Talk:Loudness war

Modern exceptions

I noticed it says in the article there are very few, there are a few including.

Franz Ferdinand - Franz Ferdinand (loud but good by todays standards) RG -6db

Eskobar - 'Til We're Dead RG -4.5db

The 411 - Between the Sheets

Mark Ronson - Version

I wonder if other people agree? AJUK ^Talk!! 09:45, 10 November 2007 (UTC)

"Forever Faithless" sounds awful - compare the version of "Insomnia" on it against the same track from the mid-90s "Now" compilation. The "Now" albums always seemed to be slightly compressed anyway but against the new Faithless CD it sounds far better. Not familiar with the other albums, but I did recently rip and encode the third Norah Jones CD which didn't have the same level of brick-wall limiting as most modern CDs. Squirrel 07:59, 10 November 2007 (UTC)

I'm unfamiliar with all of those albums, but could you calculate the Replay Gain values for each of those albums to give us an idea of their relative loudness? --MinorContributor 18:32, 10 November 2007 (UTC)

Version is badly compressed. Not as bad as some but my no means "good". Mojo-chan 18:04, 14 November 2007 (UTC)

I only have Stop Me and it looks like this [1] I assumed that was what the rest of the album must look like. AJUK ^Talk!! 19:31, 14 November 2007 (UTC)

Most classical recordings use very little or no compression, or at least none that is audible. I believe many recordings, especially live recordings, use no processing at all. And they almost inevitably have a much larger dynamic range than any pop recording. So for an example of a recording free of compression, limiting, or clipping, you could use most decent classical recordings, especially those by the late John Eargle for Delos. Karlchwe (talk) 17:49, 11 March 2008 (UTC)

Sadly, while Franz Ferdinand's first album was very well produced, their second one is the typical super-loud junk. -- Stormwatch (talk) 10:28, 13 April 2008 (UTC)

Yes I had trouble understanding why they did that, thought they kind of proved a point with the first album it was so popular; and did much better than the second! AJUK ^Talk!! 19:27, 20 April 2008 (UTC)

Rolling Stone article

There's a new Rolling Stone article about the loudness wars called The Death of High Fidelity. Please feel free to add new references from this article; I don't have time to go through the whole thing myself right now Illuminatedwax (talk) 15:03, 28 December 2007 (UTC)

I read that article and it's on target. We could use it for more refs. Thanks for bringing it to WP.Jrod2 (talk) —Preceding comment was added at 17:18, 2 January 2008 (UTC)

The Loudness War & Live Music

Since (often digital) mixing decks for live music these days often have all this 'technostuff' built in to allow exactly all this sort of boosting & compression & limiting, 'Live Music Events' (with or without 'lipsynching') are now experiencing the same problems.

I suggest a new paragraph, at least - or is it worth a separate page?

FoolesTroupe (talk) 02:15, 13 January 2008 (UTC)

Why would that do that? AJUK ^Talk!! 12:41, 13 January 2008 (UTC)

I don't think it's worth a mention or a separate page. My career's in the live field exclusively, save five years of university schooling in studio techniques. As a longtime live guy, I just don't agree that boosting and compression and limiting are the reason for WAAAY TOO MUCH LOUDNESS at concerts. The main reason is that loudspeakers are more capable than they were twenty years ago; that and Joe Average Sound Guy mixes until it sounds "edgy" which means he's tickling the higher distortion levels of either the upper performance envelope of the PA (which is already too loud) or the upper electrical limit of some other part of the mixing chain. My personal experience is that when I'm asked to mix at a level I know is too loud for the audience's safety, I try to use the "technostuff" for good purposes--I use it to generate a higher level of harmonic content (distortion) within the mixer so that the client gets the impression that the PA is really bumpin' into its hard stops when it actually has quite a lot of "go" left. I make the mix artificially distorted (for these certain clients) while producing a true sound level that measures safe by objective metering tools. Of course, this kind of audio trick can rob dynamic depth from the performance but in my estimation it's better to satisfy an insistent client with "technostuff" tricks than to harm the thousands of people in the audience with too much real volume. Unfortunately, my viewpoint is not the norm. Tons of sound engineers get and keep their jobs by pushing the PA to its limits--and it's not the boosting and limiting tools and tricks that allow this, it's the sheer power of modern loudspeakers.

And all of this is original research. Not ready for Wikipedia prime time... Binksternet (talk) 07:15, 13 January 2008 (UTC)

"I just don't agree that boosting and compression and limiting are the reason" Not JUST ... I have heard live concerts - some from miles away :-) - and heard what seemed like the good old 60/70s 'overdrive a valve amp till the speaker cones breakup' sound - but watching the meter on instantaneous - it was obvious that so much compression and limiting was used (and the bass boost was obvious!), that what was happening was probably limiting as far up in the chain as in the mixer - and of course with well over 120dB(C) ON SITE - the sound guy was deafened (or at least aurally impaired!) anyway! :-) You know - those sort of 'live concerts' where the attraction and 'drug' is the bass boost and volume - see the Radio National reference about "The Bottom End" - which is not OR - it's a Radio show - but a transcript on line is possible, especially if enough requests reach them.

They are not all as good as you (and me!) out there... :-) :-P

FoolesTroupe (talk) 11:09, 13 January 2008 (UTC)

This would be worth a mention somewhere, but not in this article, which is about a specific subject relating to problems with fixed ceiling media. Tacking on other subjects relating to loudness creates a WP:COATRACK. Incidentally, that like to "The Botton End" is similarly off-topic, and not very informative. / edg ☺ ☭ 11:24, 13 January 2008 (UTC)

I have listened to "The Bottom End" twice, and it IS highly relevant, AND extremely informative - have you listened to it?

FoolesTroupe (talk) 02:33, 15 January 2008 (UTC)

New articles on The Loudness War

Check this out, it is really good and will really help the article: 1) The Baltimore Sun Audio gain in volume signals loss for listeners --Goferwiki (talk) 02:47, 14 January 2008 (UTC)

New Image

I don't think this is a good example as there appears to be no degradation of sound quality untill the final image and that is good by todays standards, I think normalizing a track so that the peak level is at full volume is a good thing. Although I am surprised there would be a need to strangle Beatles records. AJUK ^Talk!! 12:38, 17 January 2008 (UTC)

Since the final image shows apparent degradation of sound quality, this is a good example. The progression toward this brickwalling demonstrates how the decisions being made to make CDs louder. My only concern about the image is that the years may not be clearly readable, making it hard at first to tell which is the first image and which is the end result. While more extreme example of a flattened before/after waveform can probably be found, I think the replaygain images demonstrate this fine. / edg ☺ ☭ 14:58, 17 January 2008 (UTC)

Changed colours and added text saying that this illustrates the trend of increasing loudness. Aquegg (talk) 07:30, 18 January 2008 (UTC)

I think it looks fine, if anything, slow it down to 2 sec/frame and create another frame where the signal, for the most part, looks like solid block. Good job, man! Jrod2 (talk) 03:17, 19 January 2008 (UTC)

Why are there no examples of GOOD production?

Interestingly, after having read the article and some of this discussion, there are no examples of any albums that are well respected for having good dynamic range in this day and age. I was reading the Rolling Stone magazine yesterday and it quoted Robert Palmer and Alison Krauss's new album as being particularly good. Can anybody find any others? Alexthecheese (talk) 11:44, 21 January 2008 (UTC)

This isn't a shopping guide, and endorsing releases (among thousands of available CDs) is probably not a good idea. The point of this article is that newer CDs (including some or many marketed as "remastered") have this problem. For what it's worth, the 1983 Japanese Abbey Road on Toshiba-EMI is sought by Beatles collectors for being well mastered; a similar Toshiba release of Dark Side Of The Moon exists from that period. edg ☺ ☭ 13:19, 21 January 2008 (UTC)

I agree, we can't be starting a list of "full dynamic range mastered CD releases". There are several thousands of them out there. The article is called the "Loudness War" and that should remain the main topic. Unfortunately, it's hard to find in today's releases, good pop CDs. They still have an element of loudness in almost all of them. Case in point, the 2004 release of Brian Wilson's "Smile" album. It blows out of the water the 2001 re-mastered "Smiley Smile" Beach Boys album, both in sound quality as well as in musical content. That CD is what Brian wanted all along back in 1966. Yet, the mastering guy for "Smile" compressed the loud parts to make the average levels sound louder. However, compared to the remastered "Smiley Smile", the "Smile" album sounds full with dynamics and a perfect clean sound as all the signal to noise floor level issues from the original recordings are gone.

I personally think (and lots of people agree) that the label Capitol made 2001 remastered "Smiley Smile" sound worse than the original LP release in 1967 and the first CD version in 1990. Compared to Brian Wilson's "Smile", it sounds like crap (the mastering guy pushed the level and hammered it against the headroom's ceiling, emphasizing the noise problem from the original recordings) and everyone should return those CDs to the label for a full money refund. Jrod2 (talk) 17:13, 21 January 2008 (UTC)

I concur that a list of good albums isn't what this article is about. Alexthecheese, I suggest for you that albums which have won awards for Best Engineered and such would be among the top tier of dynamic and well-mastered albums, though a degree of flirting with the edge of hard-limited loudness will be observed in several of these recordings starting from about '95 on. It's a start, anyway. Binksternet (talk) 18:03, 21 January 2008 (UTC)

They gave an award to Hail to the Thief!! That is brick wall limited, obviously they don't care about compression, looks like its not the audiophile award you would expect it to be, unless the award was based on the vinyl version? AJUK ^Talk!! 12:02, 18 March 2008 (UTC)

I think it would be a much better an example that despite not being compressed was still hugely successful, ie Franz Ferdinand 77.99.57.229 (talk) 11:42, 1 March 2008 (UTC)

I listed a few a few months ago, no one used them. 77.99.57.229 (talk) 12:11, 26 January 2008 (UTC)

Motown not the instigator

Well ... this discussion certainly throws a whole new light on the old Gordy Records slogan "It's what's in the grooves that count(s)". However:

Years before the Motown label existed, Imperial Records was in the habit of cutting their singles very "hot". Compare an original 45 of Ricky Nelson's "Be-Bop Baby" (or the Majors' "A Wonderful Dream", or the Showmen's "It Will Stand") with its contemporaries - the difference is blatant.

Invictus Records (where the famed H-D-H team settled after defecting from Motown) boasted openly on its labels of a sound tailored to stand out on AM radio. (someone please pull out an old Chairmen of the Board single and supply the exact wording) My guess is that the secret recipe included serious compression, along with some brightening of frequencies above 3 kHz. —Preceding unsigned comment added by 209.155.146.2 (talk) 01:25, 23 January 2008 (UTC)

That is fascinating, but sounds like Original Research to me. The reason motown is listed is because one of our resources lists it as an offender. That would be awesome if you found some resources that backed up your information. Feel free to add it if you do. Illuminatedwax (talk) 19:16, 29 January 2008 (UTC)

a seminar in britain?

I ran across a news article [2] that "a seminar against loudness war is arranged in England next Tuesday." The article continues to cover the loudness war but doesn't give any further detail on the seminar (it is implied that it is for professionals). I couldn't find any more information about this on a quick google. This could be a good addition to the article. Anyone have more information? --MinorContributor (talk) 09:51, 5 March 2008 (UTC)

I think this is it. --MinorContributor (talk) 21:11, 6 March 2008 (UTC)

Broadcast compression not part of Loudness War

Right from the start there is a mis-characterization of the term 'Loudness War'. While it is true that Broadcast practices compound the problem, the term refers solely to the choices made during the process of producing, mixing, mastering, engineering (whatever you want to call it) the original recording. And it is not that compression is bad, it is that over-compression is bad. The distinction seems arbitrary until you look at an illustration such as the excellent one at the top of the page. Also, the point of the Loudness War is to call into question the motive for compression. If it is an artistic one, it is OK. Rock music as a genre is always compressed more than classical music; it is supposed to be. Rock music is loud on every beat, but there should be room for an extra loud part of the song. The current practice is to eliminate that extra loud part of the song by making every part equally loud.

If you look at the sources, you will see that there is a tacit acknowledgment that it is necessary to compress Broadcasts due to the limited dynamic range of a Broadcast. The mistake is in comparing the music being mixed in the studios to how the current hits of the day sound over the radio. Apples and Oranges. Trying to make it sound in the studio as it will over the radio is a mistake. A mistake which will make it sound worse when it gets played over the radio later on. Now there are other motivations to over-compress, I am just addressing how Broadcasting get thrown in the mix.

Stephen —Preceding unsigned comment added by 67.190.211.111 (talk) 04:53, 8 March 2008 (UTC)

Actually, if you check the links given, you'll see that the term "loudness war" is widely used to describe the effect of different radio stations each trying to be louder than their competition. --Slashme (talk) 08:16, 19 March 2008 (UTC)

I guess my retort is that radio stations have to compress because they have a lower signal to noise ratio. The concern nowadays is that digital compression is more powerful than any form of compression from the past and does not have any built-in constraints. This medium which has the potential for greater dynamic range also has the potential to offer less dynamic range than any other medium. I think that if you check the links you will see that the main concern is that if you compress music as much or more than the radio will with its own compressors, the radio's compressors won't know what to do with the signal and will mess it up. The concern is not that the broadcasters over-compress, but that they need a relatively uncompressed signal to work with.

Stephen —Preceding unsigned comment added by 67.190.211.111 (talk) 13:57, 21 March 2008 (UTC)

The "Loudness War" started in broadcasting and spread to the recording companies. The rationale given in broadcasting (and while it's OR, I was there, then) was that if our signal sounded louder (even if it didn't sound as good) the station could make more money for its commercials. Those making the decisions to increase compression had no understanding of signal-to-noise or distortion. htom (talk) 17:11, 21 March 2008 (UTC)

Not completely true. Radio stations also use those so-called Optimods to do a sort of station branding, i. e. to make their station recognizable and distinguishable from the others. This is called sound processing, which has already aroused some anti-movement by people who do not like this. This effectively means, that a good ol' 80's song might sound in over 10 different versions over the radio: due to additional compression (yes additionally to the overcompressed audio source - call it AUDIO TORTURE!), exciters, stereo broadening tricks, and sometimes /(reportedly) slight reverberation applied. -andy 78.51.74.120 (talk) 08:37, 23 April 2008 (UTC)

Examples

What are the examples of? Albums that "break" this mold? Albums that verify this? This needs explanaiton. -24.6.56.121 (talk) 07:23, 18 March 2008 (UTC)

Yes, I tend to think that you have to demonstrate that CDs can be mastered without digital distortion. That is CDs that have not succumbed to the Loudness Wars. Trying to be louder that your competitors is OK unless you destroy your music in the process, which is what is happening today. Even my recent McCartney CDs are distorted. Anyone like me who listens to music and asks himself what type of Bass he was using, what does he have in the drumset will realize that these distinctions are being lost.

Stephen —Preceding unsigned comment added by 67.190.211.111 (talk) 13:45, 21 March 2008 (UTC)

Another useful source

A good piece on this that I notice is not cited in the article: Douglas Wolk, "Compressing Pop: How Your Favorite Song Got Squished", p. 212-222 in This is Pop (2004), Eric Weisbard (ed.), Harvard University Press, ISBN 0674013441. If someone is working further on this article, wants to read this, and cannot access a copy of the book, feel free to get hold of me (email, please), and I'll scan the relevant 11 pages and send them to you. (Not particularly a topic I want to write on.) - Jmabel | Talk 05:52, 30 March 2008 (UTC)

MP3s etc.

I'm a little surprised that the article does not really discuss that MP3s etc. carry even less information per time than CDs, so that as more and more people listen to music via their iPods and the like, the temptation to ignore high fidelity must get even greater for the sound engineers. I don't have anything citable on that, though it must be out there; if someone does, it would make a good addition to the article. - Jmabel | Talk 05:59, 30 March 2008 (UTC)

I respectfully disagree. Up to a point, increasing sound pressure level at playback increases the detail one can hear in music. Loudness war methods exist that can retain the high frequency detail while maximizing level within a hard limit (such as 0 dBFS.) MP3s, on the other hand, sacrifice detail and nuance for smaller data sizes. When you turn an MP3 up louder, you don't hear more detail. I tend to think these are two completely different topics. Binksternet (talk) 08:18, 30 March 2008 (UTC)

Well then, whatever there is (citably) to say about how MP3s etc. fit the picture still belongs in the article, because they are becoming how more an more of the world hears music. - Jmabel | Talk 17:22, 30 March 2008 (UTC)

Be bold!  :-) If you have something to add, add it. Perhaps the Other formats section is the place to start. Binksternet (talk) 18:20, 30 March 2008 (UTC)

I may be bold, but I'm not bold (or foolhardy) enough to make substantive edits on a highly technical topic where I have only layman's knowledge. - Jmabel | Talk 23:39, 30 March 2008 (UTC)

I appreciate the caution. MP3 is not a direct cause of compression and clipping. The audiophile objection to lossy compression is reasonable (up to a point), but that particular complaint is not relevant to this article. What might be relevant is the possibility (alluded to in the Rolling Stone article, if anyone wants to search for a source) that industry types consider louder files optimal for MP3 listening. / edg ☺ ☭ 00:28, 31 March 2008 (UTC) (currently listening to Tool in ogg vorbis, q5)

People often confuse dynamic range compression with the compression used to produce lossy files. The two can't be compared by any means, because of their entirely different natures. This article discusses the use of mastering techniques used to artificially control the loudness of a record, and that has nothing to do with the format of the sound or anything similar.

Those who despise lossy compression demonstrate their lack of knowledge about it. When the lossy file reaches its point of transparency (wich varies from person to person), it is virtually impossible to tell the difference between the compressed file and the original. For most people transparency can be reached at 192 kbps using constant bitrate (personally I can't understand why people are still using cbr). Improper encoding is the reason why a lot of people think of mp3s as an equivalent of bad quality, contempt against them (mp3 files) has no objective basis whatsoever. Besides, there are superior lossy formats if one is concerned about quality. Edd3 (talk) 20:58, 21 April 2008 (UTC)

Thanks for clearing that up. I should have said MP3 is not a direct cause of dynamic range compression and clipping. Not distinguishing that from lossy compression made my comment rather unclear. Perhaps we should make that distinction explicit in the article as well. / edg ☺ ☭ 21:48, 21 April 2008 (UTC)

MP3's use of a lossy compression algorithm is designed to greatly reduce the amount of data required to represent the audio recording and still sound like a faithful reproduction of the original uncompressed audio. Whether your conversion fills up all 8 or 16 bits of resolution, your headroom is unchanged, thus, loudness has nothing to do with the data lost or the perceived loss of sound quality. In addition, Kbps sampling rates are irrelevant to loudness as well. Jrod2 (talk) 04:44, 22 April 2008 (UTC)

And yet tracks with large amounts of dynamic range compression won't encode as transparently to MP3 (or AAC, ATRAC, WMA or any other lossy format). Even a CBR file is "variable" of sorts, in any given fixed size frame there will be more bits allocated to certain frequency bands. A snare hit with a transient followed by rattling snare wires will get allocated more bandwidth than say the bass (which needs less bandwidth to encode).

Unfortunately when everything's compressed into 1dB of dynamic range there's no way for the encoder to tell what needs more bandwidth. I did a little test encoding the 1981 and 2005 versions of ABBA's "One Of Us" (from The Visitors), the same track I used for screenshots, using lame in VBR mode (flags -q0 -b32 -V0 -mj). I'm not at home to check the actual figures at present but the average bitrate on the uncompressed 1981 master was significantly lower than the heavily compressed 2005 remaster. From memory the average bitrate was around 220kbit for the 1981 version and 280kbit for the 2005 version.

I read an article recently that explains this but a lot of it I belive is to do with the MP3 encoder seeing the heavily compressed signal as noise and therefore being unable to drop the bitrate on those frames. Material with a wider dynamic range also compresses better (in terms of psychoacoustic rather than dynamic compression) as it enables the encoder to utilise the bit reservoir more effectively.

I'll stop there as this is starting to get into the realms of original research, but I'll have another dig around for that article. It may even have been one of the links from this page. Squirrel (talk) 08:22, 23 April 2008 (UTC)

Music Types

Is it worth mentioning that this effect is slowly working its way into music types it never used to effect very much, I have rhythm and Stealth by Leftfield and Play by Moby both from around 1999/2000 and they are both really dynamic. AJUK ^Talk!! 13:00, 31 March 2008 (UTC)

Thanks

I just want to express my gratitude to the writers of this article. It is very much possible that without this article I had never realized why listening to music is not as enjoyable as it was before. Fortunately I read the article. Now at least one reason is obvious, yet more of them may be hiding. I realized that I have some commercial music CD copies that are technically inferior compared to the real potential of the CD according to the standard. I don't think that it is my fault that I expected approximately all new CD copies to be at least as high quality as the old vinyl LP copies. I expected that, but I was wrong. However the real error was done in the record industry. Consumers are victims of the greed of the record industry. 130.230.31.119 (talk) 21:44, 13 April 2008 (UTC)

Not only are we victim of their greed but their madness. Jrod2 (talk) 22:36, 13 April 2008 (UTC)

Yeah, as I said before they sell without this treatment anyway. AJUK ^Talk!! 19:14, 20 April 2008 (UTC)

Thanks from me too. Some time ago, I had started a German translation of this article, which has greatly evolved from "beta" stage the last weeks/months. There are a lot of German firms who produce hi-fi equipment for audiophiles, this is a good audience to address to. -andy 78.51.74.120 (talk) 08:44, 23 April 2008 (UTC)