Loudness War

aka: Record Of Loudness War

Top: Dynamic range. Bottom: Dynamic pain.^note Chinese Democracy was actually released four years after American Idiot, but the engineer apparently knew what he was doing.

"I like my music like I like my life — everything louder than everything else."

— Meat Loaf

Before we begin, this article isn't about cranking up your stereo; it's about the music itself and the media it's etched or printed onto, and how it can make the act of cranking up your stereo have no effect. This non-effect is called clipping and is usually a result of excessive loudness.

Ever since the dawn of humanity, it's been a known fact that people respond more easily to louder audio stimuli. Thus the illusion that "LOUDER = BETTER". With the advent of music recording technology, the music industry sought to capitalize on this; if our album is louder than their album, and the listeners don't adjust their volume dials to compensate (or can't, if they're hearing it in a bar, a gramophone with no volume dial, etc.), our album will sound better! However, vinyl records suffer from physical limitations^note for example, too much bass can bump the needle out of place, and too much high end can destroy the printing needle when creating a record! that necessarily constrain the physical (and perceived) loudness of a record. The search for loudness became easier with the introduction of the compact disc in 1982; now working with a digital storage medium, sound engineers found themselves liberated from the former physical constraints on perceived loudness.

Any recording medium has an absolute limit when it comes to amplitude, and compact discs are no exception. Once audio engineers managed to max out, the only solution to out-loud those other CDs was to break out the dynamic range compression, which squashes every bit up to the same volume level, causing listener fatigue.^note In addition to dynamic range compression, sound engineers can use overdrive distortion (or outright digital clipping) to further increase the average amplitude of a track. Another trick is to increase the "perceived loudness" of a track by emphasizing frequency ranges the human ear is more sensitive to (highs and high-mids) and de-emphasizing others (very low sub bass and sometimes mids). The often-used analogy here is that of attempting to read a text written entirely in capitals in a huge font. In movies, Orange/Blue Contrast has a similar effect; since the contrast is pleasing to the eye, producers will push at as far as possible, often over-saturating the film in these colours and squashing out more subtle colours entirely.

Compression is not inherently bad. In fact, it is an essential part of the sound engineer's toolkit, used to level out the inconsistent volume in a track.^note This is practically guaranteed to happen when recording live material. Especially bass guitars. Furthermore, it can be used creatively to manipulate the sound of a single instrument^note especially drums, where compression is often used to emphasize the body or tail of a strike or even a full mix, "gluing" the tracks together. However, if used ineptly, it can result in unpleasant changes like undesired pumping, outright distortion, or a more abstract loss of "life" or "impact".

It may be worth noting that, due to the general trend of increasing loudness of commercial releases over time, a release that would have qualified as an example of this trope in 1995 would not necessarily qualify as an example of this trope in 2005 or 2015 (unless it's by Oasis). Similarly, an album that would qualify as an aversion of this trope now would not have qualified as an aversion in 1995. An album that comes out to DR8 generally qualifies as good mastering by today's standards, but would not have been two decades ago. When looking at examples of this trope, it's probably worth noting that they're graded on a curve, so to speak.

Note that this is a separate issue from the loudness or intensity of music by genre. When loud music is used as a weapon, that's Loud of War. See also Ridiculously Loud Commercial for similar techniques used in advertising.

---

Example subpages:

• The Worst Offenders

Other examples:

Compression for stylistic reasons

• One of the first examples is The Rolling Stones' single "Have You Seen Your Mother Baby, Standing In The Shadow". Although it's never been explained why, contemporary recording standards suggest that it must have been deliberate.
• One of the most infamous examples is Iggy Pop's 1997 remaster of Raw Power by The Stooges. At DR1, the album is constantly in the red and occasionally averages -4 dBFS ("Your Pretty Face Is Going To Hell" in particular goes down to DR0). The album was intended to sound loud, blunt, and unsubtle, but there's a difference between "exhilaratingly loud" and "headache-inducing, unlistenable sonic mess;" for reference, the previous CD edition, which used David Bowie's mix from the LP release, clocks in at DR11. Even his bandmates James Williamson and Ron Asheton think he fucked up horribly here. This is often cited as an example of an artistic use of the phenomenon, as Iggy Pop has been cited as saying he wanted to "recreate the feel of an old, worn-out vinyl". Given the overwhelming amount of crackling, distorted midrange on the disc, it's easy to say this goal was at least somewhat accomplished. However, given that the results often cause listener fatigue, most people wish he had used some other method of creating this feel.
  • There is hope, however: Sony released the "Master Edition" in April 2010 with the original David Bowie mix reinstated and slightly beefed-up to fix some of the valid criticisms of his original mix (inaudible bass and the like), clocking in at DR10. It just might be the first step towards reversing the trend in general. Even better was a 2012 vinyl remaster, which is discussed further under "Aversions".
• Strapping Young Lad invokes this for the sole purpose of creating a gigantic and unrelenting wall of sound. It's unpleasant, but that's the point. (Note that while all SYL CDs are horribly brickwalled, only Heavy as a Really Heavy Thing and The New Black are clipped). Averted somewhat on the vinyl editions, which are significantly quieter.
  • Devin Townsend's solo album Terria is notable for having a very compressed bass drum which gives the album a unique feel. Ironically, this is one of his quietest masters.
• Many acts on Ed Banger Records, like Justice, SebastiAn and Vicarious Bliss, who use absurd amounts of compression◊ to create a signature sound, and in Justice's case, using digital noise and clipping as an instrument.
• Virtually any recent act that psych-rock producer David Fridmann has worked with, like Flaming Lips, MGMT and Tame Impala. A prominent example is MGMT's Oracular Spectacular: They asked Fridmann to turn up the gain to make it sound "dirty", turning "Kids" in particular from this to this.
  • In defense of Dave Fridmann, Kliph Scurlock, The Flaming Lips' drummer has this to say, implying that it's the bands who push for brickwalling instead of Fridmann himself (taken from this post):
  "I've seen things written here and there that it's all Dave and that he a 'habitual ruiner' of recordings, but I'll go on record right now as saying that Dave never tries to 'pull a fast one' over on bands and we know exactly how things sound when we leave his studio. In fact, he's fought for this song or that song to be quieter overall on more than one occasion."
• The alternative rock band Sleigh Bells went for an intentionally compressed sound, although their guitarist Derek Miller later admitted that his own music's loudness was annoying to him.
• Queens of the Stone Age's Songs for the Deaf. The entire album is supposed to sound like it's being listened to on a car stereo.
  • For the curious: album gain -10.88 dB. And Rated R two years earlier has an album gain of -9.92, which casts doubt on the "car stereo" theory.
    • This is likely due to the fact that Songs for the Deaf has an extremely dark, bass-heavy mix, and ReplayGain uses loudness contour analysis to determine how loud an album actually sounds vs. how loud it actually is. Lower-end frequencies don't trigger higher values nearly as much as upper-midrange signals—Songs for the Deaf is notably abundant in the former and lacking in the latter. Recent updates to the ReplayGain system have leveled out in this area, becoming closer to the absolute-volume RMS standard.
• Intentionally invoked by a number of Noise and early Industrial artists, most notably by Whitehouse on Birthdeath Experience (1980) and Right To Kill (1983), on which everything is "in the red". Hell Is That Noise, indeed. Averted by legendary harsh-noise artist Merzbow, whose production tends to be frighteningly clear (though see below).
• Neutral Milk Hotel's In the Aeroplane Over the Sea, inexplicably, is really badly clipped, especially considering it's a folk album. May be a side effect of the lo-fi aesthetic Jeff Mangum was going for (the band actually lists "white noise" as a musical instrument), but it's still really, really weird. Despite the fact that a rather substantial portion of the album is just acoustic guitar and voice, if you look at the zoomed-out waveform of the album there are almost no dynamics to be found.
• Experimental hip-hop group Death Grips. Exmilitary, their debut album/mixtape has an album gain of -13.04dB. One song, "Spread Eagle Cross the Block", has an insane ReplayGain value of -14.64dB. No Love Deep Web, their third album also uses very heavy distortion but not quite to the extent of their debut (album gain of -11.62dB, the loudest song at -13.13dB). Their other album, The Money Store is still loud but not as loud as the other two.
  • They've also released the stems for Exmilitary on their website (as Black Google) to encourage remixing (with the side effect that people can make their own de-loudness war'd mixes) and released an "unmastered" version of The Money Store which is nowhere near as loud.
• Japanese noise rock band Guitar Wolf intentionally try to be as loud as possible. You have been warned.
  • Incidentally, they apparently aren't too fond of the loudness war, because back when they started out, their music being as loud as it was made it pretty much unique, whereas now their music sounds almost normal. It's also worth noting that Jet Generation came with a sticker saying, "Warning: this is the loudest album ever recorded. Playing at normal volume may cause irreparable damage to stereo equipment. Use at your own risk."
• Ulver's Black Metal album Nattens Madrigal has a ReplayGain value of -16.22 dB, which indicates a gigantic amount of volume compression (especially considering that there are probably five minutes' worth of ambient passages between the tracks dragging the values down somewhat). This was likely a deliberate aesthetic choice to make the album's production even colder. It's also worth noting that, while Nattens Madrigal is notorious for having horrible production (there is an Urban Legend that it was recorded in a forest, which the band has dismissed as impossible), the band took care to avoid any clipping when mastering it. Unfortunately, they were not as careful with the band's later album Blood Inside (recorded in a completely different style, as Ulver are wont to do), which is significantly less compressed but still has clipping. Most of the rest of Ulver's discography averts this trope, being mastered at more reasonable levels.
  • The remaster of Nattens on Trolsk Sortmetall averts this. Unfortunately, the version of Bergtatt on the same release is louder than the original.
• Folk Metal band Moonsorrow consistently use compression and distortion to give a larger than life sound to their albums. This is most noticeable on their debut Suden uni, their 2007 album V: Hävititty and their 2008 EP Tulimyrsky. The albums have been remastered and re-released on vinyl, where they are still clipped, but it's less severe/noticeable. They have abandoned this with their latest album, Jumalten aika, which comes in at DR9 and sounds great.
• Frequently intentionally invoked by Hellektro bands such as Suicide Commando, whose Implements of Hell album is almost solidly brickwalled as well as having horrendous clipping on par with Death Magnetic.
• Subverted with Rachael Please, whose tracks are ridiculously clipped in order to get the effect he wants.
• Intentionally used by most dubstep artists, including Skrillex(except for the vinyls) and Bassnectar.
• While the production on Merzbow albums tends to be frighteningly clear, in the mid-90s he made what are quite possibly the two loudest albums ever in Venereology and Pulse Demon, whose dynamic range comes down to DR0 and which contain what is generally held to be the loudest track ever in "I Lead You Towards Glorious Times". Given that the majority of his releases are much more clearly mastered, it's safe to say this was intentional. To be clear, many (though by no means all) of his later recordings are still pretty loud, but not that loud.
• Venetian Snares & Speedranch's 2001 release Making Orange Things destroys the competition when it comes to brickwalling. One of the tracks, "Pay Me for Sex", has a ReplayGain of -19.92dB and is currently one of the least dynamic albums ever released according to the unofficial dynamic range database (a couple of Merzbow's mid-90s releases, discussed immediately above, are even lower). In fairness, this is very much an intended effect as the album is made up of very distorted synths and harsh noises.
• New York-based Psychedelic Rock band The Psychic Paramount's CDs are generally some of the loudest ever. Gamelan into the Mink Supernatural is DR0 and only a shade quieter than Venereology (see a couple of entries above). Their other CDs aren't quite as loud, but still usually in the DR3-4 range (Origins and Primitives, which contains a number of acoustic tracks, is better, being overall DR6). As they are also a Noise Rock band, it's safe to say Artistic License probably applies.
• The unusually low dynamic range of David Bowie's ★, coming in at just 5, is suspected to be a case of this. "DR5" happens to be an abbreviation for the surface receptor protein death receptor 5, which mediates apoptosis (cell death). Monoclonal antibodies targeting this receptor protein have been under testing as an experimental treatment for cancer patients; it's likely that the choice to master ★ so restrictively was done primarily as a reference to this treatment, seeing as how Bowie himself was dying of liver cancer during the album's production.
• Deliberate over-amplification is part of the voice treatment for the Daleks, to help create the idea of a deranged alien mutant locked in a tin can, screaming into a microphone millimetres away from its mouth.
• Low's 2018 album Double Negative features very obvious distortion as part of the album's aesthetic. Two of the songs are DR0 (the ending of Tempest, in particular, couldn't be more clipped if the band wanted to). Unlike many other extreme examples, it's a mostly mellow listen.
• This trope is a signature element of the somewhat awkwardly-named "Ear Rape", in which a portion of audio suddenly sounds extremely compressed, with overbearing clipping and distortion, to sound as loud and unpleasant as possible without outright destroying one's speakers.
• In the Vinesauce fangame Vine Realms, Pantherk's sole voice clip (consisting simply of him saying his own name) is heavily compressed to the point of audibly clipping, nodding to a moment in one of Joel's streams where he modded the Mortal Kombat fangame the clip was sourced from by boosting it to the highest possible volume. Here, the poor mastering is similarly exploited by having Pantherk's voice be the loudest thing in the entire game, much louder than every other bit of audio present.
• Psych / Garage Rock musician Ty Segall's 2008 self-titled is absurdly loud on purpose. His albums on Drag City are far more reasonable.
• Both versions of the rock instrumental Telstar were extremely compressed on purpose.
• Hyperpop, the online electronic music genre and associated subculture, is known for extreme amounts of compression on everything. Most musicians make their songs full of dense layers of electronic noise, so the loudness is stylistically appropriate, if sometimes a little overbearing.

Amusing casualties

• Because the dynamic range of most digital music is squeezed right near the top of the format's capacity, the available volume settings in many modern mobile devices such as the iPod are literally incapable of playing correctly recorded digital audio at the proper volume. This means that even with the internal amplifier at maximum volume through the included headset, what should be normal passages are quiet, and intended quiet passages are inaudibly soft.
• It seems half the comments on the YouTube showing of Star Trek: The Original Series (only available to Americans) are about the advertisements being way, way, WAY louder than the low-levelled audio. Video advertisements being infamously louder than TV programs (with much of the same brickwalled audio quality that even the uninitiated can recognize as 'sounding like a commercial'), and the TV program being leveled at about what you might get in the 1960s (or maybe a DVD), suddenly the advertisements get perfectly normal people screaming to all-out boycott the products advertised. A similar issue comes up in Australia every few months.
• Recent NHL Hockey games have touted being able to import custom audio for use as in-game music or sound effects. Unfortunately, the pre-existing audio is amplified so much that the feature is nearly useless. While you can increase the game's music volume in order to hear the imported songs, not all of the pre-existing sounds can be replaced, meaning you run the risk of blowing out your eardrums.
• Grand Theft Auto V's Self Radio custom station added in the PC version is badly wrecked by the game's horrendously compressed audio. Depending on the song, it can either be quiet to the point of being drowned out by engine noises or louder than the game itself!
• Games in general can be subject to this if the programmers don’t think about headroom when writing the audio management code, even if the individual sound effects aren’t themselves clipped. It works like this: games have to play multiple sound effects at once, so they end up stacked on top of one another. But if you just put in sounds that are normalised to -0.1 dB without adjusting the levels, they’re all already close enough to the 0 dB line at their peaks and troughs, so you’ll wind up with nowhere to go when you have more than one or two sound effects playing (or even, potentially, one sound effect playing on top of music), which results in an integer overflow – in other words, clipped game audio.
  
  The solution when making sound effects for a video game is to leave a lot of headroom with each sound effect (and the game soundtrack), so that when they stack on top of one another (which is inevitable, particularly in some game genres), they won’t clip too often. It’s probably impossible to stop the game audio from clipping at all, but the less often, the better, of course. The ideal solution is to simply program the game to reduce the levels of all the sound effects (and the soundtrack) by 12 dB or so before putting them together. One other solution, if the game’s programming isn’t that sophisticated, is to normalise sound effects to a much lower level than one would if were mastering for CD or the Web. This means a bit lower signal-to-noise ratio, but if you use 16- or especially 24-bit audio, this shouldn’t be a problem; even uncompressed 44.1-kHz, 24-bit sound effects won’t be too large compared to most of the graphical assets used in most modern games. (Game audio is also one case where 24-bit audio may actually be something more than a placebo, since the game has to adjust sound effect levels on the fly; adjusting the levels of dithered audio again reduces the effectiveness of the dithering to some extent, and 24-bit audio doesn’t need to be dithered because it already reproduces more dynamic range than humans are capable of hearing.)
  
  This has long been a problem affecting the video game industry; for instance, Marathon (1994) and its sequels (1995 and 1996) clip pretty badly whenever the player is battling enemies. Some of the sound effects are also themselves clipped, which was for stylistic reasons in certain cases (the VacBobs from the third game, who are supposed to sound like they’re talking over a crappy radio system), though by no means all.

---