• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Loudness compression, loudness wars.. What exactly it is and why is it happening?

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,190
Likes
2,318
Location
Zg, Cro
#1
It would be very nice if someone of you takes a little time and explain this loudness thing to the rest of us that don't really know what it refers to. :)
 

DonH56

Technical Expert
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
3,398
Likes
4,486
Location
Monument, CO
#2
I'll assume you know what loudness means, simply how loudly the sound that you hear (the volume, or sound pressure level -- SPL).

Dynamic range is the difference between the softest and loudest sound.

In general, given two identical recordings but with one just a little louder, people will choose the louder one as "better". This can be taken to extremes, of course, but it is true for very small differences (~0.1 dB) in loudness. Manufacturers of music (etc.) sources know this, of course, so by making their offerings a little louder than the other guy's they may have a market advantage.

Now, how do you do that? If you want to make the overall level (sound) louder, generally a couple of things are going to take place:
  1. Limit the loudest sounds, using a hard limiter (clip the signal, which adds distortion that will make it sound even louder albeit "harsher") or soft limiter (compress the loudest signals by "rounding them off"). This way you can turn up the volume and put the loudest signal right at the maximum limit of the bits available (I am assuming a CD or digital format but the same idea works for analog recordings like LPs or tape; raise the volume to the very highest you can get away with).
  2. Raise the level of lowest signals so they are not so quiet; that is, boost the loudness of the softest signals.
The net result is loudness compression -- the overall signal is now much louder, but the dynamic range has been compressed so the difference between softest and loudest signals is less than the original, and the loudest sounds are pushed right to the maximum the recording media (CD, hi-res, LP, tape, DVD, BD, whatever) can handle.

This can be useful in a few cases. If the softest signal recorded is below the noise level of the media, say -80 dB, so it cannot be reliably captured or heard on a tape or LP, then boosting that signal so you can capture (and later reproduce it in playback) makes sense. Similarly, if the signal is too loud and would saturate (overdrive) the recorder, limiting (compressing) it means you can still capture it with (hopefully) minimal distortion. 120 dB dynamic range sounds really great on paper, but in the real world most of us do not have quiet enough rooms or powerful enough amplifiers with speakers sufficient to reproduce (playback) such a large dynamic range. In a car it is even worse, so compressing the signal's dynamic range (loudness compression) again makes sense.

The problem, as in many things, is when it is done to excess. Some recordings have been compressed so much that the dynamic range is only 20 dB or less, then pushed to the "top" of the recording media's dynamic range so the result is a song (or whatever) that is just LOUD. But louder wins, right? Well, not when the dynamic range is squashed so there is very little difference between loud and soft, and is so limited at the top end that it all sounds distorted. Nobody really wins a loudness war.

HTH, IME, IMO, FWIWFM, my 0.000001 cents (microcent), etc. - Don
 

Dialectic

Major Contributor
Patreon Donor
Joined
Sep 26, 2017
Messages
1,082
Likes
990
Location
New York City
#3
In the mid- to late 2000s, when the iPod and those f***ing earbuds were at the peak of their popularity, concern began to spread among recording artists and record label executives that listeners would tune out soft songs and listen to other, louder songs in their playlists. The result was a competition among mastering engineers (and sometimes mixing engineers) to make tracks as loud as possible, or:

THE LOUDNESS WARS

An album often cited as an exemplar of the loudness wars is the Red Hot Chili Peppers' Stadium Arcadium.

At a technical level, albums mastered to be as loud as possible undergo extensive dynamic range compression, usually in the mastering studio (the last step before the album goes to the pressing plant or to iTunes) but sometimes earlier in the production process. Historically, dynamic range compressors were analog boxes, some of which were said to sound good and add pleasing coloration to the sound. However, during the loudness wars, VST plugin dynamic range compressors have become more common. These plugins apply simple algorithms with aggressive compression and no pleasing colorations. While popular music recordings have conventionally undergone some dynamic range compression since the 1960s, the VST plugins now in wide use are capable of applying much more dynamic range compression than analog compressors; a simple algorithm can boost the level of an entire track to the maximum.

As a consequence of this extensive use of VST plugin dynamic range compression, there is pervasive digital clipping on most CD releases of popular music from the last ten years or so. Those of us who know about audio engineering know that clipping is almost always bad because it profoundly distorts a waveform. Any halfway decent recording engineer who has used DAW software in the last two decades knows to set levels low enough to avoid clipping. The bizarre consequence of the loudness wars is that recording artists and executives now want clipping, and mastering engineers now tout their ability to offer loud tracks over their ability to make albums sound good.

An important question is what the hell is happening to the overall waveform when massive amounts of dynamic range compression and digital clipping are added to a track. Here's what the difference looks like in DAW software:



The green waveform is a version of a Metallica song on the Guitar Hero game, in which no digital dynamic range compression was applied. The blue waveform is from the CD. Note that, on the CD version, the dynamic range of most of the track is pushed to the limit. The level is so high that peaks are simply lobbed off (or clipped).

Here's a powerful demonstration of the audible effects of such compression and clipping by the great Bob Katz.

A few small areas of the recording industry have remained mostly free of the baleful influence of the loudness wars:
  • classical music;
  • jazz; and
  • anything on vinyl.
Perhaps the strangest consequence of the loudness wars is that lots of young people who listen to popular music now insist that vinyl sounds better than digital. With respect to what they listen to, they're probably right because vinyl masters usually are prepared separately from digital masters and do not undergo the same degree of dynamic range compression.
 

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#4
regarding "anything on vinyl" ... As you stated, most releases on CD over the last 10 years have been compressed/limited to a greater degree; the same can be said for vinyl during that same period.

Actually, 30 years ... many if not most CD remasters from the 90's, display more compression than the originals.

This is the very reason many philes look for the originals, both CD & LP.
 

Timbo2

Senior Member
Joined
Feb 27, 2018
Messages
328
Likes
200
Location
USA
#5
The green waveform is a version of a Metallica song on the Guitar Hero game, in which no digital dynamic range compression was applied. The blue waveform is from the CD. Note that, on the CD version, the dynamic range of most of the track is pushed to the limit. The level is so high that peaks are simply lobbed off (or clipped).
That particular Metallica album made me stop buying any new popular music until last year. It was unreal how positively awful it was. On the plus side it was so bad that the Wall Street Journal actually picked it up and made it a story.

I'm a subscriber so here is a Link. I'm not sure how many clicks or how long it will remain valid. After that it will be paywalled again.
 
Last edited:

Dialectic

Major Contributor
Patreon Donor
Joined
Sep 26, 2017
Messages
1,082
Likes
990
Location
New York City
#6
regarding "anything on vinyl" ... As you stated, most releases on CD over the last 10 years have been compressed/limited to a greater degree; the same can be said for vinyl during that same period.

Actually, 30 years ... many if not most CD remasters from the 90's, display more compression than the originals.

This is the very reason many philes look for the originals, both CD & LP.
Vinyl may be more compressed now, but it generally does not contain digital clipping in the way that recent CDs of popular music do.
 

bennetng

Addicted to Fun and Learning
Joined
Nov 15, 2017
Messages
612
Likes
483
#7
When analog media like tape and vinyl clip, they will not have a perfectly chopped waveform like digital, it is the limitation of such media since they have phase and frequency response issues and these imperfections made the analog waveform "look" nice even when they sound horrible.

Lossy codec also alter the waveform so they can also make loudness war songs look "more dynamic" when you look at the waveform.

Not everyone have cutting lathe or open reel tape recorders to do experiments but for those who have cassette tape recorders, record a loudness war flac on a cassette, re-digitize it and the waveform will look super nice but when you listen to the recording it still sounds like garbage.
 

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#8
Vinyl may be more compressed now, but it generally does not contain digital clipping in the way that recent CDs of popular music do.
Correct. Vinyl by&large escaped the war; but few of the Remasters I've recently purchased (last ten years) have measured DR=original, and sometimes the results have been quite disappointing, esp considering the ~$30 cost.

When researching titles, this helps http://dr.loudness-war.info/ provides useful info on music/DR content.

When analog media like tape and vinyl clip, they will not have a perfectly chopped waveform like digital, it is the limitation of such media since they have phase and frequency response issues and these imperfections made the analog waveform "look" nice even when they sound horrible.
Vinyl miss-tracks, then it sounds "horrible", otherwise ...
 

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#10
That particular Metallica album made me stop buying any new popular music until last year. It was unreal how positively awful it was.
The Page/Marino remasters were awful, so much added compression and lots of clipping ... yet many claimed they were "better". Threw me off purchasing any remasters.

That said, a few Remaster-Gems (dynamic-range terms) out there ... even the originals can't duplicate.

case in point: Ken Scott's 2003 SACD reissue/remix of Ziggy Stardust, brilliant!!! (one of the very few SACD's which actually delivered what HRez initially promised)
 
Last edited:

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#12
I laughed so hard, I cried; watching Hetfield "struggle" on stage during Metallica's 2017 Grammy spot.
 

DonH56

Technical Expert
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
3,398
Likes
4,486
Location
Monument, CO
#13
The Loudness War explained in under 2 minutes:


A much longer article that I wrote coming from 10 years as a pro recording/mixing engineer:

https://www.computeraudiophile.com/ca/ca-academy/dynamic-range-no-quiet-no-loud-r643/
We need a bigger thumb's-up icon for this one!

This brings back so many memories of my (very limited -- infinitesimal compared to Mitch) studio days and getting to play with A800's and even my own modest 1" 2-track Studer. Tape slamming started small then grew to be a big thing, but the compression caused by tape saturation (unless carried to extremes, and not sayin' that didn't happen) is nowhere near as obnoxious as digital clipping and hard limiting that was and is used far too often.

Everyone should at the least watch the short video Mitch made; it explains the problem much better than words ever could, and is guaranteed to leave you shaken and angry at the sound of the (vast?) majority of recordings available today (new or re-mastered). At least based upon my collection (small now compared to my old vinyl collection, <1000 CDs.) I have old pop records and modern CD versions and the CDs are just terrible, and it has NOTHING to do with digitization, quantization of time and amplitude, or anything except dynamic range compression that squeezed the life out of them. (OK, some of them are EQ'd pretty badly IMO, but that is easy to fix.) Sheffield Labs is one of the few that have remained faithful, haven't really kept up with the rest. I have three and four different CDs of some treasured old albums I bought in the (mostly vain) hope of finding just one that didn't have compression and gain-riding that killed the dynamics.
 
Last edited:

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#14
... is guaranteed to leave you shaken and angry and the sound of the (vast?) majority of recordings available today (new or re-mastered). At least based upon my collection ...
My experience exactly. While digital is a superior medium, digital mastering went from good to mostly horrendous. Vinyl has different issues, but its not been nearly as crippled dynamically.

Hence, when done right, an original LP - properly ripped to CD will "sound" (and measure) better than many, if not most, CD remasters. The problem is: when many pro-vinylheads compare LP to digital they generally don't understand and/or refer pressing details, but this doesn't stop 'em from issuing the standard pro-vinyl statement claiming "no CDP player can do that" in terms of dynamic slam. Well, no shit, Sherlock! Depending on the pressings, when comparing early higher DR valued original CDs to the same high DR original LP, its often (but not nearly always) a different story.
From an inventory point of view, I've far more vinyl with higher DR capability than what's within my digital collection, which is sad indeed ...
 

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
7,238
Likes
7,948
#15
A group of friends I record sometimes went into a studio to record a Christmas CD. Mostly very old traditional Christmas songs going back to the middle ages. All acoustic group with some singers. The results were bizarre. Despite being told several times they didn't want the typical compressed super loud recording that is what they got back.

Death Magnetic was famous for its terrible compression and on the software that generates DR levels it was a 7. This Christmas CD of an acoustic group was a DR6. The first person to pick up the preliminary mastering believed something had broken in their stereo system and wanted to hear it somewhere else. The violinist listened to her favorite song and said, "I know I was playing on this one, but I can't hear anything that sounds like a violin." I don't know how anyone would think the result was good in any way.

So after complaining, the second time around you could only say it was less terrible having heard the first version. This one was a DR8 or 9. So they played some recordings I had made which have just a little compression so you don't have to listen in a completely quiet room. Guy says, "I didn't know you wanted that kind of sound." The third version was really nice. More compressed than I would have done it, but not by much and the fellow had good gear and good mastering skills. I could only think of how those skills were wasted doing flat-line masters for everyone.

Here is the waveform for Christ Child's Lullaby from that first mastering.
Christ childs lullaby compressed.png


And the waveform of the one used from the 3rd version of the mastering.
Christ childs lullaby 3rd master.png


BTW, they are about 30 seconds different in length because during some editing the first time he repeated 30 something seconds in the middle by accident.
 

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,375
Likes
1,142
#16
THE LOUDNESS WAR IS OVER

It’s quite typical that people debate an issue most fervently right after the issue was in retreat. This happens all the time, to all of us.

Loudness was at its strongest about ten years ago. That’s right; a decade ago.

Since then loudness seems to have gone down, towards the levels of ye olden days. If Deruty and Pachet (2015) are right loudness will be «normal» between now and 2026. See figure below.

However, this doesn’t make existing recordings and masterings that have too much loudness great again. Debating things that cannot be altered is also one of those human pastimes.

Source: http://ismir2015.uma.es/articles/136_Paper.pdf

 

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#17
Dennis, Mark Waldrep (AIX) tell similar stories about his Bad Company days.
 

TBone

Major Contributor
Joined
Mar 16, 2016
Messages
1,001
Likes
155
#18
Loudness was at its strongest about ten years ago. That’s right; a decade ago.
It's not as bad as it was, but its still bad ... ex: Adele's 2016 recording has values as low as DR4 ...

how many more recent examples would you like?
 

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,375
Likes
1,142
#19
It's not as bad as it was, but its still bad ... ex: Adele's 2016 recording has values as low as DR4 ...

how many more recent examples would you like?
I refer to the research piece I quoted. Anecdotes are interesting, but still anecdotes.

;)

I meet this sort of argumentation every day when people are confronted with data. So I hear quite often that the data must be wrong. Or the method. Just in order for people not having to change their opinion.

«When the facts change, I change my opinion. What do you do, Sir?»
 

mitchco

Senior Member
Joined
May 24, 2016
Messages
307
Likes
824
#20
dynamicrangeday_loudness.jpg


Another interesting graphic...

As much as I would like to believe the loudness war is over or subsiding, my recent purchases of modern music in the rock, R&B and alt categories show otherwise. I just took a quick snap of a few tunes from 2017-18 and unfortunately still unacceptable DR:

dr issue still.JPG


What are you gonna do eh?

PS @DonH56 - thanks for the kind words. I did not make the video, credit goes to Matt Mayfield, in which he did a fantastic job in not only showing the difference, but one can easily hear the difference, even over Youtube, with its own compression and listening on laptop speakers.
 
Top Bottom