded this way (listen to some Elvis), but now, common practice is for each part to be recorded and compressed individually, to some very pleasing effects, in my opinion. As a listener, you don’t have to keep changing the volume while driving 70 mph on the expressway (turning it up at the quiet parts and back down at the loud parts) like I frustratingly do when listening to the orchestral Gladiator soundtrack.
On the other hand, too much compression can sound equally bad. Have you ever been frustrated while watching a movie on TV by the fact that the dialogue almost sounds louder than the explosions? That’s because the mix has been so heavily compressed that the movie plays at one unwavering volume (don’t ask me why they do this). Another example can be heard if you listen to talk radio, and as soon as someone stops talking, you hear the hiss of the room and its equipment as though it’s suddenly as loud as someone’s voice, as in this clip:
I’ve found that there’s almost never any convincing excuse for the way I behave.
The hiss is always there, but relative to the voice, it’s very quiet. But if the signal is heavily compressed, the quietest sections of the recording (the sections where the hiss is all that the microphone is picking up) become just as loud as the loudest sections (when someone is screaming and the hiss is drowned out). This is somewhat necessary in radio because radio is live. You can never know how loudly someone is going to talk, and you can’t risk overloading the signal, breaking through that ceiling and causing horrid distortion. So you compress the hell out of everything as it happens.
So, good-sounding compression is really a balance. If everything were purely at the same volume (and sounded good that way), getting recordings to be super-loud would be easy: you’d just boost all the way up to the ceiling. But when you’ve got little bursts of peaks here and there with everything else at a lower level, you have to compress those peaks a lot if you want the volume of everything else to increase a lot. If you look at the waveform again for “Suck My Kiss”…
…you’ll notice that there is a lot of dark blue surrounding the light blue waveform. It may surprise you, though, to know that some of those snare and bass drum peaks actually reach the ceiling; it just happens in tiny moments that you can’t see. But the overall recording doesn’t sound that loud compared to today’s recordings. That’s because there’s a difference between the volume of the peaks and the perceived volume of the sound as a whole. Our ears play tricks on us, and how loud the song feels is actually at a spot in the waveform that is below all of the peaks and valleys, well into the “meat and potatoes” section, if you will. You can imagine how much of this waveform would have to be destroyed to get the core to be close to the volume at the ceiling, but you don’t have to imagine it. Here is the waveform for “Know Your Enemy,” from the latest Green Day album:
It looks like we’re zoomed in, doesn’t it, like we’re not seeing the whole thing? You can easily imagine all those spikes extending past the boundaries to their natural stopping places, right? Nope. Those peaks are no longer a part of the music. YEAH, BUT LOOK HOW LOUD IT MUST BE! BODACIOUS! Yes, that’s true. Of course, in reality, as soon as the song comes on after “Suck My Kiss” has finished, you’ll just end up turning down the volume until it sounds about the same as the previous song. So let’s see how the two look at similar perceived volumes, shall we?
Yes, that’s right. This is the music you’ve been listening to the last few years. At the beginning, all we hear is drums, and you can really see that every snare and bass drum hit has been chopped off well into its peak, and then, when the rest of the band comes in, forget about it (The clip of “Suck My Kiss,” by the way, starts with the full band playing…loudly). Once the awesome illusion of LOUDNESS is out of the picture, this looks comical, doesn’t it? I mean, it looks like your dad took a pair of hedge trimmers to your favorite song. Make no mistake: you’re looking at the actual music having been destroyed. But this is no amateur garage-band demo. No. This is a professionally released master track, a finished product produced by some of the top sound engineers in the entire world, reproduced in a supposed pristine lossless audio format onto a shiny new compact disc, or God forbid, a vinyl record. AND THEY’RE EXPECTING YOU TO PAY FIFTEEN DOLLARS FOR IT. That would be like someone expecting you to pay full price for a Porsche that looks like this:
Of course, it’s still a matter of degree, and in all honesty, most of the core of the waveform is preserved in this particular recording (no, really, technically, most of it is), so the actual audible differences are still relatively subtle, believe it or not, especially since the guitars are already intentionally distorted, and everything is already intentionally compressed. The mixing and mastering engineers worked very hard to conceal the effects of their brick-walling. As a result, I can hear only a bit of over-compression and yes, even distortion in this song, but most people probably won’t care or even notice. I’m not trying to complain that this song sounds terrible when you listen to it. It actually still sounds great, way better than I could do with my equipment and experience, but it could sound at least slightly better.
Now, as I said, it’s a matter of degree, and people keep trying to push things further and further. With Metallica’s latest album, Death Magnetic, the Loudness Wars finally reached a point where my enjoyment of the songs is actually noticeably affected. This is the loudest album I have ever heard in terms of perceived volume. Look at the waveform of the loudest section of the song, “The Judas Kiss”:
Yeah…So, I can no longer imagine how this waveform would extend beyond the “ceiling.” I quite simply can’t see enough of it. And you see that “quiet” section in the middle? All that’s going on there is a single vocal track and a few half-closed hi-hat cymbal hits. And it’s still smashing its way repeatedly into the struggling brick masonry above. Throughout this entire album, I personally can hear a constant, steady, and unmistakable crackling distortion. And don’t think I’m just hearing the guitar distortion. At the very beginning of the above song, the tom fill in the drum part sounds like Nintendo. It’s horrible, and I’m not the only one who thinks so. Upon its release, there was actually a backlash from enough fans to create some internet media attention. Just for fun, here are the waveform comparisons for “Suck My Kiss,” “Know Your Enemy,” and “The Judas Kiss,” respectively, all at the same perceived volume:
Now, even if a recording isn’t pushed quite as loud as Metallica’s latest effort and the difference is more subtle, there still is a difference. I’ll give you an example with a simple drum-track, since drums are usually the part of the mix most obviously affected by limiting. The first recording has not been brick-walled at all, while the second has, and very heavily. Listen especially to the bass and snare drums. A word of warning: these sound clips are a bit louder than the ones in the first half of the post, so you may want to turn down a bit. However, you may find that you’re able to hear more of a difference when listening to these clips on the louder side of your comfortable listening level.
No Brick-Wall Compression
Heavy Brick-Wall Compression
The differences are subtle, but in the brick-walled clip, the bass and snare drum have lost a good deal of their volume and impact. There is not as much “thump” or “pop,” and the snare is quieter compared to everything else. In contrast, because the different sounds are spread out over a wider spectrum of volumes, the first clip has a certain “openness” and clarity to it. The second clip sounds, I don’t know…messier. Still, there is a quality to the brick-walled version you might like more. Namely, the compression has the side-effect of increasing the volume of any echo-ee, splashy reverberations, making certain drums sound “fatter.” However, there are other ways of beefing up the drums without squashing them like a pancake, such as adding the signal from an ambient microphone into the mix:
No Brick-Wall Compression, Ambient Mic Signal Added
Heavy Brick-Wall Compression, No Ambient Mic Signal Added
I could go on making all sorts of adjustments to either track to get it to sound the best it can, but I think the important thing to take away from these examples is that generally speaking, any time you use limiting compression, you sacrifice that openness, punch, and clarity that you hear in the first clip of each pair above.
To be fair, a master track will usually sound best with at least some limiting, albeit at a happy medium between the two extremes illustrated in the above clips. And it may very well be that, in certain cases, I happen to like the sound of my drums when they’re crushed to absolute death (to get that consistently “gritty” sound). The point is, I can always do that as a choice, rather than being forced to try to make everything sound as good as possible under the circumstances of my musical creation being sonically squashed.
Now, that’s just drums alone. On a full mix, the effects of brick-walling are even more subtle, BUT you have to work much harder to make it subtle. Every mix has its own threshold for brick-walling; that is, every mix will allow you to brick-wall it in a way that’s barely noticeable, but there will be a limit to how loud you can make it before it starts sounding like that absurdly brick-walled clip of “My Creed” I posted earlier. The only way to raise this threshold to the point where you can get your mix to sound as good and as loud as “Know Your Enemy” is to do all kinds of work on individual tracks and the mix as a whole to optimize its ability to be brick-walled.
And that’s exactly why I’ve decided to stop doing this: I often do have to go to great lengths to get a mix loud, and I just don’t see the point anymore. After all, I’m certainly not doing it because it necessarily sounds better that way. And quite frankly, I’m sick of dealing with it. Sometimes, I’ll mix a song, then push it through a brick-wall limiter, and it will still sound good. Other times, for the life of me, I won’t be able to get a mix to be as loud as my previous songs without it obviously distorting in a really unpleasant way. I’ll have to go back and forth, changing equalizations, messing with compression settings for individual tracks, applying volume automation meticulously to the lead vocals – it becomes an entire project all on its own. Every step of the way, from the bottom up of each individual track and the recording as a whole, I have to make decisions not merely based on how to make the mix sound good (which is hard enough), but also based on how to make the mix sound passable when it slams into a brick wall just so it can be as loud as the current “trend” in commercial recordings. Because if I don’t think about it from the beginning, the end result might end up sounding something like this:
Only good radio is a ham radio, I say.
My thinking now is as follows:
- If I make my mixes super-loud, they will, in almost all cases, sound at least slightly worse.
- If I make my mixes super-loud, it will take me a lot more time to finish them.
- If I absolutely can’t get people to want to listen to my music, making it louder will do absolutely nothing to change their minds.
So yes, “My Creed” sounds “quieter,” but the good news is, you can always turn up the volume on your sound system. The listener controls the listening volume. He always has. And I’m sure you don’t ever turn the volume all the way up, do you? You’d blow your speakers. So I think I have a bit of room to spare up there by that poor ceiling. And as a matter of fact, the version of “My Creed” that I posted is moderately brick-walled, and it’s still slightly louder than “Suck My Kiss.” But I see no benefit in over-doing it anymore.
Also, I’m going to re-upload everything very soon so that all of my songs are at comparable volumes (quieter, but better).
Well, this has been excruciatingly long, but quite fun for me, actually. If you’ve actually gotten as far as this sentence, then it really makes me feel special that you would pass up that opportunity tonight to watch the entire Lord of the Rings Extended Edition Trilogy, which should be wrapping up right about now. That or spending quality time with your Aunt Ruthie. I apologize if I made you miss that. Oh, well, it’s water under the “Bridge.” OH MY GOD, YOU DIDN’T SEE THAT ONE COMING, DID YOU!!!!!!!
Seriously though, thanks for reading.