AES Journal Forum

The Loudness War: Do Louder, Hypercompressed Recordings Sell Better?

(Subscribe to this discussion)

Document Thumbnail

[Feature] The term “loudness war” refers to the ongoing competitive increase in the loudness of commercially distributed music. While this increase has been facilitated by the use of dynamic range compression, limiting, and clipping, the underlying cause is the belief that louder recordings sell better. This paper briefly reviews some possible side effects of the loudness war and presents evidence questioning the assumption that loudness is significantly correlated to listener preference and sales ranking.

JAES Volume 59 Issue 5 pp. 346-351; May 2011
Publication Date:

Click to purchase paper as a non-member or you can login as an AES member to see more options.

(Comment on this feature)

Comments on this feature

Page:   1   2
Kevin Gross
TC-NAS Chair
Kevin Gross

Comment posted January 13, 2012 @ 17:17:13 UTC (Comment permalink)

The accomplishment of this paper is that it *finds no evidence* that louder recordings sell better. Unfortunately the author demonstartes an anti-loudness-war prediposition and so the reader can't assume he's highly motivated to seek such *evidence*. Nevertheless, hopefully this research will shift the loudness discussion away from sales and towards sound quality. I assume Vickers is expecting that to lead us away from hypercompression but it also could lead to an admission that targeted listeners expect and perhaps even prefer the sound of hypercompression. That would paint the loudness war a different color, wouldn't it?

Earl Vickers
Author Response
Earl Vickers

Comment posted January 17, 2012 @ 14:34:24 UTC (Comment permalink)

Yes, if a large majority of listeners preferred the sound of hypercompression. To the best of my knowledge, no such study has been done.

As discussed in more detail in the convention paper from which this Journal article was hypercompressed, it's likely that preferences may vary for different listeners and listening situations. If that's true, then the optimal scenario might be for the mastering engineer to apply a little less compression, to satisfy listeners who prefer more dynamic range, and allow the playback device to perform additional compression when desired (for mobile, background, or late night listening, masking of ambient noise, or according to personal preference).

A somewhat less optimal, but still acceptable, scenario would be to apply a compromise amount of compression that would preserve some dynamics and still work pretty well for everyone. Widespread adoption of automatic loudness normalization could help facilitate this.

The worst-case scenario would be that decisions are based largely on unproven sales assumptions, instead of sound quality. I suspect this may be the current situation. If, in fact, hypercompression drives away more customers than it attracts, this may even be a lose-lose scenario for both sales and sound quality. There is little solid evidence one way or another.

My article was primarily intended to question assumptions and call for further study, though it may have veered a bit more toward advocacy than was prudent. At any rate, if anyone is aware of evidence I've overlooked, or would care to contribute original research on the topic, I'd be interested to see it.

Earl Vickers
Author Response
Earl Vickers

Comment posted February 7, 2012 @ 14:31:38 UTC (Comment permalink)

At the 2011 New York AES Convention, Susan Rogers gave an interesting presentation (audio available from the AES store), that referenced and occasionally questioned my papers on this topic. While I disagree with some of her conclusions, I'm rethinking the issue of dynamic range compression and hearing loss.

My paper made the traditional assumption, as embodied in the ISO1999 standard, that hearing damage is proportional to cumulative energy exposure; this would suggest that compressed music is likely to be more damaging, since it spends more time at higher levels. However, Rogers quotes Gerald Fleischer, "Strategies of the Hearing System Against Noise and Auditory Damage," in Reflections on Sound, Svensson, P. (Ed.), 2008, who states that hearing damage is more likely to be caused by short impulsive noise than by ongoing continuous noise, due to the ear's protective mechanisms. To the extent that dynamic range compression reduces the relative level of impulsive peaks such as drum hits, it may in fact serve a protective function. However, this is a hard thing to test. At any rate, this effect applies mainly to fast (microdynamic) compression, not to compression with slower time constants (equalizing the levels of verse & chorus, etc.).

Regardless of the effect of hypercompression, I think the weight of the evidence strongly suggests that extended headphone and ear bud listening at high levels can be harmful, particularly in regard to tinnitus. (More at .)

Default Avatar
John Groper

Comment posted February 9, 2012 @ 14:36:14 UTC (Comment permalink)

With regard to Earl's comment on the ISO1999 Standard assuming "... that hearing damage is proportional to cumulative energy exposure..." I would have to respectfully disagree with the assumption in the Standard. While the assumption may be true in many cases, it cannot be true in ALL cases. Take, for example, hearing loss by military personnel exposed to extreme SPLs of short duration ( eg gunshots or explosions.) This is referred to as "acoustic trauma." A single incidence of acoustic trauma may result in hearing loss.

Some years ago, while searching for personal hearing protection, I was shown a product that (according to the salesperson) contained a fast-acting diaphragm to close off thed ear canal and protect against acoustic trauma for shooters. The theory he gave me was that these diaphragms acted more quickly than the ear's natural protective mechanisms. In the absence of extreme high level sound these devices were supposed to open up the acoustic pathway to "near-normal" to allow normal listening levels. I have done a brief search of the web and do not find these devices anywhere at this time. It may be that they worked for awhile, but degraded over time. With no way of knowing if the diaphragms still functioned as designed, the user might be hesitant to continue long-term use. I certainly would!


I recognize that the comment may be a bit "off-topic," but I couldn't let the assumption pass without mention.

Default Avatar
Willard Ford III

Comment posted May 29, 2012 @ 11:37:43 UTC (Comment permalink)

Though the author *finds no evidence* that louder recordings sell better, if he was actually an engineer working in studios regularly he'd know the answer...YES.,

There is a simple and very basic reason that louder recordings sell better. The fact is, the record company won't accept the master unless it is compressed to pump out the volume on FM radio or iTunes.  If the record company doesn't accept the master, the record doesn't sell.  Simple.

Earl Vickers
Author Response
Earl Vickers

Comment posted May 30, 2012 @ 13:44:28 UTC (Comment permalink)

The question of interest to me was whether, all other things being equal, consumers prefer louder, hypercompressed music. While I don't see evidence of a correlation between hypercompression and better sales among the albums that reached the Billboard top 200, the albums that charted did indeed tend to be more compressed than those that did not. This correlation does not necessarily imply that consumers prefer hypercompressed music. In fact, I suspect your suggestion is correct,  that hypercompression is the price for getting on the charts in the first place. (This correlation could also relate to genre and other factors.)

So have industry gatekeepers rigorously studied this issue and proven that hypercompression results in greater sales, given the same level of promotion? Or is this largely groupthink, superstition, loudness envy, old wives tales (apologies to old wives), and self-fulfulling prophecy on the part of the record companies? I don't know, but to me the fact that there seems to be little or no correlation between sales and compression among the heavily-promoted top 200 suggests the latter. Given that compressed music doesn't necessarily sound better on FM or through chains of digital codecs, and especially as the world moves toward loudness normalization, hypercompression may yield little or no sales advantage (again, given the same amount of promotion).

Default Avatar
Robert F. Auld

Comment posted July 9, 2012 @ 19:57:47 UTC (Comment permalink)

Regarding this comment: "The fact is, the record company won't accept the master unless it is compressed to pump out the volume on FM radio or iTunes..."    That may be the case for some record companies, and not for others.  It may also depend on the genre of music—a jazz label may not have this attitude, a hip-hop label probably would.  

But one thing is clear—there are people in the record companies, as well as studio clients, who need to be educated about what happens to music that is played on terrestrial radio.  As most on this forum know, radio stations have dynamic processors that limit peaks and level out average volumes.  The more aggressively loud stations use these dynamic processors heavily, other stations (say, NPR affiliates) less so.  But in every case, anything played over the air will be peak limited and have its average level adjusted to match other programming on that station.  So, if you submit a hypercompressed CD to a radio station, when played on the air its level will be lowered to match other programing.  And a CD with wide dynamic range will be compressed and boosted to match other programming.  

Robert Orban (designer of the most popular broadcast processing line) is emphatic that music that has been only moderately compressed sounds better when sent through his processors than music that has been hyper-compressed.  Frank Foti (designer of another popular line of broadcast processors) agrees with Orban.  Studio clients and record companies need to be told this—if necessary, over and over.  Making it as loud as possible does not make it sound good on the radio--quite the opposite.

As for iTunes and related download services, they are a somewhat different beast.  However, most designers of digital audio compression codecs agree that music submitted for encoding should have about 2 dB of headroom.  Music that goes right up to digital zero can, in some cases, cause internal distortion in these codecs.  And at Fraunhofer (the designers of the MP3 codec), they have stated that hyper-compressed material does not encode and decode well, as the MP3 codec was designed around a pyscho-acoustic model that assumes normal dynamic range.  Put highly compressed material through the codec and that model starts to break down, and you get audible artifacts.  Again, clients and record companies need to be told this—if necessary, over and over.

Finally, iTunes and other media players (like WinAmp) are increasingly using normalization algorithms to adjust the relative volumes of different tracks.  A listener who uses this feature will not hear a hyper-compressed track as being louder than one that is not.  In short, the increasing use of loudness normalization renders the loudness race irrelevant.   Again (notice a pattern here?), clients and record companies need to be told about this.  

Kevin Gross
TC-NAS Chair
Kevin Gross

Comment posted May 30, 2012 @ 13:43:12 UTC (Comment permalink)

It is an interesting observation that could suggest additional research. After controlling for other variables, are louder stations more listened to? Out in the real world, is excessive loudness fatiquing? Do people only listen to louder stations only on the way home from work or do they listen to them all day? Is length of commute correlated with station preference? Is loudness correlated with road rage?

Earl Vickers
Author Response
Earl Vickers

Comment posted July 11, 2012 @ 19:37:54 UTC (Comment permalink)

Regarding the issue of whether record companies will release CDs that aren't hypercompressed, and whether such CDs can do well in the marketplace, the latest Electronic Musician magazine had an article on Jack White: 

"One stipulation White had for mastering engineer Bob Ludwig was that he wouldn't use any dynamics processing in the mastering process. 'There was a study from Earl Vickers [ loudnesswar] about the loudness wars,' [Vance] Powell says. 'As the loudness war escalated, record sales went down. I'm not saying we're killing the loudness war. But I think it's a very bold move for Jack to say, 'I realize there are records out there that are going to be louder, but I don't care....'"

He asked Ludwig if there was a way to make it louder without changing the dynamics, and Bob said yes. "So the master came back, and it sounded great. There's nothing squashed or lost in the dynamics, and it still sounded really loud."

To me, the music has a more open quality than many recent CDs - it breathes and feels less fatiguing. At any rate, this seems to be an existence proof that it's possible for artists to hit #1 on the Billboard charts without hypercompression. 

Page:   1   2

Subscribe to this discussion

RSS Feed To be notified of new comments on this feature you can subscribe to this RSS feed. Forum users should login to see additional options.

Join this discussion!

If you would like to contribute to the discussion about this feature and are an AES member then you can login here:

If you are not yet an AES member and have something important to say about this feature then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.

AES - Audio Engineering Society