If you are not aware of it, at this very moment, war is being waged on the consumer in terms of loudness. What do I mean by that?

Simply stated, there is a very strong push within the marketing arms of the music industry body to make music (or as many industry types refer to it, "product") seemingly more enticing by making the music louder. The idea here, and it's certainly not conjecture, is that most consumers, when presented with two versions of a recording, will prefer the louder version - and...by louder I mean that one will play louder for a given setting of the volume (gain) control on the preamp / receiver / mp3 player.

This phenomenon is not new to the audio community. As many of us know, when comparing speakers (or headphones for that matter), it's important to normalize one speaker's loudness to the other. In point of fact, I'm really talking about normalizing their efficiencies which then manifest themselves as loudness. For example, if when comparing two speakers, and one of them is 10 dB more efficient than the other (i.e. the SPL observed at 1m with 2.8 VRMS input of speaker "B" is 10 dB higher than that of speaker "A" under the same conditions), the tendency, especially over short listening intervals, is to prefer the louder of the two speakers, primarily because unless normalized, the speaker that is 10 dB more efficient will be perceived as twice as loud as the other.

This is human nature, and in fact, we probably all know that a speaker's efficiency is one and only one ingredient in a soup of factors that make a speaker palatable. I think we all have experienced this phenomenon at one time or another. Anyway, by normalizing the loudness (again, the efficiency) of the speakers, you're able to make a fair comparison between them because you are removing a source of bias (differences in loudness) and comparing more of the salient differences between them. In effect, you are comparing the timbre of loudspeaker "A" versus loudspeaker "B".

I'm not saying that efficiency is bad. All that I am saying is that is must be accounted for if fair comparisons are to be made. Also, I don't really want to go off-topic here, but I wanted to use the speaker analogy to make what follows a bit more understandable. So, let's get back to it.

What controls the apparent loudness (for a given setting of your gain (volume) control) in a digital recording? basically, it comes down to a few key factors (dynamic range etc), but fundamentally, the way in which one makes one version of a given track louder (at the source) is to re-scale the digital file.

Remember, in a digital file, the maximum value (that is, all bits hi) that can be achieved is 0 dB. That's the ceiling. Thus, the signal will always be some number of dBFS below 0 dB. How much the signal (let's just call it 'music' instead) is below 0 dBFS is really up to the recording Engineer and the Producer, but ultimately, it's the mastering Engineer who holds sway (as they have the final say as to gain, EQ, effects, etc). Depending upon the music, some music will spend its life no higher than 10, 15, or 20 dB down from 0 dBFS. Mind you, there are many ways to look at the value - we could consider its peak value (which is of great concern) as well as its mean (average) value, or any number of statistical analyses of the music (i.e. crest factor, percentiles, etc) pertaining to its relationship to 0 dBFS.

Let's revisit the way-back machine when the CD format was new (some of us remember that era).

At the time, many record companies simply re-issued LP titles (in CD format) from their catalogs without doing much to the signal at all (for better or worse). At the time, it (apparently) was the norm to respect the limits of the digital medium (at least in terms of full scale) and, like a physician's first axiom, "do no harm" to the signal.

As that market changed and mp3's et al became ever prevalent in the marketplace, many of these record companies sought to find sources of 'new' revenue based on their holdings. There's nothing wrong with that - the music business is, after all, a business. However, due to the changing landscape of the industry, the rise of the internet, and most importantly, affordable recording gear for 'indie' artists, the revenue streams were quickly going dry.

Here's where the trouble starts...

Company "A" holds the rights to Artist "B's" most popular recordings, and while "The Artist" may indeed receive revenue (Royalties...pardon me if this is not the correct legal jargon) from subsequent re-issues, "The Artist" may in fact have no say in how the re-issue is remastered.

Uh oh...there's trouble afoot...

What if Company "A" sees an opportunity to make something "old" become something "new" by virtue of remastering...and what if...perhaps...Comapny "A" ' s main interest is revenue and not necessarily fidelity?

Now, before people start 'hating', please keep in mind that remastering is by its very nature not necessarily a bad thing. There are many, many exceptiojnally talented mastering Engineers out there for whom the music is their "raison d'etre" and thus, show great respect for the music, the artist, and overall, for the artistic process. However, like most of us, we work for someone...someone who ultimately signs our paychecks, and as such, there are times where though we may disagree (even vehemently) with what's been asked of us, we do it because we have to do it (Milgram had something to say about this, but I digress).

Back to Company "A"...

So, Company "A" realizes a few things about the present market conditions... First, most people just don't care about fidelity, or at least, this is what they hold as a belief. Second, that by making the "old" become "new", the profit margin is huge - all the costs associated with having developed the original "product" have long since been recovered - coutless times over in fact.

So, Company "A" opens up the valuts and pulls from it a version of Artist "B" 's recording, and they notice that in order to get to a 'reasonable' (read "marketable") listening level with this version, the volume control has to be advanced rather high. The light bulb goes off...and the mastering engineer is instructed to raise its apparent loudness (or more likely "just...make it louder...").

So, the original version, which may have been issued with its music being - 15 dBFS is raised in magnitude by 15 dBFS, effectively utilizing all of the bits in the medium. By itself, this isn't necessarily bad, because at this point, the axiom of "first, do no harm" has been respected. Granted, the "new" is now a bit more than "twice as loud as the original (remember - for a given setting of the gain (volume) control), but the recording itself has not been otherwise altered; its dynamics have been left intact, no additional equalization has been done, and most importantly, no additional distortion has been caused by the process.

However...upon auditioning this first attempt...someone with control in Company "A" says ... "I like it, but can you make it a bit more [insert adjective of choice here]?"

This is where the real trouble starts. If the Re-mastering Engineer stands on principle and says "anything else that I could do to it would alter it, and most likely, not for the better". Mind you, I am not saying that remastering, with fidelity as the ultimate goal is a bad thing - not at all. What I am saying is that taking existing great recordings, driving them into distortion (literally, driving the digital medium into distortion - the very thing that digital was envisioned and engineered to prevent), and level-compressing their dyanmic range...well...those things are, in my opinion, bad.

Regrettably, more and more of this is occurring. I have noticed (in my own CD collection) that certain remastered issues of some of my favorite CDs actually sound worse than the originals, and often times for this very reason. But don't take my word for it - there is plenty of grist for this mill (i.e. "The Loudness War") out there on the web, you need only web-search, and you will find a mountain of posts and data on the subject.

Several of the websites (and if I can dig them up, I'll add them to this post) compare the original .wav files to the same song on a "remastered" disc, and in many, many instances, you can see that the remastered version has wavefors that are visibly clipped. I'm not joking - visibly clipped.

Granted, it does happen that the occasional extremely short-lived peak in some recordings does hit 0 dBFS, but that's the thing...you see, something that is visibly distorted may in fact never be heard...especially if it is a small fraction of a second as there isn't time for your ear to truly notice it (and even more so if other bits of the music will mask it). However, when the entire song spends most of its life as a clipped signal, then something is horribly wrong.

You can prove this for yourself if you have some editing tools, many of which are packaged with CD burning / ripping software (I know that Nero has an editor built in, and I suspect others such as Roxio or iTunes do as well).

So, if you have an original version of a CD and its remastered version, you can first extract ("rip") the original version of the .wav file, and then the remastered version of the wav file and compare them in the editor. If you see that they are more or less the same % full-scale, then chances are some noise reduction and EQ were the only things done during the remastering process. However, if you see that one is visibly larger (and thus louder) than the other file, then that recording has been re-scaled in an attempt to make it sound 'better' than the original by (as a minimum) making it louder as compared to the original.

Mind you...you are very likely (depending on your musical tastes - I suspect jazz and classical are less prone to these tactics, though I could be wrong) to see that some of the remastered versions of your favorite music are in fact being driven into distortion.

I have a problem with this, simply because, I like to think that the original release was how the Artist wanted his or her music to sound; if they had wanted it distorted and level-compressed, they certainly could have done so the first-go round. Granted, this is not iron-clad, and I'm also sure that there are many artists who wish that the recording technology of their day were capable of something different, but fundamentally, to me anyway, it's about respect for the artist and that vision.

I might think that the Mona Lisa would look better were she wearing more blush (just an example here...), but that doesn't give me the right to bust into the Louvre at night dressed in black and apply some to the painting just to suit my tastes. If Leonardo were around, and he wanted to do so, then he alone has the right, but if I didn't enjoy it as much after the fact, I could choose no longer to gaze upon said painting.

The problem is pretty big though, because the "Remastered" label generally invokes a response in the customer that it must sound better than the original, because logically enough, why would they remaster it if it was only going to make it sound worse?.

There has been a lot of talk recently in the industry (with a lot of push-back from the record companies) to label remastered versions with indicators of their fidelity. That is, things such as mean dBFS levels, crest factors (the relationship between peak and RMS levels), but so far, none of these seem to be gaining traction. Moreover, I am not sure just how many consumers would be anything but befuddled by such indictors on the packaging.

UPDATE: Here's a link that does a pretty nice job of explaining (graphically) what's going on: http://spectrum.ieee.org/computing/software/the-future-of-music

old_school_2'

That was a very interesting post. To that end I, thought you might be interested in a short paper on this issue that I delivered at CES to a joint meeting of the CEA Audio board and the Grammy awards Technical commitee. I delivered the paper in the role of my "day job" at Atlantic Technology:



A Major Issue Facing the Recording Industry in 2012 and Beyond
© 2012 Peter Tribeman, Atlantic Technology. All rights reserved.


Throughout the history of the modern music recording industry, one theme has been consistent: Each succeeding recorded medium has delivered markedly superior sound quality compared to the one it replaced. As the 78 was replaced by the LP and the LP was replaced by the CD, each time there was a significant improvement in the frequency range, signal-to-noise ratio, distortion, and dynamic range of the new medium.

There have certainly been “side media” along the way, like 8-tracks and cassettes, which may not have advanced the state-of-the-art from a sound standpoint. But the main recorded medium in widest use at any given time has always featured a major sound quality improvement among its central virtues.

Sadly, today, this has come to an end for the first time since recorded popular music has been a major social/economic factor in society.

Today’s downloaded, disjointed “songs” (virtually no one listens to entire albums any more and longer-length classical works are pretty much ignored) place a far higher premium on user convenience and accessibility than on ultimate sound quality.

A further indication that sound quality is nowhere near the top of consumers’ priority lists is evidenced in the practice of downloading at lower sample bit rates in order to maximize the available memory space on their chosen playback device. Today’s recording industry is demonstrably uninterested in offering widely-available commercial recordings that can legitimately claim “lifelike, accurate” sound as an attribute.

One of the biggest impediments to offering recordings with a wide frequency range and a wide dynamic range is the very real limitation of the environment and conditions under which much of today’s music is played back.

The dynamic range in today’s recordings has been compressed into near non-existence, for purposes of seeming “louder” when played back on TV and radio: As observed on http://tvtropes.org/pmwiki/pmwiki.php/Main/RecordOfLoudnessWar :

“The absolute peak of loudness started slowly creeping up in 1995, when Vlado Meller mastered Oasis' (What's the Story) Morning Glory? to -8 dB RMS. Since then the tendency has been to make records louder, while Meller has gone on to ruin more albums like Californication. The resulting loudness war is due to a variety of factors, such as commercial concerns, stupid executives, following the leader or listeners / musicians who are unaware of this phenomenon, can't tell the difference and don't care, or actually LIKE how it sounds.

Most of today’s music is experienced under what would be considered “portable,” “mobile,” or “background” situations. The days of sitting down in front of one’s “stereo system” on a Saturday afternoon and listening intently for hours in a quiet living room to the latest “audiophile” recordings is long past. Decades passed.

These days, people listen to their iPods while on the treadmill at the gym or walking from their office to lunch around the corner, or when driving their car. Home listening consists mostly of docking one’s musical storage/playback device in a low-fi plastic “docking station” with 3-inch speakers, simply to fill the background with indiscernible sound while cleaning the kitchen after dinner.

All of these common listening scenarios have high noise floors, which mask frequency extremes and render wide dynamic range recordings unlistenable by virtue of the high inherent noise floor swamping the low-level passages of the program material.

The impatient, inattentive, convenience-oriented nature of today’s music consumer, coupled with the typical noisy environment in which music is “consumed” leads to wide frequency range/wide dynamic range recordings being not only unappreciated, but in fact, outright unusable for the average listener.
There is a segment of the music-listening market to whom recording quality still matters. And many people would opt for better sound once they experienced it if it were readily available.

Should the recording industry ignore this segment altogether? Is there a way to encompass both the quality-conscious and casual music listener?

Fortunately, the solution is easily implemented and technologically feasible. My suggestion is that recordings/downloads be populated with two different versions:

1. A “popular” version with the normal compressed dynamic range suitable for use in cars, gyms, and other noisy environments, and

2. A second, selectable “audiophile” track with uncompressed wide dynamic range, lower distortion and no “brick wall” clipping.

This will satisfy both market segments and broadens the music’s appeal well past its current consuming audience.

Peter T:

You make a lot of salient points... and I think your suggestion as to 'branding' (or delineating) the type of recording (in terms of fidelity borne of the dBFS criteria and such) may actually be a very simple but effective way to hedge our collective bets on the future of fidelity. There is one slight danger though, and that is bands that, despite having the technology available to make clean, uncompressed recordings, choose deliberately to make a heavily-compressed and distorted file. In such cases, unfortunately, there would probably not be such an immediate audible difference between the heavily compressed (i.e. via codec) version and the full fidelity version.

Nevertheless, one of the things that is at the core of your suggestion, if not otherwise overtly stated, is an answer of sorts to one of longest-standing complaints about the world of compressed audio: Truth in Advertising.

I have groused about this issue in the forum that I host on another site, but I'll breifly re-hash it here: the trend...going several years back...to call a low bitrate compressed file "CD quality" and... to sell it as such. Frankly, as an Electrical Engineer and as someone who has spent most of his career in signal processing, I see this as nothing short of fraud. I know... strong words, but I really feel that way.

Ironically, my background / training / career make the math-based descriptors (i.e. track by track crest level, mean dBFS level and so forth clear to me, but alas, Greek to the average consumer. Thus, this is why I think your idea could actually work.

That is, it gives the Record Companies what they want (revenue), whether working from a remastered back-catalog release or a new release, as well as giving the consumer a bona fide choice, and indeed, they could be priced differently (and while philosophically I'm opposed to that, I would accept it as a sort of Faustian pact if the end result was that truth in advertising were to be achieved). Even if priced alike (and let's face it, somewhere there must be a model / analysis by the Finance people about bandwidth / storage costs and utilization rates) the important take away from this idea is that indeed, customers are not misled, but are given a clear-cut and fact-based choice.

Speaking personally, given a choice between an album in FLAC at a 30 - 40% price premium over an mp3 variant, I would take the FLAC version almost 100% of the time (the notable exception, for me, being content that doesn't really benefit full fidelity (I'm thinking about some of the earliest recordings such as those from an Edison phonograph etc)). Otherwise, I would opt for FLAC every time. Regrettably though, the world's largest purveyor of brand-named mp3 players have chosen not to support FLAC, but instead, their own proprietary version of something (at least by what I have been able to figure out to date) that is very much akin to FLAC (but FLAC, being open-source, spells revenue trouble...doesn't it?). On the other hand though, they do indeed offer what appears to be an uncompressed (rather, fully reconstructable) version at a price premium, so I actually applaud them for this. Yet, on the other hand, said company has recently introduced their 'plug in' to recording studios in order to ensure that the Mastering of said recording has been "optimized" for their music outlet. As someone who works in sound quality and live recording, this seems wholly ludicrous (but that's another topic, and another post, which I'll probably address in the coming days).

Again, having worked for many years in the field of signal processing and binaural methodology, accuracy and realism have always been central themes of import for me, and while I realize that that which I do makes me (perhaps) overly sensitive to fidelity, nevertheless, I still feel very strongly that consumers should at least be allowed to choose...and more importantly, to know what they are buying. As a minimum, they should not be lied to.

As far as the other points pertaining to listening to a track rather than an entire album, I don't know that will ever change. Let's face it, some albums come across better when listened to as a unit, but others are indeed spotty, and for this reason, only "the hits" are likely to be listened to with any real degree of frequency.

Mark

I agree that the advent of the various compression and equalization schemes and even the much adored RiAA curves used on records are all variants that are used to tell the consumer they are too ignorant to choose the unaltered master file even if its quality is superior to the adulterated version. Giving the masses the ability to choose the higher quality recording would likely not change the status quo significantly but would be a boon to the "audiophile" who likes to keep the audio pure. Just knowing that the recording was unaltered would make the audiophile giddy on one hand and depressed on the other. Giddy to think they had the best possible source recording and depressed knowing it can never be reproduced perfectly. Nothing makes an audiophile happier than chasing his own tail.

As the cost of memory and other storage media relentlessly goes down, I think the need for compression (in terms of file sizes and songs fitting on a device) will go away. The uglier form of compression which enables incompetent sound engineers to cram a "louder" average signal on a given medium is a thornier problem that (sadly) seems to be getting worse and (even more sadly) that many younger folks don't even seem to care about.

My personal solution is to use FLAC for the home where I can slightly compress file sizes without any signal loss, allowing me to serve content off of a reasonably cheap hard drive. For "mobile" (car/gym/ipod/etc), I stick with 320kbit MP3 compression which is a decent balance of file size and fidelity.

Best,

Hmmm... well, I applaud you for your enthusiasm as well as your insights and delineation between compressed dynamic range (level compression) and digital compression. I think you are dead-on in terms of the 'economies of scale' of memory (especially on the portable front where SSHD'd solve a longe-standing issue (battery energy density)); there really is no 'need' to compress (frankly) to anything of a higher compression rate than what FLAC would achieve. Frankly, .wav would suffice given the onslaught of cheap memory, but one thing that FLAC has over .wav is that it adheres to the metadata tagging standard(s) whereas .wav really doesn't. So, it would seem given the points that you have made and the tagging / album art support that FLAC provides, the future looks bright.

However...

When you have artists who insist on partaking of the loudness war...and essentially destroying - to varying degrees - any semblance of fidelity by means or strident level compression, then in the ears of the typical consumer, FLAC would seem to have no real - check that - immediately noticeable improvement in fidelity over the lossy formats. I say this because, well, let's face it - most of the music that generates revenue has dynamic range nowhere near something like that of Holst's "The Planets". As such, when you compare a heavily compressed (level...EQ etc) track...as mastered by those who are willing (all too willing) to not only participate in but fully embrace the loudness war...in a lossy format versus a lossless format, then I would wager that most consumers would say "but...I really can;t hear a difference between the two versions...tell me why (again) I should download the bigger file...especially if I have to pay more...when in fact, it sounds no better than the lossy (or simply 'mp3') version".

You can't blame them for this...because truly, if it sounds like crap after mastering, it will still sound like crap in an uncompressed version. Therein lies the problem. On the other hand, classical / operatic content has and will likely continue to embrace formats such as FLAC because fundamentally, the market there is (more often than not) targeted to those who value the aesthetics of the music and appreciate the art inherent within.

Like any product, there truly has to be a demand, and for most pop / rock artists (especially those beholden to a 'big' label) the label calls the shots. This is why I applaud artists like Neil Young (and while he may get some of the technical matters wrong when he discusses the topic, it's forgivable because his cause is just, and his heart in the right place). Frankly, the rise of Indie 'labels' and Artists who can establish and maintain creative content rights for initial and subsequent reissues...they will be the only ones who, it seems to me, are capable of stemming the tide of the loudness war. For those artists, the fidelity of their music will be valued - as a minimum, because they refuse to level compress (unless perhaps for effect) and understand what music really is.

Who knows?

Agreed 100%. Sadly, there's far too much "popular music" these days that sounds like utter fertilizer right off the digital master so there's no perceived sonic benefit to avoiding file compression at all. Fortunately, I confine most of my purchases to content that maintains what I perceive to be decent sound quality, though it is becoming harder and harder with "big label" content.

Best,