The “loudness war” is not actually about loudness

I try not to talk about the “loudness war”, because it tends to feed the trolls. Let’s remember that “war” here refers to competitive mastering practices, not an argument between analog and digital lovers!

I recently came across an article titled Is the loudness war a myth? on an otherwise excellent blog. I’d posit that the loudness war is NOT a myth, but its not exactly about loudness either.

The “digital loudness war” refers to the pressure put upon music producers, and especially mastering engineers, to achieve a high average (Root Mean Squared) sound level at any cost. Basically this came about as a result of the “shuffle” feature on digital music devices - nobody wanted their song to be the quiet one. (The actual loudness war is much older, going back to the days of AM radio, but that’s a different story.)

The article basically argues that because the macro-contrast between quiet and loud sections of a song are preserved, the loudness war isn’t a problem.

“Crest factor” is simply the difference between the peak (instantaneous) and RMS (average) levels in an audio track. Since the peak level of a digital file cannot exceed 0 dBfs, the dynamic range of any peak-limited, or normalized, recording can be simply stated as its RMS value. In the 90’s, this would be anywhere from -20 dBfs to -15 dBfs (15 to 20 dB crest factor). In the 00’s - half that. Reduction in crest factor has absolutely been happening and distortion is the result.

It’s easy to see why this is a problem: The quest for higher RMS has led to clipping the stereo buss, as well as a host of other loudness-oriented (and distortion causing) production techniques. Despite the fact that the CD format offers enough dynamic range - that is, signal to noise ratio - to not require clipping the peaks in order to maintain a decent noise floor, that’s exactly what’s happening on 99% of modern music (a statistic I made up, in case that wasn’t clear). So now, listeners complain about 16 bits not being enough and demand 24 bits. But, as long as you smash the music up against the digital ceiling, 24 bits won’t be enough either. The problem is not bit depth, but how those bits are allocated.

In the end, the loudness war is not about the contrast between quiet and loud sections: it’s about preserving the transients without distortion.

In a way, its not about loudness at all, but about quality.