This goes to ele—augh, no, aside from over-compressing, we need to stop overusing that joke. Photo (CC-BY) Orin Zebest.

You’ve heard the gripes, and heard and seen the somewhat unscientific demos. Now it’s time to examine the over-compression of music with – science! Earl Vickers of STMicroelectronics examines the Loudness Wars in an academic paper, as noted to us by reader photohounds.

The Loudness War: Background,
Speculation and Recommendations
[PDF Link,]

The paper comes from last November, but it’s as relevant as ever. It’s not just the usual take, either. Its history begins with Phil Spector and vinyl, considering the impact of broadcast TV and not just the music industry. It notes the evolution of compression technologies, particularly multiband technologies.

Most importantly, though – and I’ve spoken regularly to mastering engineers about this – the paper turns to the issue of listening fatigue. Here’s one whithering criticism of the industry on that: some engineers even believe that thoughtless over-compression could be to blame for the decline of the entire industry.

Mastering engineer Bob Ludwig stated, β€œPeople talk about downloads hurting record sales. I and some other people would submit that another thing that is hurting record sales these days is the fact that they are so compressed that the ear just gets tired of it. When you’re through listening to a whole album of this highly compressed music, your ear is fatigued. You may have enjoyed the music but you don’t really feel like going back and listening to it again.”

2008 Metallica, unsurprisingly, more apocalyptically-loud than a 1909 Edison cylinder … for what it’s worth.

You’ve seen much of this before, but rarely in such well-annotated, comprehensive form.

Best of all? The conclusion applies lessons from Game Theory to work on making the loudness wars come to a conclusion.

Here’s another thought, too: with artists increasingly self-releasing or releasing through more specialized labels, greater access to music online, direct-to-consumer distribution, and online replacements for conventional terrestrial radio, many of the factors that produced some of the oddest hyper-compression at the top of the charts are fading into the background.

Pax Musica for the loudness wars, anyone?

  • Simon

    This is so damn true …

  • Simon

    Maybee the iPods headphones are also hurting record sales. I Think the bad ass quality headphones need highly compressed material cause' any other, not compressed material does sound bad on that bad quality headphones.

    What if everyone would have B&W N803D Speakers on top of his head … ??

  • Rob Henke is making a conscious effort to fight this tendency. His last record as Monolake, called Silence, Has very little compression, almost none, and he explains his mixing technique in a thorough essay on the album's liner notes, and also on his website. 

    We probably all agree we shouldn't overcompress, but when mastering time comes, we all secretly pump it up in fear of being softer than your colleagues. Some periferal genres do not need this, but if you are still using the classical kick, snare, bassline and mid-high melodics configuration, you're in for the loudness war.. And this includes many more genres than I can count (even IDM).

  • Peter Kirn

    Right, but it's worth reading the whole PDF — loudness per se doesn't necessarily have to be bad; the issue is working out how much is too much in regards to listening fatigue.

  • Combined with crappy MP3 and no wonder the ear is feeling abused!
    I have stopped pushing my music into a corner and started trying other techniques. Much happier about the results πŸ™‚

  • I've made this point here before and I'll make it again: It's the shit music that gets over-limited. If you avoid bad pop-rawk, commersh metal, ticky tacky R&B, bling-centric hip hop, electro-pop and Swedish hard techno, you'll only rarely encounter over-limited music.

    What I really want to stop is that over-the-top side chain compression of melodic material keyed to the kick sound. Just effing leave it, idiots!

  • Andy

    Yeah, it depends on music style and target audience. Bad quality players are also an important factor. Personally, I react very sensitive to over-compressed music. It becomes painful and exhausting to listen to it very quickly. When I check my own tracks in the mixing stage and something sounds wrong and unpleasant then it often helps when I just reduce the compression.

  • nylarch

    >>>>It’s the shit music that gets over-limited.

    Not necessarily.   Flying Lotus records are both brilliant and so over compressed I have a hard time making it through a whole record (a bit better mix on Cosmogramma).

  • it's not a joke, you see, if it goes to 11 then it's louder…

  • I'm looking forward to reading this thesis. Thank you for posting it. It analogously does follow the prisoner's dilemma schema. Remember, this is for non-repeated games, no history, and seemingly no grim trigger for future iterations. And information will not remedy the problem in a one-instance game: people can change their minds on a whim.

    I'm looking forward to seeing Earl Victers' recommendations. One thing that comes to mind is creating a market for natural dynamics: radio stations that feature it (and don't make exception during commercial breaks), XM satellite channels, labels-like on food ("no HMMO's, no sodium lauryl sulfates etc."), and unplugged concert going.

    On this last note, I'd argue that it is our relation to music that has become warped, and a lot of this goes back at least as far to the proscenium (they go farther back for theatre than for orchestra) replacing earlier concert halls and the expansion of the orchestra to accommodate pieces like Berlioz' New World and to dissolve socializing during the performance among the elite classes in France. Much of the demand to change the earlier architecture comes with scaling. If we could rid ourselves of the bigger is better and too big to fail economic pressures, then you wouldn't have loudness wars.

  • I look forward to reading the paper but want to contribute this theory: 
    The compact disc itself has contributed to the decline of music sales, especially in regards to long form (album) sales. This isn't a 'vinyl sounds better' argument but instead I'm focusing on the length of the modern album. Until the cd came along albums were limited to 35-40 minutes. Occasionally there would be double albums but even The Beatles couldn't pull that off without some iffy material. 
    Even if not every song on a 35 minute album wasn't the greatest you usually had a decent quality/filler ratio. Now enter the cd, 50-60-70 minutes starts happening, and  people start expecting it, they feel ripped off if a disk is 35 minutes. But how many bands can fill 60 minutes with quality material? Even with bands I like once they get beyond 11 songs I'm ready for something else. 
    So fans start to notice that the quality/filler ratio has gone down but there may have been just as many good songs per release as before, but the need to fill the disk with filler weakens the album. 
    Speaking of albums, my new album 'Synaesthetic' is available for download at Bandcamp. 54 minutes of good electro-ambient music. And only 3 songs!

  • It's Amazing how much work was put in to these technolgies to get dynamic range and now we use none of the range  but we have devices that can handle it. sad. And it seems like it is getting worse

  • Simon

    Keeps getting worse ? I guess the same but keep asking me how it can get even more worse ? Maybe someday we will all be listening to pink noise !? Only rhythmic patterns fade between the noise …

    Really, how can it get more worse ? What comes next ? no dynamic ?

  • No dynamic sounds good! Ask Merzbow..

  • I like chapter 4.2.1 of the pdf: "Avoid Loudness Envy"

  • r

    I'm always amused when researchers pretend that game theory has anything to do with social behavior. Nice work otherwise though…

  • tom1

    persever dynamics!

  • heinrichz

    great article..but this is wishful thinking..
    "Here’s another thought, too: with artists increasingly self-releasing or releasing through more specialized labels, greater access to music online, direct-to-consumer distribution, and online replacements for conventional terrestrial radio, many of the factors that produced some of the oddest hyper-compression at the top of the charts are fading into the background."
    ..the loudness wars are not going to go away once hearing is gone.

  • @cooptrol +1 for Merzbow reference, my friend!

    But yeah, I think there's a fine balance to be found when dealing with the usual nth-derivatives of rock genres and the sounds involved. You'll want to compress and pump it up a bit, but not to the point where your crashes are making everything buzz, à-la "Californication" (RHCP) or "Whatever People Say I Am, That's What I'm Not" (Arctic Monkeys) – regardless of how awesome those albums are.

  • JoshuaB

    I just graduated from an audio engineering school. As you can imagine this is a subject that came up quite often. We literally had a teacher ask us to help turn the tide being that we are the  'new generation' in the industry. Fight the good fight! Bring dynamics back to music. If it isn't loud enough let the listener turn it up!

  • greg

    This is a seriously shitty paper.
    No original research, assumes that the music industry is in fact in decline, and even assumes a reason for it without a citation . . . nevermind there are more people successfully supporting themselves as musicians right now than there has been in the history of humanity.
    Perhaps you need to stop reading press releases and commercial journalism, and start looking at numbers, slick.

    When there are citations, they're bullshit. "Many experts *suspect* . . . " "portable media players *can* . . . "
    And what the hell does game theory have to do with this at all? What on earth does the full page of discussion about it have to do with compression, the reasons for its increases, in either sense, or anything in the abstract?
    Musicians don't have tools to compare their work with and without effects?
    Isn't that what a bypass switch is for?
    One can't bypass effects from the (real or virtual) mixer?
    Would someone care to tell me how he determined a correlation, much less causality, between Norah Jones getting an award and the (completely unverified) claim that her record was noteworthy for its lack of compression?
    Furthermore, is he really trying to tell me that The Eagle's Greatest Hits is something recording artists should try to emulate?
    Does he think 10,000 signatures on an *online* petition means anything?
    Who is this? Does he play golf with my dad?
    Why did I have to read 16 pages before I got to anything, uhm, factual? Even when you get there, the research is flawed to the point of being useless, by his own admission.
    Why doesn't someone put his or her figurative dick where his or her word processor is and give us a couple statements of fact without qualifiers or negations? Does that require something besides looking at spectrums of a seriously limited sample of commercial recordings on your desktop? Is that too hard?
    Why doesn't this *interview* someone who makes decisions on how loud albums are instead of musing on why they might be making them louder?
    Finally, On page 20, he seriously suggests that the same consumers who can't tell the difference between a compressed and uncompressed track casually are going to look at a database.
    I want my 20 minutes back.

  • Andy

    Greg, I'm under the impression that you didn't get the whole text. He didn't just talk about compression (it's not a "how to become a better engineer" article), he try to explain why we all have to suffer from over compressed music. And here the game theory kicks in because it delivers an approach to explain the obvious constraint of pushing the loudness to the max. He also try to explain why we have to take the psychology of perception into account when we try to understand the effect of over compressed music.

    Norah Jones (or her engineer) most likely got an award because her CD doesn't contain over-compressed shit AND it reaches a high level of mixing / mastering. It's not about the lack of compression. It's about the voluntary and intended renunciation of deadly compressed trash.

    Just read it again and you may understand.

  • Eddie

    Many of your complaints have no solid basis. I was planning on reading the paper during the weekend but based on your comments decided to check it out today. Game Theory is nicely explained with specific regard to the loudness war:
    “a group of people facing a social dilemma may completely understand the situation, may appreciate how each of their actions contribute to a disastrous outcome, and still be unable to do anything about it”

    "Many mastering engineers are reluctant to apply excessive compression, but they often feel pressured by musicians and record company executives who are motivated to stay competitive with others, even at the possible expense of audio quality"

    To verify that Jones' record is not hyper compressed you simply have to listen to it. Not complicated at all. There's no correlation or causation involved regarding sales. Earl mentions a possibility of the record selling that many units because of its low level of compression compared to other albums from that year. The award is more closely related to the use of compression since it was a Best Engineered Album award.

    Without addressing each of your points, I would recommend that you spend more than 20 minutes reading the paper and understanding it. I know I've benefited from spending more than 20 minutes reading comments by Bob Katz, Bob Ludwig, and others regarding the loudness war. For example:

    Regarding your question "Who is this"? here's Earl's resume:
    No mention of him playing golf, though πŸ˜‰

    I have no relation with Earl whatsoever, other than being a satisfied customer (SFX Machine Pro). I believe that I also asked him a question once via email and his reply was fast and friendly. I look forward to spending a couple of hours over the weekend reading this paper.

  • "Not necessarily.   Flying Lotus records are both brilliant and so over compressed I have a hard time making it through a whole record (a bit better mix on Cosmogramma)."


  • greg

    Oh, I'm sure he's great at his job, I just walked away from that paper (not necessarily part of his job) feeling like I hadn't learned much . . . except a couple albums I have virtually no interest in are heavily/lightly compressed.

  • Greg,
    This was intended primarily as a review article, and as indicated in the title of the paper, many of these topics are still highly speculative. I've tried to highlight areas where additional research is needed.
    The section correlating Billboard sales rankings to dynamic range is original research. I think the paper also provides new ways of looking at a number of these topics. The game theory section is greatly abbreviated from what was originally planned, due to the length of the paper, and it is certainly not rigorous; however, I think it provides a useful way of viewing the situation.
    While several of your complaints have some validity, many of them misrepresent what I said, and I'm sure that given more time, you could have expressed them without the gratuitous insults and obscenities. It sounds like you have a number of good ideas for how to write a better paper. I encourage you and others to help perform listening tests and other needed studies.

  • ridiculous

    good luck with your quiet songs

    I'll be out bumping the the clubs with more bass 

    half the reason I dont enjoy listening to jazz is because they eq all the bass out and dont master it loud enough. Its like the And Justice For All of music genre's

    speaking of Bob Katz, I totally respect him, but his primary focus is acoustic and classical music mastering. Two dead genres last time I checked or turned on the radio.

    Not Techno Dubstep Drum n Bass, which is even engineered differently than those stale prog rock dweebs who did too much blow back in the 80's. Why not interview Simon from Exchange or the guys at Dubplates and Mastering Berlin 
    they are dons, and what they do translates brilliantly on million dollar sound systems.

    absolutely pay for good mastering, but worrying about this nonsense and how it relates to electronic music is so ridiculous

  • Andy

    Yeah, wobble bass for everybody. We all need jazz like boom boom boom techno techno boom boom boom. Why? Because acoustic music is a dead genre. LoL You're really the man … πŸ˜‰

    Seriously: dance music has to be loud, but to compress the shit out of it is contra productive.

  • Jonah

    @Peter KirnWith more labels there is more to "competition". Each of these labels are more noise, more perceived need to be loud just to be heard over everyone else's noise. 

    Would a coalition among music forums, blogs, etc. to display a prominent link to a site that educates  about over compression change things? Have the site Display The Beatles iTunes sales figures? πŸ™‚

    The thing is, loud generally does sound better, at first blush. And if loudness doesn't sound better it is more attention grabbing and stimulating.

    Overcompressed, flat commodity music is salty, greasy fast food. And fast food sells.

    Our system is based on an endless loop of new, different, more. This has brought more music technology prices down, but 

    The paper didn't really cover this when it talked about the "Remastering" problem. I think mp3s show that fidelity isn't a concern to most music consumers.

    So, you gotta sell the same music to people that already have it in 3 other formats. You also don't want customers to listen to the whole song, at full quality or at their leisure (I mean I can't put a song on my MP3 player to test and see how I like it jogging for example). How do you get their attention and show increased value? You make it louder.

    Electronic dance music especially falls back on shouting at you to cover up deficiencies and it works!

    The music commodity industry doesn't have ideas, but it takes what's stimulating and pushes it to the extreme and to the point of parody. First from rock, now from electronic music. What's left after that?
    PS. "Wall of Sound"  wasn't meant to be literal. Just listen to "Be My Baby"! It has great dynamics.

  • The point about listening to music at low volume is probably the most beneficial aspect to having music heavily compressed/limited.

  • I'm a little bit disappointed, there's no science in this paper.

  • Peter Kirn

    Well, parsing some of what people are saying, I see two issues:

    1. Being anti-*loudness* doesn't make much sense. Compression and multi-band compression are tools to achieve sonic aims. So that's why, say, you might enjoy FlyLo — or not, but at least it's an artistic intention. By the same token, though, thoughtless over-compression distorts (literally and figuratively) artistic intent, and can dissatisfy your listeners.

    2. I don't agree with others who assume independent electronic releases will be locked in just the same loudness wars. For one thing, as with the examples in the club — an over-compressed track may sound awful in the club, and that's dance music. I think your market is people's ears, and not having to go through narrow channels like radio may mean that you actually design your sound to those ears.

  • By the way, I've posted a video based on the AES presentation at&nbsp ; .

  • Irr Relevant

    Hey Greg and all…

    There is a whole lot of original research and papers on this topic on tc electronics website:

    especially those by Thomas Lund ( and others)

    And I'd like to point out the EBU's latest effort concerning the loudness practices, EBU Recommendation R128

  • It is plain simple to me, i dont like music where people use compressors as musical instruments.

  • bryan tewell

    I believe the loudness wars stem less from perceptions by industry executives of being able to more effectively catch a listener's attention, but have a lot more to do with trying to more realistically capture the experience of going to a live (loud) concert. I know this may seem a bit counter-intuitive especially with advances in studio production technology that usually suck all the realism out of recorded performances, BUT, a lot of these advances have also helped push the loudness wars forward. For example, drum replacement and doubling may make drum performances sound (or sometimes even exist as) manufactured (fake) performances, but the main advantage and purpose of such a production technique is increasing the punch and volume of the drums without making them noisy or over-bearing in a mix. With this, you start to feel the drums a lot more in general settings (including everything from a high-fidelity audio environment all the way down to ipod phones,) Just as you would in a real-life performance context. I know to many this mindset may seem extremely counter-intuitive, but I submit to the world that this process of mixing is actually better suited for certain types of music (not all, and honestly, not very many types of music). It brings the listener of a recording closer to the experience of seeing some kinds of music live through the excitement and energy of the music being really loud… I bet you don't have many people arguing on this side of the fence.

  • Dynamic music will make a CD sound far more like a live performance, when you attend a concert it may be loud but the snare drum still snaps the bass cuts through the mix and everything has a balance.

    I'm not sure having the drums and bass sound like they were recorded in a library makes it sound live at all.

    Listen to AC/DC back in black either the CD from the 80's or the vinyl and then listen to a modern band like Foo Fighters or Pendulum. Then crank up the volume I can guarantee black in black will sound far more "live".

  • Do you remember the pepsi challenge? Turns out in a sip test, people prefer pepsi over coke. The secret? More sugar? Stands to reason then that coke should consider trumping pepsi in the sugar content then, right? Turns out in a more sustained test, say bringing home a six pack, more people end up preferring coke.

    The loudness wars is much about that initial reaction from consumers and vying for their attention. I remember once when a tv manufacturer came out with a limiter on their set's volume. That idea was shut down within the year; I speculate they had too much pressure from other interests. And so for those of us who prefer DnB to jazz, I  just want to point out that some agonizingly annoying ad is going to come on your tele or radio station and blow the volume of your track hell out of the water (not that either genre has much coverage there anyway).

  • Eddie

    Finally I got to read the article, watch the video, listen to samples, and look at additional research.
    Regarding the questions “Is there a problem?” and “How do we solve it?” I would say that there is indeed a problem, and we are unlikely to solve it.
    Our best bet would be to educate the artists and consumers. This is easier said than done. One obstacle is that most people don’t know how to listen and/or what to listen for. Even musicians.

    Why should people with better ears pay a premium? If we want to win the war on loudness and educate people, I would go in the opposite direction. Allow artists and engineers to release better, more musical representations of their art, without any price increase. In fact, make it cheaper –or even free- as educational material.
    Make better mixes available so that listeners can compare and contrast both versions. Just as movies have a director’s cut, allow an “engineer’s cut” even if it’s just a sample of the whole song.

    Many engineers today will do whatever the artist and/or label tells them. They have bills to pay, mouths to feed, and clients to keep happy. But perhaps they could be allowed to create their own final mixes that are independent of the commercial product.

    Some challenges:
    Consumers are easily manipulated, and by now they tend to believe that louder is better. This is reinforced since louder does seem better as an initial impression, even if it’s not better for long-term listening or when listening more critically to the audio quality. Examples of consumer manipulation include the food industry and the “nutritional supplement” industry. Both are billion dollar industries that often sell questionable products. They can even market as “healthy” things that are unhealthy, and people buy the products. For example, instead of showing sugar as an ingredient they may call it evaporated cane juice. Low-fat, reduced-fat, no sugar added, natural, etc., may be included in food that is not healthy, yet consumers are deceived into thinking that they’re eating healthier.

    Here’s a 122-page report… (I have not read it yet).
    Could something similar be done in music? Can we push for health-related research in this case?

    With food supplements it’s worse. Progress has been made with the USP Verified symbol. At least this allows consumers to know that they’re getting what is described in the package (e.g., Vitamin C, Garlic, etc.). In this multi-billion dollar industry it was not uncommon for some products to be different than what was described. And of course, the USP verification does not address whether or not the product actually works as described.

    In spite of labels and available information, consumers in these industries make sub-optimal choices.

    Could a label be used for music (eventually)? “Prolonged exposure may lead to listening fatigue.” Or a positive one: “Mixed and mastered with your prolonged listening pleasure in mind.”

    How do we educate the listening public, when musicians and engineers may also need educating? I've encountered plenty of musicians and some “engineers” who think that louder is better, compress everything, and in spite of working in the music BUSINESS don’t have the proper ear training in this regard. If these people represent a challenge, how about the average listener? Worse yet, how about the fanboyz/grlz?

    Look at Death Magnetic’s reviews on Amazon. A majority of the reviewers gave it 4 or 5 stars. Do they care about sound? There are those who complain about the sound quality, and interestingly some people will disagree with them. Similar to Springsteen’s Working on a Dream. That album has a horrible sound.

    Indeed there’s a problem. Fortunately there’s awareness, technology (TT Dynamic Range Meter), and people who want to make a positive change. There’s also humor:

    Posting -as educational material- comparisons on YouTube (e.g., Wicked Game) may also help. Just a few seconds of each version. This could help educate people. For my own edification I heard It's Oh So Quiet after applying heavy compression and making the volume of the original and the compressed file the same so that they would be comparable. Those can be ear-opening exercises for the general public, since the effect is evident. No room for subtleties πŸ˜‰

    Thanks for the paper and video Earl πŸ™‚

  • yo

    To be honest we're getting into an age of the loudness war is getting louder by the day! 4/4 music designed for a club sounds better over comp'd & slightly distorted however in a broadcasting environment it will sound like crap thus radio edits nowadays usually retain the dynamics for that purpose than their original counterparts. Get use to it folks the loudness thing is the norm now for dance music. 

  • AntoxaGray

    And yet many producers still do that because that’s what mastering tutorials on youtube say.