Syndication
Articles
Newest Mics
Metallica Wins the Loudness Wars
Saturday, December 20th, 2008 | by matthew mcglynn
There was a great End Rant in Tape Op about the latest Metallica album, Death Magnetic. Mastering Engineer Paul Abbott of Zen Mastering wrote,
The outcry over the un-listenably loud level of Metallica’s Death Magnetic CD has become one of the more visible occurrences of the “loudness wars” in the mainstream media… now, a lowest-common-denominator threshhold has been crossed and the problem has been pushed in our faces.
I don’t pay much attention to the mainstream media, and I don’t pay any attention at all to Metallica — in fact it’s safe to say that I get 100% of my Metallica news from Tape Op, which thankfully means I don’t get very much at all. So I’d missed the outcry. Hell, I’d even missed the release of Death Magnetic.
The story caught my interest. As Paul noted in his rant, the loudness of the CD was made more notable by the near-simultaneous release of an alternate mix — with much more conservative limiting and compression — via Activision’s “Guitar Hero” game. Suddenly, A/B comparisons illustrating the casualties of the loudness war can be easily made. For example, see below.
An associate of mine provided loaner copies of both mixes of the album. A few minutes of research at Last.fm revealed that the most popular track is “The Day That Never Comes,” so I loaded both versions of that song into Pro Tools.
Visually, the difference is obvious. Once the drums come in, the audio level is practically pegged at 0dB for the rest of the song. There are no more dynamics to be had.
Sure, it’s a metal song. Maybe it’s just really that loud, you might be thinking. But no, the Guitar Hero mix shows plenty of dynamic range. And, not to spoil the surprise, it sounds a lot better too.
To be clear, the core problem with the Death Magnetic CD is not that it’s loud. The problem is that it sounds like crap.
Let’s do some comparative listening. There are two ways to do it: matched volume levels (in which the two versions sound equally loud), and matched gain levels (in which both versions are output at unity gain, and the CD version is a lot louder).
 |
Equal Volume: 320 kbps MP3 Caution! The second half of the “equal gain” clip is much louder than the first half! Turn down your playback volume! Both clips contain an excerpt of the Guitar Hero mix followed by the same section of the CD mix of “The Day That Never Comes.” |
Here’s a picture of the waveform for the two mixes of this excerpt. The Guitar Hero mix (on the bottom of the screenshot) shows moderate changes in dynamic level, and plenty of transients. The CD mix, on the top half of the screenshot, is brickwall-limited — the entire segment is at maximum volume.
To match volumes for the “Equal Volumes” clip, I used a VU-meter plug-in in RMS mode, and reduced the gain on the CD mix until the two versions, panned hard L and R, were closely matched. The gain difference was staggering: 10.7dB! (see screenshot) A 10dB change requires 10x as much power to produce, and equates to roughly doubling the volume.
The “Equal Volumes” clip contains the Guitar Hero excerpt (at unity gain) followed by the CD excerpt (at -10.7dB). You won’t notice a change in playback level, but you will hear distinct differences in the quality of the sound. The distortion on the CD mix is apparent. Every snare drum fill makes my headphones sound broken. The CD mix is aggressive, for sure, but it doesn’t sound good.
Zooming in to examine the waveform, it becomes obvious why the CD mix sounds the way it does. The relatively smooth waveform of the game mix is replaced by a jagged, exaggerated line. The rounded forms of the game mix are gone — there is nothing round about the CD mix, which is comprised almost entirely of vertical lines. The CD mix is comprised entirely of transients.
Curiously, the CD mix even has transients where the game mix does not — see the vertical spikes in the top waveform that have no corresponding spike below. That’s just noise: artifacts of the compression and limiting process.
Consider that the vertical changes in the waveform correspond to vibrations in the air, and therefore to the movement of a loudspeaker cone. The CD mix requires that your speakers spend the entire song oscillating between their maximum and minimum excursion points. Your eardrums, too. No wonder the CD mix is so fatiguing. Listening to the CD mix is literally a lot of work.
Paul Abbott, in his Tape Op piece, quoted a statement by Death Magnetic mastering engineer Ted Jensen in which Jensen claims no responsibility for the sound of the CD. I found the origin of this statement — a private email attributed to Jensen, posted without his permission to a Metallica forum and quoted extensively around the web.
I wouldn’t have been inclined to pin the blame for the CD mix on the mastering engineer, anyway; the few MEs that I know are uncompromising in their pursuit of fidelity. The only way they would output something sounding as bad as Death Magnetic is if the mixes they were given sounded worse. Which was actually Jensen’s point.
Lars Ulrich
Listen, there’s nothing up with the audio quality. It’s 2008, and that’s how we make records. [Producer] Rick Rubin’s whole thing is to try and get it to sound lively, to get it to sound loud, to get it to sound exciting, to get it to jump out of the speakers. Of course, I’ve heard that there are a few people complaining. But I’ve been listening to it the last couple of days in my car, and it sounds fuckin’ smokin’.
There you have it.
(Hat tip to Aaron Lyon for suggesting this story, and to TurnMeUp.org for general awesomeness.)
Tags: lars ulrich, loudness war, mastering, metallica, paul abbott, rick rubin, tapeop, ted jensen
Posted in Uncategorized |
You can leave a response, or trackback from your own site.
Brandon
December 25th, 2008 at 4:03 pm
It must be nice to have the certainty that Lars seems to have about everything. Of course, you’ll forgive me if I’m a bit leery of leaving judgments of fidelity to someone who’s been playing drums at stadium volume levels since before the advent of “quiet” in-ear monitoring.
He and the rest of the band are certainly well within their rights (and perhaps *are* right) to stand behind their record. But his pronouncements since the Napster kerfuffle all too frequently imply a derision, maybe even contempt, of the fans who have made Metallica a world-renowned band.
Aaron
December 30th, 2008 at 9:36 am
Really nicely done, M. I want to see the game version waveform next to the zoomed CD waveform, which is astonishing! Your observation about why it’s so much work to listen to a mix like this is spot on. I was surprised earlier this year to find Springsteen’s new album similarly smashed and instantly unpleasant because of it. Likewise the recently released Oasis CD. Haven’t looked for RockBand versions of these…
Julian
December 30th, 2008 at 8:11 pm
What I don’t understand is why there are two different mixes (it must be two different mixes if the trashing didn’t happen in the mastering) in the first place. I mean, if Metallica like their “new sound” or what Rick Rubin has done to their music so much, why did they agree to release a different mix for Guitar Hero??
matthew mcglynn
December 31st, 2008 at 3:13 am
Aaron, the zoomed-in waveform pic is there; just click the thumbnail to see it. or, here: http://recordinghacks.com/wp-content/uploads/2008/12/waveforms.png
Julian, it’s a good question — I thought at first that maybe the mix engineer produced one with and a second without such heavy limiting/compression, but it seems from some comments posted elsewhere that the mixes are actually different. Maybe nobody considered that the GH version would be something people actually listen to, meaning, it wasn’t given the same scrutiny by the producer?
Aaron
December 31st, 2008 at 8:53 am
Ah! Now I see the expanded screenshots. I missed that. UI suggestion: link color border (burgandy) for hot images, grey border for not. Cheers, -a
S Jansen
January 17th, 2009 at 10:51 am
Finally, an “in-yer-face” example for the general public of what many of us (who actually care about sound quality) have been lamenting for several years!
I can *almost* go along with the handful of producers who say that the mixes today have to be compressed like this in order to be heard above the (noisy) party or automobile environments. But not really.
I think the simple solution is to offer two mixes to the public: a “quality”/audiophile version, and a “fouled up”/party version. The industry could offer two separate cd’s (remember way back when one had the choice of a mono or a stereo record album?), or they could simply stick the two different cd’s in the exact same case. “Here’s one for your home stereo (quality), and one for your car to take to that rowdy party later (fouled up).”
Double-sided cd’s?
But, of course, the industry doesn’t care about giving us what we want, or giving us actual choices…it seems that this band doesn’t, either.
I can almost picture Lars whining, “Those internet pirates are just gonna turn this album into mp3s and ruin the great sound and rip us off - let’s ruin the sound before they even get the chance! That’ll show’em!”
Yup - they showed everybody, alright. Just look at those naked, ugly square waveforms. Ugh.
John S. Allen
February 22nd, 2009 at 5:13 am
Nice demo. I’m a member of the Boston Audio Society, which addressed this issue in a recent issue of its publication, the BAS Speaker. The limiting in the Metallica recording is far more aggressive than anything illustrated in the Speaker article, too.
One quibble, You say that “The CD mix requires that your speakers spend the entire song oscillating between their maximum and minimum excursion points. Your eardrums, too. No wonder the CD mix is so fatiguing. Listening to the CD mix is literally a lot of work.”
More to the point, clipping the waveform introduces new frequencies into the mix. These are the “crap” which makes the snare drum go “splat,” and makes listening hard work by obscuring the sounds of the instruments.
Radio stations maximize volume so they pop out of the road noise in your car and you won’t tune away. That’s understandable if not forgivable. But the CD experience is typically not one of channel-flipping, so what’s the point? Every playback system has a volume control. Most listeners are skillful enough to set the volume as they like.
The dynamic range of the Guitar Hero mix isn’t all that wide, either. The peaks even in the quieter sections are only a few dB down. It sure does sound better, though.
David
September 27th, 2009 at 9:11 am
Lars, sorry dude — you’re just wrong here. I’ve been a fan for years, but you can never justify loudness and clipping like that. It’s an insult to the faithful who buy your music. It’s a big “F” you, we don’t care what you think. The loudness war is greed. It’s engineers tripping over themselves to out-shout the other guy, the other song. It’s sacrificing the art for a bit more money. So some kid isn’t “disturbed” by having to reach down to his nano to turn up the volume a bit. Absurd. This is a great article, well presented and the second time I’ve stopped in to re-read it. A must read for every musician and every (open-minded) sound engineer. IMHO.
John
November 14th, 2009 at 3:07 pm
For cryin out loud guys…the Guitar Hero mixes are NOT MASTERED…Hello? They come packed as separate tracks–otherwise you couldnt play the game. I think even the kids know they are stems–how people in the business dont know this is beyond me.
Alex
January 1st, 2010 at 8:09 am
John, mastering a recording should make it sound brilliant, not turn it into a square-wave ear rape like the entire Death Magnetic album sounds like, it has NO dynamic range at all. Try listening to Pink Floyd who where way ahead of their time in mastering music with fidelity, it sounds absolutely brilliant and so much dynamic range. Majority of their songs do not go higher then -20 dB average which leaves room for very powerful sound effect. Death Magnetic is pretty much at 0 dB with no headroom, no dynamic range at all and the audio hurts my ears beyond belief. It’s one of the few albums I’ve heard that makes my ears feel tired and hurting.
Mastermeister
February 13th, 2010 at 4:44 pm
From many things I’ve read and from interviews, it’s clear that Lars fits the clinical definition of a Malignant Narcissist. Owing to that and years of stage sound levels far in excess of permanent hearing loss because he’s too vain to wear hearing protection, he is most likely also deaf, meaning that damn near anything turned up loud enough so that speakers emit smoke, will think “it sounds fuckin’ smokin’.”
Mastermeister
February 13th, 2010 at 4:51 pm
Alex has a point worth re-pointing: dynamics is a tool, a color on the palette of musical expression. To abandon it is to toss out an important musical element of expression. A good analogy might be to go on stage drunk and smacked out on speedballs. I did live sound once for Townes VanZandt, and watched as he nodded off on stage and fell off his stool. It did not earn him any praise for genuineness.