"This sucks. Can we leave?"
"I was going to ask you the same question. Why do they do that?"
I was in Austin's Continental Club, catching one of the many bands that were performing at this year's South By SouthWest music-biz conference. (Robert Baird reports from SXSW on p.154.) The Weary Boys were playing that most basic of American musics, bluegrass, on mainly acoustic instruments: guitars (one electric), fiddle, and double-bass. Yet the sound of those instruments via the club's PA system was relentlessly loud, compressed, clipped, and distorted to the point where any light and shade in the music had been obliterated.
And what is music, if not light and shade? Music without dynamic contrast is merely noise, something to be endured as background, not sought out as a destination activity. Perhaps that was the case in the Continental that night. As my friend and I left the club, it struck me that no one else seemed bothered by the sound. Perhaps they weren't listening. No one was dancing. Or perhaps they were all so accustomed to it—this is what music is supposed to sound like—that their expectations had been diminished.
But why should people who pay to hear music have become accustomed to such assaults on their senses? Could it have something to do with FM radio universally sounding, as Barry Willis writes in this issue's "Industry Update" (p.18), like "a sea of tightly compressed noise"? Why do they do that?
Could it be that, as I wrote in my December 1999 "As We See It," which was triggered by the relentlessly loud, compressed, clipped, and distorted sound of Carlos Santana's Supernatural and the Red Hot Chili Peppers' Californication, that the promise of CD as a high-fidelity medium has been lost in the spiraling downward desire of record company execs to make everything they release louder.
And louder. And louder. And LOUDER TO THE POINT WHERE THERE IS NO LONGER ANY DYNAMIC LIGHT AND SHADE NO AURAL PUNCTUATION TO GIVE CONTRAST NO LIGHT AND SHADE TO PROVIDE RELIEF FROM A RELENTLESS ASSAULT ON LISTENERS' SENSES WHICH SCREAMS AT THEM IN THE SONIC EQUIVALENT OF UPPER-CASE LETTERS SO THAT IN THE END THEIR ONLY RECOURSE IS TO TURN IT DOWN OR EVEN OFF IN OTHER WORDS TO LEAVE THE CLUB.
This typographical analogy was used by Rip Rowan, editor of www.ProRec.com, in his essay on the subject, which I read just before I listened to this issue's "Recording of the Month," the 30th Anniversary rerelease of Pink Floyd's Dark Side of the Moon. As Jon Iverson describes in his review (p.131), the multichannel SACD remix has been done with care and taste, and the two-channel SACD version is a straight digital transfer from the analog master tape. But the CD layer...?
JI describes it as sounding "less detailed and more congested...more forward than the SACD layer, and much louder than my original Harvest/EMI (Japan) CD," though he did find that it substantially cleaned up the haze of the earlier CD release and added LF impact.
I'm not quite as charitable. Looking at the waveforms of the CD and SACD layers of "Money," you can see both that compression has been used and peak limiting applied to chop off the tops of spiky transients during the CD-layer mastering. Both processes allow the signal's average level, hence its apparent loudness, to be increased. This is not to the extreme degree of the Santana album: the average RMS power of the SACD RotM (calculated with Cool Edit Pro) is -17.7dB, that of the CD version -14.2dB. Spectral analysis of the two versions indicates no significant differences in EQ (though the LP has more HF energy). But the L/R balance has been shifted by almost a dB, a small DC offset has appeared, there are now 362 clipped samples, compared with none for the SACD, and there is evidence of clipping below 0dBFS.
Yes, it sounds congested. More important, "Money" is an exercise in graded dynamics. The relatively quiet opening cash-register riff is followed by declamatory verses that, while loud, leave room for the saxophone and guitar solos that follow to grow, with then a release before the final verse. This is what you hear on the original LP, the original CD, and the two-channel SACD layer. But the mastering of the new CD version makes the initial verses as loud as the solos, leaving the music nowhere to go. Yes, it's punchier. Yes, it's louder. But it's also less interesting, less subtle. Why do they do that?
In his online essay, Rip Rowan offers an insightful history of why the people who give us the music are taking it away. A more thorough explanation can be found in the new book from A-list mastering engineer Bob Katz: Mastering Audio: The Art and the Science ($39.99, 2002, Focal Press, ISBN 0-240-80545-3). In the readily readable manner familiar from the essays on his website, where this book can also be purchased), Bob takes us through all the steps along the road from session tapes to the CD in the store. "Making good sound is like preparing good food," he writes. "If you overcook, it loses its taste."
Bob devotes not just pages but many chapters, and even an appendix to how mastering engineers can avoid overcooking the music entrusted to their charge. Yet the reality is that the needs of the record company and casual listeners are incompatible with those of critical listeners. A skilled mastering engineer can produce a recording that meets both sets of needs, but only if he is allowed to do so by the people who pay his bill.
Rip Rowan concluded that "LOUDER IS BETTER is definitely a self-correcting problem. Because this stuff just plain sounds bad, and sooner or later (hopefully sooner) people are going to realize that the music doesn't 'rock more' or 'cut through better' but that it's just plain annoying." Bob Katz agrees: "We need to educate producers that fatiguing, hypercompressed CDs will not be auditioned more than once....Teach them that a decent amount of dynamic range helps make an album more enjoyable, lively, even clear in most cases, and that sound quality suffers as the average level goes up."
To judge by the sound of the DSotM CD layer, there is a lot to be learned.