Art Of Darkness

To observe that music is mastered too hot these days is a little old hat, so I will not belabor the point here. But I did want to say a quick word about Pitchfork's recent article "The Dark Art Of Mastering Music". Never mind that the piece is shallow and fawning in its appeal to mysticism, and how the Loudness War is only indirectly referenced by way of mentioning one of its highest profile victims. Forget all the words that ultimately end up saying very little about the process and why music lovers should care and focus on the provided audio clips.

This is just a screenshot. You can't click those buttons.

They are not explained at all in the body of the article, and the only information provided is that one is the Unmastered Version and the other is the Mastered Version. No spoiler alert is necessary here: The Unmastered Version plays rather quietly compared to the Mastered Version. Of course, they sound different in more ways that just absolute volume, but that is besides the point. The problem here is that in order for this to be anything like a real comparison a reader/listener might learn something from, they should be volume matched.

Before going any further, let's just look at the waveforms for these audio files:

You don't have to be a rocket scientist to see the problem here (er, hear). Without getting technical, these are basically time vs. amplitude plots of a digitally represented audio signal. Recall that sound is a pressure wave which vibrates the particles of the medium through which it is traveling (e.g. waves displacing particles of air and hitting our ear drums). It is obvious that when these digital audio files are converted back in to sound in the real world through your headphones or speakers into your ears, they are going to sound vastly different. I'm not talking about a slight treble boost or gentle tweak of the mid-bass -- this is a dramatic difference in loudness.

So why does this matter? It is nigh impossible to perform a meaningful comparison between these two "before and after" audio clips when one is so much louder than the other! Specifically, the article discusses mastering engineer Joe Lambert moving the vocals forward in the mix and having to boost the bass to compensate. But how can a discerning listener hear this when everything is so much louder than the Unmastered Version?

The proper way to do this little test is to turn down the mastered version so you can switch back and forth between the two. In order to match the volume for comparison, I attenuated the mastered version by 10.5dB so the perceived loudness was closely matched. Here's what it looks like now:

Hey, wait a minute -- where dem peaks go!? They were compressed out of existence. But when I play the files back to back, they both sound equally loud. (Perceived loudness has less to do with a loudest audio peak and more to do with how often it approaches it. That's an oversimplification, but them's the breaks.) Now I can truly compare the two versions of the song.

So what's the verdict? The Mastered Versions sounds as flat as it looks. The Unmastered Version affords the music room to breathe, and thus has more impact. Granted, one must turn up the Unmastered Version relative to just about every other mastered piece of music out there, but so what? I have a volume knob and I'm not afraid to use it. Therefore, it should come as no surprise that I prefer the Unmastered Version. In my view, there is less music in the Mastered Version even though more of the space is filled up in the waveform plot. This article inadvertently proves that much of what "mastering" is these days is harming music.

The transient peaks in music matter because when I turn up a dynamic (i.e. unmastered) recording, I can achieve a comfortable listening level while still retaining those peaks. With the mastered version, the engineer has turned everything up for you and chopped off the peaks. And you're never getting those back. You've cut off the top of the mountain range, and now everything looks the same. There is less contrast. There's no shadow in your photograph or favourite film. Everything is unnaturally bright and the image becomes tiresome and uninteresting after a while. There is too much sweet and not enough sour, and other questionably relevant analogies.

The lesson here is when comparing the elements of a mix before and after mastering, one does not simply set the same playback volume for both versions since one is likely to be considerably louder than the other. And it matters. Broadly speaking, I wish more music escaped the mastering stage looking and sounding more like the Unmastered Version in this example. Digital audio is capable of reproducing a startling amount of dynamic range.  "CD quality" 16-bit audio has a theoretical dynamic range of 96dB, or as much as 120dB if properly dithered. That's an awful lot of canvas upon which to paint. It is a shame that by the time the music reaches your ears, only a fraction of that space has been truly taken advantage of.