Economics of the Great Compression

After getting a lot of generous and smart comments on my initial post about compression and mid scooping I decided to take another crack at it. The basic fact, described here, is that modern commercial music is heavily heavily compressed and has its midrange scooped out. The question I’m looking at is why.

The late Walter Sear, of Sear Sound studios in NYC, was a big believer in analog technology and he posted a series of articles on his website about recording and the technology and economics of the industry.

Sear claims compression started when radio broadcast engineers wanted to send their signal further. “By compressing the program material at the transmitter, their broadcast range could be increased, they could reach more listeners in the ‘fringe’ areas, and hence, could charge more to their advertisers.” Everybody pretty much seems to agree about that and that compression could make your station “pop” when someone twiddled the dial.

He goes on: “So, now that the example was set by the broadcast station limiters and compressors, the audio engineers discovered a new tool. Compress, compress and compress.”

Before that, recording engineers constantly turned the the signal level up or down:

And the engineers were trained musicians….You put the score across the [console] and you followed the score. And you knew the human voice — you knew the singer was going to run out of support at the end of a phrase, and you ran up the fader a little bit. And you saw that she had a high F to come in on and you pulled the fader down, knowing she was going to blast it. We were musicians. We were as much a part of the performance as the musicians.

Sear argued that modern use of compression marks the decline of skills. It’s a plausible argument: compression compensates for a musician’s poor control over dynamics. The better the musician, the better and more consistent his or her control over dynamics. No rock band gives the engineer a score: the aesthetic of rock and jazz and R+B prizes spontaneity over scripted dynamics. Sear is lamenting a decline of meaning of skill on the part of musicians and and on the part of engineers. A machine, the compressor, replaced individual skills.

The alternative says it’s a question of taste: that compression “just sounds better.” And really, Sear is kind of contradicting himself in that quote: if compression is so bad, why were engineers in effect doing it themselves, by “riding the faders?” His own account suggests that compression is just an automated version of what he was already doing, and that what he was doing sounded good.

But that’s not what he means. Sear also makes an argument that human hearing is not designed to emphasize loud, “steady state” noises. “When we were trying to survive in the jungle, it wasn’t the steady state wind noise that was important,” Sear says: “It was when the tiger stepped on the twig which snapped that was important for survival.”

I’m usually extremely skeptical of “when we were in the jungle” arguments. It’s just too easy to make that stuff up. But the “steady state” argument is kind of interesting. The steady state sound of a big waterfall fills the mind: it’s intense and overpowering. I wonder if the taste for compression comes from exactly the fact that it’s “steady state” and overwhelming? Because Sear isn’t wrong–what compression does is bring all the same components in a pieces of music to the same flat dynamic range–not the same relative volume, but the same dynamic range, so every instrument comes at you in the same steady roar. For a great example, listen to Lady GaGa’s Bad Romance.

He’s also right that setting up a machine is cheaper than hiring an engineer who’s also a musician, at a professional studio staffed by guys with physics degrees. That’s expensive. Hiring good musicians to play the songs on the Byrds’ first album, instead of using the Byrds themselves, was expensive.

The more modern engineering method, hiring one guy with a computer and software to record each person individually, fix their mistakes, and recombine it into a sonic whole, in which each instrument comes at a uniform level of compression, is much cheaper.  And the music that results is typically sonically forceful: it overwhelms the senses, like a waterfall.

I don’t know the exact reason all modern music is compressed, but it’s seems plausible that it’s a product of economics. If somebody’s primary exposure to music was songs on the radio, heavily compressed, then that listener wanted to hear that same sound when he bought the record. At each stage, compression both builds a taste for compression and makes production cheaper, which builds a taste for compression which makes production cheaper, and pretty soon you have a mass taste.

Economic causes mus be crucial, but also a desire to have music be overwhelming, so it blots everything else out and fill the mind. Whether that’s a good thing is hard to say. Does anybody other then the owners of recording studios yearn for the days when the only way to record was to rent really expensive studio time? I like cheap recording technology, but it’d be nice to walk away from the sonic waterfall now and then.

14 Comments

  • I’m digging your second at bat here. My feedback for this post goes like so:

    1) You’re wrong. I don’t know about you, but I don’t have a sophisticated understanding of mixing or mastering techniques. Still, I think it’s safe to say that producers are awesome. They aren’t overcompressing like crazy as you suggest and it seems a clumsy and sweeping insult to intimate that there is a “sonic waterfall” of music that isn’t being produced with exceptional delicacy.

    You make it sound like compression is an epidemic that has herded taste, sensitivity, and discretion off a cliff. What is likely closer to the truth is that much of today’s best pop music is produced with a very balanced use of all mixing and mastering tools.

    I suspect you are just sweeping together the dust of history in ways that don’t bunny up for you when it might be more accurate to just say you don’t like the production value of club music.

    2) Perhaps you’re also clueless and out of touch. I had a jazz history professor in 1994 who was trying to make an NBA reference and the best he could come up with was Kareem Abdul Jabbar. Do you listen to any music made nowadays? We need to make sure you’re not just making sweeping judgments that are based on 5 minutes of Kiss FM on the way to Harris Teeter.

    Let me know if the following tracks sound like somebody “compressed the shit out of them”. Or if you’ve heard any of these folks before. This is all shitty YouTube audio, but in terms of the mix we ought still be able to hear something of the original production value.

    Tightrope by Janelle Monae
    http://www.youtube.com/watch?v=pwnefUaKCbc

    Stillness is the Move by Dirty Projectors
    http://www.youtube.com/watch?v=YMPF6lpM0XM

    Stillness is the Move by Solange (this one suffers the most from YouTube’s file size compression but I’ll email you the mp3)
    http://www.youtube.com/watch?v=34STLHtu97A

    Nantes by Beirut
    http://www.youtube.com/watch?v=PCkT4K-hppE

    Cruisin’ by D’Angelo
    http://www.youtube.com/watch?v=JAfuUZRou7g

    Kilometer by Sebastian Tellier
    http://vimeo.com/3486497

    She’s Got Me Dancing by Tommy Sparks
    http://vimeo.com/3884738

    End Love by OK Go
    http://www.youtube.com/watch?v=V2fpgpanZAw

    3) Typo: “Eco­nomic causes mus be cru­cial”

  • Tom, our accordion player just sent this Wikipedia link in an email:
    http://en.wikipedia.org/wiki/The_loudness_war

    This is the strongest evidence I’ve seen in support of your compression plague argument, but I still have reservations. Those images of the Red Hot Chili Peppers song do show a change over time, but that song didn’t have much dynamic range in the performance to start with, did it?

    More importantly, most any track that has been normalized looks pinned to the top of the screen when you look at the waveform of the whole 3.5 minute song in an image that’s only 570 pixels wide.

    I just emailed you a few waveform screenshots of two tracks I like to give you a bit of counter evidence to the Loudness War waveforms from the Wikipedia article. One is Tightrope, the Janelle Monae track linked above, and the other is Get By by Talib Kweli. Even in the zoomed out versions there is more visible range, and when you look at just look at a 20-second segment the dynamics are even more obvious. Feel free to post these if you like.

  • That’s very interesting–we have to run out but I’ll post that later.

    Janelle Monae–I’m incline to suspect that they are consciously compressing her less to get that old school soul vibe which she has along with the future girl schtick. I like her by the way

  • Mark Bower wrote:

    Compression is applied not only at the recording stage to replace the fader-riding engineer, it’s also applied at the mastering stage to make an overall “hotter” recording.

    There are some differences between analog and digital. Tape apparently has some “natural” compression that is not quite as pleasing as what happens when recording digitally. I don’t quite understand this part of it, but apparently some still record to tape initially and then go digital in the later stages of the process to get the benefit of a “warm” analog sound.

    Economics plays a significant role when it comes to the difference in the amount of compression in audio formats. There is a trade-off between sound quality and file size. Make an MP3 from a CD or a .wav file and listen to both. As faster downloads and larger, more efficient storage become cheaper there may be some decompression.

  • Meredith wrote:

    I think there’s another key point here: that, since the early 1980s, pop music is visual.

    In your last post on compression, you mentioned Lady Gaga’s music, which replicates that flat sound you would hear in a dance club. (I think that’s part of it too, at least with dance music: simulating for kids in suburbia the experience of being in a big-city dance club.)

    Lady Gaga’s music is undeniably visual, ie that it is designed to support an elaborate spectacle. I think this is especially true of the song “Bad Romance.” The first time I heard it, I felt like I was watching it, not listening to it. I realized that all of the breaks & bridges in the music were probably keyed in the live show to dramatic lighting cues, costume changes, and dance moves. By listening to the song, I could picture the live show, even though I have never even seen the woman perform, not even on television.

    I don’t understand the mechanics of this stuff, but I do think the argument needs to take into account the visual reorientation of pop music. It used to be that videos were simply visual interpretations of a song, and live shows were simply a chance to hear a song live. What if the process has been reversed, that the visual orientation of our culture has caused pop music to be written and produced in such a way as to set up and sustain the inevitable spectacle that emerges from it? There’s still an economic component to this, in that new artists get really shitty contracts to have their music distributed, and the live show becomes the only way to make money. I’m sure Lady Gaga loves her fans, but I suspect she also is trying to make bank while she can, and touring is far more lucrative than putting out new music. Plus she’s all spectacle–the “music” was always designed to deliver the visual, not the other way around. This is true of videos too. A video is essentially an advertisement for a song, but I think with the proliferation of MTV and later the ubiquity of hand-help screens, the consumer’s expectations for their music have changed. Videos are also landscapes for product placement and corporate sponsorship. And they are products themselves–allegedly some wealthy businessman paid Lady Gaga $1million to appear in the “Allejandro” video.

    Anyway, thanks for giving me something interesting to think about!

  • Tom King wrote:

    I think part of what’s happening is that technology is so much a part of the modern pop music songwriting process that production and songwriting are blurred. Those special effects are the song just like the amp and distortion pedal are a huge part of the sound of an electric guitar. Those things are not ancillary. What would the Edge have been without a delay pedal? He’d be washing dishes at a pub in Dublin.

    In the old days music was mostly acoustic, so with advances in recording technology there was an obsession with obtaining fidelity to the original live sound. We used to call the stereo a “hi-fi”. Today, however, music is heavily produced with signal processing tools so to what is the fidelity? Today the production is as much a part of a band’s sound as the dry recorded instruments and bands that are purely electronic have no acoustic presence to match to. What does fidelity mean in that context? Instead of the live sound being the ideal, the recorded versions of the songs are what musicians now strive for in the live environment. More and more studio gear, like Autotune, has made it into the live performance environment for realtime use. Is this good or bad? I don’t know; it’s just the current aesthetic. And yes, I think it self perpetuating. Musicians imitate the songwriting and production of the bands they like.

    I would still think that modern acoustic music like classical, some jazz, and bluegrass would still need to use the old fidelity model.

    I always wondering if production has evolved intentionally to accommodate certain listening environments like radio, earbuds, and clubs. I think this has always been going on to some degree. Look at the Selmer guitars! That’s an example of instrument that was modified to be loud so it could match its environment.

  • Sear sees it as a process labor historians call “deskilling,” in which the marketplace replaces skilled workers with machines, and he’s like a cobbler, and so are skilled musicians: somebody with less skills,or no skills at all, takes your job. But CDs don’t get any cheaper.

    But I think the observations here are right, that a lot of different factors combine to produce a style that’s then self-perpetuating

  • […] would be worth com­par­ing this kind of com­pres­sion to com­pres­sion in mod­ern music,: in both cases the dynam­ics are lost but the mes­sage becomes more “punchy.” […]

  • The analog vs. digital recording issue is actually related to distortion as much as compression. Compression is of course a form of distortion, and certain types of what most people think of as “distortion” (e.g., overloading one or more of the stages of a guitar amplifier) are in fact used as much for their compression characteristics (“better sustain”) as their other distortion characteristics.

    Analog tape in particular distorts and compresses a signal in certain, often pleasing, ways that digital recording doesn’t, particularly when “pushed” a bit (overloaded slightly). Tube amplifiers are another famous example; they differ from transistor amplifiers not only in subtle ways, but in relatively obvious ways such as the harmonic content of the distortion.

    As a former rock and pop recording engineer (both studio and live), I can’t agree with the idea that compression was ever used for de-skilling the recording process, or even that compression and fader riding are the same thing. I certainly used both in my time. (And no, a score isn’t necessary when you already know the music–you’ll note that most rock musicians don’t work from a score, either.) Compressers still take some skill to set up; it’s not like you just flip a switch or even turn a single knob. (For even the simplest musically useful compressors you need to set both a threshold and a ratio; there are many more settings on more sophisticated models.

    One key thing that compression can do that can’t be done manually is change the gain on the signal much faster than can be done manually. And in that comes what might be the key insight: in electrical and electronic recording, life is all about gain: managing your gain and signal levels is a constant and continuous concern. While there are technical reasons for this (no recording system has anywhere near the dynamic range of a human ear), it was inevitable that in a world dominated by amplifiers very quickly people would start to consider the artistic effects that could be achieved through varying gain in various ways and at various frequencies. In some instances gain control will be used indirectly in multiple stages in fairly sophisticated ways, e.g., using a compressor to control the level of the signal used in turn to control a gate on the reverb signal of a snare drum, which reverb signal is then mixed back in with the original drum’s signal.

    Dealing with dynamic range is not even something limited to audio: it’s been a constant concern of photography since its inception (eventually leading to interesting stuff such as HDR imaging), and there have been similar discussions about the artistic merits of those sorts of techniques.

    Compression has certainly been used for purposes that have dubious artistic value, but to a recording engineer it’s just another tool for manipulating sound, and is no more artificial or strange than the whole idea of recording and playing back music in the first place. And while there are areas of the recording world where you’re trying to achieve close fidelity to what you’d hear in an entirely acoustic original performance, much modern music never existed or could exist in such an environment in the first place, and talking about it as if it ever did is as strange as wondering why a rock guitar doesn’t sound like a classical guitar.

  • As to why there’s such a broad taste for heavily compressed music now, I think it’s a matter of imitation. The biggest names in pop/rock used it heavily & were successful, so the following generations are doing it. The people making pop music now are primarily the ones who listened to 80’s pop, no? And 80’s–let’s face it–are compressed like no other, in imitation of 60’s. Like fashion, music makers are imitating what preceded them the generation before, around 15-20 years mas o menos.
    So you can blame Michael & Madonna & Depeche Mode.

  • And then before that, Phil Spector. 🙂

  • Ну как прошли праздники ?
    Кстати со старым новым годом 🙂 Будет повод еще раз праздновать праздник и посмотреть фильмы 😀

  • A lot of specialists tell that personal loans aid a lot of people to live the way they want, just because they are able to feel free to buy necessary stuff. Moreover, various banks give auto loan for different persons.

Leave a Reply

Your email is never shared.Required fields are marked *