As an audio mastering engineer, I know firsthand that everyone wants their music loud. In fact, I have a form at my MusicGuyMastering.com where potential clients can upload a song for a sample master, and the form includes a “Notes” section. The number one note/request I get from clients relates to loudness in one way or another. It makes sense; it’s a biological fact that we hear the same song side by side, one at a slightly louder overall volume, and we’re conditioned to think the louder version is better. With that in mind it was probably inevitable that the Loudness War happened, so let’s talk about what it is and everything to know about it as concisely as possible.
What is the Loudness War
The Loudness War is a term used to describe the general increasing of volume in commercial music releases over the course of history. “The course of history” is admittedly a broad spectrum, but it technically goes back to the 1940’s when an instance of this was first recognized. The term especially came into vogue in the 90’s and early 2000’s when the industry was actively pushing louder records to the point that it became competitive.
Obviously every song technically has a ceiling for how loud it can be. The ceiling that is audio clipping, or turning up a waveform to the point that it begins distorting, is non-negotiable. Here’s an illustration of what roughly happens when you turn a waveform up above that ceiling of 0dB:
Therefore, the way to get more perceived loudness (see LUFS vs dBs) out of a song is to reduce its dynamic range so that it’s closer to the point of clipping on average throughout the song.
This is typically done with compression which uses an infinite ratio, otherwise known as a limiter.
With that infinite ratio, the compressor threshold becomes which nothing can penetrate. Set the threshold so that 10dB is being compressed, with an infinite ratio you achieve 10dB of gain reduction. This is in contrast to say a gentle ratio of 2:1 which would still allow that threshold to be breached by 5dB (10/2=5).
This takes dynamic audio which looks like this with its peaks and valleys:
… and turns it into this (incidentally I get A LOT of song submissions for mastering which look like this):
This is what causes those fat, sausage-like waveforms for a song – you’re getting long stretches of the output level being exactly the same.
The practical effect of this is that the listener experiences fatigue of being fed the same volume over and over again. Dynamics are natural and challenge our ears, keeping them engaged throughout the song. Without them, the song more resembles a constant drone which we eventually tune out.
In plain terms, this is why the Loudness War is universally considered to be a bad thing.
The competition of making popular, commercial records incrementally louder and louder to “one-up” the last song which was released has effectively strangled out the dynamic range which was created during the recording and mixing process (painstakingly, through gain staging).
Example of Loudness War
What really showcases the effect of the Loudness War is when you compare the waveforms of popular records which were re-released over the years.
This is taken from Wikipedia; this is the waveform of Abba’s 1980 song Super Trouper, comparing 4 different releases over a roughly 30 year span:
You can see the initial 1980 release is understandably easily the quietest out of all of them. You can also see there’s one instance it appears in that final chorus where the peak is almost hitting 0dB.
Conversely, the 2001 “remaster” cranks it up so you have multiple instances of the peaks hitting the ceiling.
Of course the 2005 version puts the 2001 to shame as part of the box set as it’s easily the most limited of the bunch. There’s very little dynamic range, particularly on that final chorus.
Ironically the 2011 or most recent example shown brings it back down and is the closest thing to the original release.
This brings us to today and specifically the entry of music streaming services.
Streaming Normalization
Because of their ease of access and size of their libraries, music streaming services like Spotify and Apple Music are how most people consume music these days.
Aside from revolutionizing the music industry in countless other ways, they’ve had an interesting impact on the Loudness War.
I touched on this in my recent overview of normalization (see what is normalize audio), but streaming services have a setting called “Normalize Audio” or some variation of that.
Typically left on by default, all songs on the platform are automatically adjusted to be louder or quieter as necessary to meet a standard average volume so that the listener doesn’t have to constantly adjust their own volume to account for the differences in volume between songs.
This feature has thankfully largely reduced the need to make commercial masters as loud as humanly possible since they all come out sounding roughly the same volume for most listeners.
I say roughly because the algorithms which Spotify, Apple Music, and the rest use to normalize audio get it incredibly close, but a song which is mastered to be -4LUFS (insanely loud) will still be louder more consistently than a song from the 1970’s for example which has a lot more dynamics.
Still, the merits of having a quieter song with more built in dynamics which might need to slightly be turned up by streaming services upon playback are easier to argue than mastering a song to -4LUFS with no dynamics and then having it turned down on top of that by streaming services.
That’s where the Loudness War exists today – some combination of nowhere else to go, a greater awareness of it as well as an appreciation for dynamics, and normalization being the reality for how most people consume music these days.
Speaking of which, I did a brief overview of what LUFS to master to for Spotify and the rest of the streaming services, so check that out as it’s an interesting read and ties in nicely to this.