in ,

Garbage In, Garbage Out

  • Garbage in, garbage out. I’ll bet you’ve used this phrase countless number of times. In fact, these four little words have collectively become one of the few generally accepted tenets of the audiophile world, especially among we digital folk. Now in case you haven’t ever heard this expression in context before, let me explain: it is typically used by audiophiles as a trite way to point out that the weakest link in your digital playback chain can obviate the benefits of the expensive analogue gear behind it.

    For example, I’ve heard audiophiles leverage this phrase to justify the cost of upgrading their DAC, since if the conversion from 1’s and 0’s to electrons isn’t up to snuff, it really doesn’t matter how much money you’ve spent on your favorite transducer of choice (personally, I’m partial to the HiFiMAN HE-1000).

    Garbage in, garbage out. I’ve also heard this phrase used to describe why you shouldn’t ever enjoy the sound of low-res MP3s or any other type of lossy format. Again, if your source material is missing bits to begin with then you’ve already lost a modicum of fidelity before you even press play. GIGO! But over the last decade or so, these words have taken on a whole new meaning for me, and as a result, my perspective on this hobby has changed dramatically. Hopefully by the end of this article you’ll understand why.


    Full disclosure: to this day, I can’t really fully explain why I love metal as much as I do. Even my wife is sometimes confused by it. I know that I gave my heart to Megadeth and Type O Negative in my teenage years. GWAR got me through my freshmen year in college. And Death was my constant study companion in graduate school. These days, when I write for my own blog, I listen to mainly black and death; sub-genres of metal that most audiophiles I meet consider obnoxious noise.

    And before you write me off as someone who doesn’t appreciate “good” music like classical or jazz, I assure you I do. My mother went to Julliard (harp) and was my early music teacher. My great-uncle (clarinet) and aunt (harp) also went to Julliard and played for the Met and the New York Philharmonic respectively. Up until I was fifteen years old, I was being groomed to follow in their footsteps. This year alone I’ve seen five operas (Otello, Tannhauser, Tosca, Lakmé, and Figaro) and three orchestral works. (I’m writing this article to a live broadcast of Rigoletto right now). As you can see, I adore classical music as well. But despite my love for Wagner, Verdi, and Mozart, my heart belongs to metal, and I’m perfectly fine with that.

    However, my love affair with heavy metal posses a big dilemma for me as a self-professed audiophile. Why? Because unlike classical and it’s “American” variant, jazz, where high-fidelity reigns supreme, heavy metal is a form of popular music (another term audiophile trade shows still don’t quite grasp yet), and as such is produced with a very different goal in mind, namely volume. The louder, the better. At all costs. Unfortunately, it’s not just metal either; all genres of popular music today including pop, rock, rap, EDM, you name it, are subject to the same production constraints as metal. The rise in prominence of this kind of production style is dubiously known as the Loudness War, which began life in the early ’90s and has been single-handedly destroying popular music ever since.


    Despite being widely regarded as an artefact of the digital age, the Loudness War is really a by product of the golden era of vinyl. Back in the ’40s and ’50s, there weren’t any kind of broadcast standards to normalise loudness on the radio like there are today. That meant different songs played on the radio were all heard at different loudness levels. Songs played from vinyl records pressed from hot (loud) masters would sound louder than their soft (quiet) counterparts. Consequently, the volume knob in your grandfather’s Edsel would get a lot of heavy rotation (literally) when his favourite Motown tune came on. Record labels of course knew this and leveraged it to their full advantage, trying to drown each other out through volume. The idea was (and largely still is) that if one label’s roster of artists sounded louder on the radio than the others, people were more likely to gravitate toward the louder set.

    Ultimately, this first round of the Loudness War reached its peak in the ’70s, when the limits of the vinyl format itself prevented labels from producing records any louder. But then the digital music resolution happened, and with it, a new and more brutal Loudness War took its place.


    As a format, the CD is superior to vinyl in practically every measurable way. And by the early ’90s, it was pretty clear that the CD and more generally speaking, digital music, would eventually become the defacto high-fidelity format.

    The first decade of the CD showed no signs of the Loudness War re-emerging, no matter the genre. From Megadeth to Madonna to even Massive Attack, CDs were mastered with fidelity first, volume second in those early golden years. But with the advent of the digital audio workstation (DAW) coupled with the fact that the CD wasn’t subject to the same intrinsic loudness limitations as vinyl, it was only a matter of time before CDs were being produced louder and louder. In fact, by the end of the ’90s, most forms of popular music were being mastered to previously unprecedented levels of volume. And though the world somehow managed to survive the Y2K problem, transients and dynamics in mainstream music unfortunately did not.

    This second Loudness War reached a breaking point with the now infamous 2008 release of Metallica’s Death Magnetic, which had its volume level pushed beyond the point of digital clipping to cause audible distortion! But here is the sad truth: by 2008, Death Magnetic was no different than many, many other well-known Loudness War causalities before it. The Red Hot Chilli Peppers released Californication in 1999, almost a decade before Death Magnetic’s release, and it too was pushed to within inches of its sonic life. Same is true for Oasis’ (What’s the Story) Morning Glory? which was released in 1995 and also heavily compressed. And the list goes on and on.


    Most Loudness War discussions centre on dynamic range compression (DRC), a technique used to reduce a track’s overall dynamic range, or the ratio of the softest sound to the loudest. This is done to in pursuit of maximum loudness. But DRC is not evil in itself, only its misuse. Which begs the question: why would an engineer ever want to use DRC anyway? Because not all sonic components of an individual track maybe recorded at the same volume level.

    For instance, vocal components maybe recorded at a much louder or softer level than their accompanying instrumental counterparts and vice versa. So in an effort to achieve volume homogeneity across all of these components, a mastering engineer may apply a little DRC to a particular instrument or vocal section so that it sits more comfortably in the overall mix. This process is accomplished by making all the quiet parts louder while lowering or maintaining the louder parts at their respective levels. The net result is a more cohesive and pleasurable listening experience, since the loud sections don’t over power the soft ones allowing your ear drums to fully digest the aural goodness hitting them.

    However, DRC is typically used in combination with brickwall limiting, a technique used to increase the volume of the record to its absolute peak level (0dbFS) before unwanted distortion occurs. The end result is a homogeneous sense of loudness (whether the material warranted this artificial boost or not). And as an engineer compresses the level of dynamic range past a certain threshold using a brickwall limiter, the music starts to lose its vibrancy and sense of realism. In fact, certain frequencies can get squashed so badly that they become completely inaudible. Ever wonder why you can’t hear bass guitars or why cymbals and hi-hats sound tinny and don’t reverb quite right on your favorite track? That’s typically the result of DRC coupled with brickwall limiting. And these production abuses are rampant in the industry; again, not just in the metal world either, but in all forms of popular music.


    So with all of the above in mind, how does one quantify whether or not an album is over compressed? Over the last few years, the audiophile community has used a software tool called a TT Meter, which measures the dynamic range (DR) of an album and pops out a score e.g. DR11 or DR5.5. DR = dynamic range.

    For proper context, DR6 is now the industry average and already considered by most sane engineers as too compressed. The recommended level according to most industry experts is DR8 or higher. Death Magnetic measured DR3. A bit of a fair warning though, a higher number doesn’t necessarily mean its sounds better, but in the overwhelming majority of cases, it usually does. I highly encourage you to both read my article on how to measure your own album collection as well as peruse the unofficial DR database online to get an even greater sense of how widespread this sonic epidemic really is. However, I can’t underscore this point enough: the application of DRC to a recording is not evil in itself, provided it’s performed judiciously. The fact is DRC is an invaluable production tool that can make a good sounding record sound even better. But because of the industry’s loudness-at-all-costs mentality fostered by the hunt for more ears (and therefore more sales), it is more often than not abused, irrevocably damaging the music in the process.


    Since the dawn of digital music, the audiophile community has been largely focused on playback, constantly arguing over sample rates, formats, and so forth, when in reality, the real focus should be on production, since a record’s fidelity has more to do with the choices made in the studio than what format it is ultimately delivered in. The fact is all recordings have a certain intrinsic production value to them, and if that value is very low to begin with, it makes no difference whether you play it back as MP3s, high-res FLAC, or even vinyl.

    It’s also why I’m not big into Pono or high-res audio in general. Not because I don’t think there aren’t any real audible benefits to be had with higher sampling rates, but because it fails to address the real problem facing fidelity in popular music today vis-à-vis its production. No format can fix the fact that Adele’s recently released 25 is a DR5 compressed nightmare. In other words, garbage in, garbage out.

    Further reading: Metal-Fi

    Alex M-Fi

    Written by Alex M-Fi

    Alex is co-founder and Chief Editor of, a website dedicated to the head-banging audiophile. His blood type is Type O Negative. Alex derives his income from writing software.


    1. Cheers Alex! Seems like we are banging our heads against our collective wall trying to change this… Perhaps someone should start a new music download site… HDR Trax… High Dynamic Range, that is! Just for those who prefer the uncompressed truth.

    2. Thank you, Alex; that would explain why the free MP3 downloads with my online classical music shop’s newsletter can sound great.
      As for the so-called ‘loudness wars’, I simply cannot comment; I have no experience of this as I do not buy ‘rock and/or roll’, ‘metal’, ‘death’ or any of these types of…er…music.

    3. Alex – while I was reading the article, only thing going through my mind was leaving a comment citing a website (metal fi) that I recently discovered, and how it marries metal with audiophilia. ….. Something that I was searching for all the time. But as I completed reading the article I saw that the article has been authored Alex from metal fi. Well that comment is not needed anymore. I totally love metal fi and the way you guys blend hi fi with metal. Listening ‘dynamic metal’ thru a high quality rig is an absolutely rewarding experience. I agree with each and every word you have said here.

    4. A couple of notes: 1. There is a difference between volume (RMS and peak levels on a recording) and DR (the difference between those two values). Vinyl has clear *volume* limitations, particularly in low frequencies. You can only go so loud before the needle can no longer track. Also, the louder your record, the less time you can have per side.

      There is however no minimum requirement for *dynamic range* on vinyl. You can smash your master to smithereens in terms of dynamic range, and press it to vinyl without any problem. All you have to do is reduce the peak levels. This is how vinyl can be cut from ultra loud, dynamically crushed CD masters that ride 0dBFS the entire time. You can’t cut the vinyl at that level, so the cutting engineer will just turn down the whole thing.

      2. While mastering engineers usually get the blame for Loudness War disasters, the mixing engineers can also be partially or entirely at fault. The engineers that worked on Death Magnetic for example claim that the mix was already destroyed when they got it, and there was basically nothing they could do.

      A crushed mix without any extra limiting applied by the mastering engineer MAY also appear “dynamic” as read by the TT DR meter, but actually sound like crap. Adele’s new album seems to be one of those examples. The vinyl has a dedicated master presumably with less limiting than was done on the CD, but it sounds pretty much the same – lousy. I suspect the mix was already crushed.

    5. GIGO for equipment made a lot of sense in the days of record players and tape decks, when their significant front-end deficiencies got grossly magnified by the amps. But digital front-end equipment made the saying irrelevant. Now it is mostly relevant to how well recordings are mastered. Alex got that right.

    6. Wow, this article makes a lot of sense and ecplains that many audiophiles tend to listen to ‘strange’ artists and music which is less truncted by DR.. What can be done to rescue real good music? Is this way of ‘garbaging’ musïc allready pesent on the original mastertapes, or is it done in the mixing and master processing afterwards?
      In other words, will Bob Stuarts MQA software unlock true mastertape sound if it will be allowed to run from 1st reel mastertape direct? Or will this create other problems and will the music ‘miss the producer’s fingerprint and work and DR and other ways of colouration of the original recording and thus will never be allowed to be released..?

      • Hey Peter, it’s unclear to me yet what MQA *really* brings to the table. However, at the end of the day, and as I said, it doesn’t matter what ultimate format your favorite set of digital bits arrives in – fidelity has to be managed at the recording stage or all bets are off.

    7. Spot on with your comments. It’s all to easy with digital mixing and editing suites to strangle the life out of a good tune. I despair at the general music listening audiences ability to listen to absolute s#ite quite happily even when an alternative is demonstrated to them. I found the aforementioned Oasis and RHCP albums simply impossible to listen to for years until I spend serious cash on upgrades, even now I feel like hunting the producer down and slashing his car tyres for such a awful job as even decent kit can’t repair compression of dynamic range.
      I have a few hi res albums that you need to get off the sofa to twist the volume pot up quite a bit if they are in the middle of a playlist with albums from the 90’s onwards but boy is it worth it for the real feel of dynamic rage. I’ve less than half a dozen albums of which I would say have been recorded well.

      Rant over, nice article.

    8. Thank you for being a voice for those who love good sound but don’t necessarily subscribe to the typical audiophile music genres/artists. I love Tool, Porcupine Tree, and much of Steven Wilson’s post Porcupine Tree music. I also love many types of electronica. This seems to be an emerging group that the high end audio industry may want to pay attention to in the coming years. The group that listens to music that has been treated with brick wall filtering and heavy compression. That type of music does not do well with the typical high end audio component. Listeners of non audiophile music would still like it to sound as good as possible, but this music has different playback requirements than most audiophile recordings.

      Perhaps a return to the days of the “Loudness” button is in order. A “Loudness Wars” button on our preamp or integrated amplifier that would make long term listening to heavily compressed, brick wall filtered music possible. 🙂

      The alternative is educating those who are in control of the compression and brick wall filters in the post production process.

    9. Great reading Alex,
      I believe this should be massively promoted. There should be billboards out there and webpages/blacklists in here (internet) with names of garbage recordings and responsible people, hopefully badly affecting sales or at least causing shame :).
      Maybe then both artists and record companies realizes that there are people who care…

      • Let me say this: I think Neil Young and Pono could have been that voice. Instead, they are focused on a problem that in my opinion does not exist. However, that’s for another day. Stay tuned Palo, and thanks for reading.

    10. That’s why I’m often bemused by many of the source and DAC makers. All I want is something that makes my garbage quality tunes sound less garbagey. Instead, I keep getting more DAPs that promise 24/192 or DSD shoved in my face. Ugh!! Instead of all those “taps” giving me a “clear window” to dirty water, how about one of you guys make me a decent water purifier instead?

    11. What a great piece – couldn’t agree more, and from a fellow aficionado of metal! It’s my love of the genre that initially led me to purchasing better gear, not all speakers/headphones can keep up with the music I like – Meshuggah, Nile, Gojira for example. To that end, my LCD 2’s have served me very well – they even make the poor recordings listenable. Nice work mate, I’m gunna check out your website.

      • Thanks Duva. I still have a pair of LCD-3s (sans fazor) that were my go to cans for all of my reviews. The HE1K has spoiled me. Rotten.

    12. I believe that the neural networks of listeners adjust and settle to a personal standard, so that the ‘best sounding’ encoding will be different for each person.
      With storage so cheap, why not develop a standard that encodes both ways? (or many ways). The listener can then select Type A or Type B (etc) listening from the medium, instead of having to adjust volume, dynamics etc with each track.
      Decoding doesn’t end at the ear.

      • That is a very interesting suggestion. Though it puts more burden on the player, and then of course you have to standardize these effects so different players don’t sound vastly different because of the way for instance, its compressor is implemented.

    Tokyolo #4: Disk Union = record store heaven

    e-earphone – Japan’s number one headphone retailer