How To Win the Loudness War

View Single Page

This story first appeared in Trust Me, I’m a Scientist on April 2, 2012.

Image courtesy of Flickr user Dennis AB

Image courtesy of Flickr user Dennis AB

Recorded music has been getting louder over for the past few decades. So much louder in fact, that everyday-listeners have begun to hear about it in general-interest papers like the Wall Street Journal, the Atlantic, the Huffington Post, and even the New York Times.

As records have increased in average level, so has the chatter surrounding these “loudness wars”. In some online communities, a shamelessly “hot” master is seen as a crime against fans. Some of the most zealous critics have even taken to publicly shaming mastering engineers or petitioning record labels to remix entire albums.

It’s easy to scapegoat the mastering engineers who oversee the final level of each album but in reality, mastering is a client service, and the engineer is only one voice among many. Somewhere along the way, “loud” has morphed to become more than a level — It’s now an aesthetic choice of its own, and has even transcended perceived volume.

This development has made arguments on both sides pretty messy, but the real solution remains simple. More on that later.

The Crusaders of Quiet

Ian Shepherd writes for the audio blog, and his most visible public role has been as a crusader for dynamic range.

Like many mastering engineers, Shepherd warns that when aggressive compression and peak-limiting are used to make tracks sound louder, it can also sap them of their life by flattening sounds, compromising musical dynamics, and introducing unwanted clipping.

He isn’t single-minded about the issue, and agrees that limiters and compressors can be used skillfully to improve “punch” and “impact”. To that end, he recommends that musicians and engineers learn to use specialized loudness meters to help them make more sensible masters.

These loudness meters allow users to measure the amount of dynamic range in their tracks, and compare it to the ratings of popular albums in the Dynamic Range Database. Shepherd recommends aiming for a minimum of at least 8db, and hopefully as much as 14db, of “DR”.

Shepherd has even created a public awareness event called Dynamic Range Day, which celebrated its third anniversary on March 16th. This year, he was joined by grammy-winning engineer Charles Dye, musician John Ralston, and production blogger Allen Wagner who helped advance the cause.

These three are the founders of, a non-profit out to set new standards in dynamics. They aim to create a certification process for artists who want to sidestep the loudness wars altogether and bear the “Turn Me Up” badge on their releases.

The organization has been active since at least 2007, but still has yet to certify any albums. It currently remains stalled in talks regarding what objective standards of loudness and dynamic range to adopt.

Deciding on hard-and-fast standards may be difficult for a few reasons:

First, there are some some inconsistencies in the DR ratings themselves. Some mind-bogglingly loud and peak-limited records by dance artists like Daft Punk and Skrillex often show a higher dynamic range than expected, thanks to their use of sparse, percussive arrangements and creatively-placed drop-outs.

Second, in all but the most extreme cases, it can be difficult to find direct correlations between DR numbers and perceived audio fidelity. We’ll explore that in a moment.

And third, although founding member Charles Dye is fairly conservative with loudness and dynamic range, some of his work, like his mixes for Ricky Martin, would fail to pass even Shepherd’s lenient standards of a minimum 8db/DR.

Once the group does get their standards sorted out, there’s still a chance they may have trouble gaining steam, as new trends in listener playback are already working to make their solution less relevant. As one mastering engineer and Mix Magazine contributor wrote, the loudness war may even already be “over.”

The Evolution of Ear-Splitting

The loudness wars have been around almost as long as recorded music. In the days of jukeboxes, artists and labels wanted their records to stand out among the other 78s, and engineers of the era pushed the medium to its very limit.

Motown in particular was renowned for putting out hotter records with each release in an effort to make home stereos scream. In the process, the tools they used to do so became an inseparable part of the sound and language of recorded popular music.

When the CD was introduced in 1982, some hoped it would bring dynamics back to music, but all it really did was give tonophiles a new threshold. 30 years later, we’ve finally found the point of total saturation.

For all the hand-wringing, average levels on CD have effectively peaked in recent years. Like many others, I believe that there’s only one place left to go – and that’s back down. But it’s important to realize that it won’t be that way for all of us. And it won’t happen all at once.

The Changing Language of Music

Music is a language, and artificial effects are its buzzwords and slang. They’re once-meaningless sounds that can say a lot – often about where you come from and who you relate to.

Heavy compression, limiting, and even pumping and clipping may have once been accidental side-effects of some kind of loudness war, but today they’re part of the musical vernacular.

Audio engineers talk a lot about the loudness wars of the recording industry, but we often forget that there was once a parallel set of “wars” between radio stations. The battles on that front may have been key in helping to prepare listeners for the most radical changes to come in the recording industry.

In the last quarter of the 20th century, radio stations continuously upped the ante – far ahead of the recording industry – by using heavy-handed compression and limiting to to help their broadcasts jump out on the dial.

As young listeners (now accustomed to the sound of heavy compression on the air) grew up to make music of their own, songs just didn’t sound like hits until they could match the sense-memories of their favorite music on the radio. A similar acclimation likely occurred again with the next generation, raised on the hot CD and MP3 libraries of their big brothers and sisters.

Peak-limiting is a sound that’s now considered normal, and has even become expected in many circles. Although though Charles Dye’s work with Ricky Martin makes significant use of compression and limiting (and would fail even Shepherd’s dynamic range criteria) it’s easy to admit that those songs sound exactly how they’re supposed to.

What’s interesting is to realize that the same holds true for many records that sound significantly louder and more aggressively clipped.

Settling on Standards

Guidelines can be helpful, but in practice it’s often difficult to draw a direct correlation between dynamic range and perceived sound quality.

When Sonic Youth‘s Rather Ripped was released in 2006, it was praised for its “hi-fi” and “punchy” sound — even though the Dynamic Range Database gives it a “bad” rating of 7DR. I don’t know of many SY fans who would use those two adjectives to describe the band’s 1987 album, Sister, even with its DR rating of 11.

Similarly, few listeners would argue that early Flaming Lips records like Clouds Taste Metallic (9DR) and Transmissions from the Satellite Heart (10DR) sound “better” than their later classics, Soft Bulletin (6DR) and Yoshimi Battles the Pink Robots (5DR).

The latter two are albums that sound loud, slammed, and beautiful. On Yoshimi, producer Dave Fridmann uses aggressive compression and peak limiting to drive some sounds to the point of intentional clipping and distortion. With a well-crafted album like this one, it becomes clear that — in at least some cases — records aren’t “clipped because they’re loud”; They’re loud because they’re clipped.

Pages: 1 2Next Page ❯View Single Page

  • Just this past AES show, at a panel with the mastering engineers, it was Tom Coyne who said it best. Just within weeks of release, after he had tastefully mastered the first release of a new artist, and had the record label sign off on it, someone from that label caved under the “loudness wars” pressure. He was advised to bring the levels up, and then had to have an assistant correct for all the “overs” that were subsequently created. He finally got the label’s approval for release again. This artist goes by the name of Adele.

  • No, whether a recording is slammed or not doesn’t determine whether it sounds good. But the exact same recording with peaks left intact would absolutely sound better. It’s easy to compare slammed re-releases and see which sound better. It’s nearly always the original unless the original release was flawed in some way.

    If peak limiting was a desirable sound rather than just a way to bring up levels, then why do all vinyl releases not have peak limiting?

  • FlameBait

    This appears to be correct. Thanks for the heads-up, Mr. Sears.

  • FlameBait

    I refuse to pay for overmodulated music. I use the Dynamic Range Database as my guide.

    To clarify: I may enjoy this new song. I may stream it. I may put it in my YouTube playlist. I may use a device, scheme, process, or app to acquire my own digital copy/version of this new song. But I will *not* pay for it if the final production causes me ear fatigue. (Are you reading this, Daft Punk?)

    So if you want me to transfer some of my income, to your account… clean it up. Now.

  • FlameBait

    Sadly, Mr. Coyne has left us. Quality-minded music aficionados will have to find more working sound engineers to take a stand against brickwalling.