In last week’s Carl Newman thread, our passionate and highly knowledgeable friend Homefrontradio, the Thunder Down Under, wrote about his disgust with modern-day dependence on compression. Among other things he wrote:
Basically, as you play the song loud, there’s a random noise field being generated on top of the song that’s a is physically-disturbing to how our ears are designed to hear music.
Maybe that’s not his most representative comment from this thread, but it struck a chord with me. It reminded me of our ongoing examination of Lou Reed…as his music was meant to sound! Seriously, I highly recommend going back and checking out what Townsman Homefrontradio has to say on these issues of compression. It’s great stuff that musicians and fans of new and old music can get into. And learn from. And trust me, Homefrontradio’s displayed a gear geek side, but he’s got a sense of humor about his quest and he’s what we call Good People!
As a man possibly unfamiliar with our rock legal system, we chatted a bit off list about the implications of this RTH SUMMONS. It was important that he understood that this was an opportunity to better explain himself, teach us, and most likely learn something himself. He wrote of an old recording project he’d been revisiting with a friend:
A mate and I were recently passing back and forth pictures of song waveforms with our comments, (I’ll see if i can find them), and learning from the process what makes songs work and not work, and why some remasters sound worse than original cd releases.
Our experimentations with remastering our own songs lead to a lot of confused emails along the lines of “Vocals suddenly take up so much space!” and “Wow! I played bass piano in the bridge? I’d completely forgotten!”
The end result, he remixed almost *everything* he’d done recently because he was so impressed with the sound difference not clipping makes. As for me? I’m glad i learnt this at the *start* of the project i’m embarking on.
I’m sure I’m not alone in wanting to learn more about his experiences and how they might be applicable to my own home recording efforts. I’m sure Homefront’s not the only one in the Halls of Rock who’s wrestled with compression, waveforms, and the almost insurmountable challenge of discussing this stuff in a public forum while retaining a social network outside one’s extended music nerd community.
Here’s what came to mind for me, and you may have your own questions for Homefrontradio to answer. Most likely, Homefront, will have questions of his own.
Interesting stuff, Homefront. I wish I better understood what I should be liking based on the natural construction of my ears. As an essentially low-fi, ’60s pop-based fan and musician, I wonder what it is my ears desire. I’m always skeptical of the stuff I read from obviously knowledgeable and passionate guys like yourself regarding compression and loudness. I like hearing an old vinyl release of “Satisfaction” as much as anything. I believe, in technical terms, that record may be considered to sound like shit. But I like it. Loud and in no way remastered!
Am I missing something? Am I cheating my ears of what they truly need to hear? What’s an example of an album that meets your expectations regarding these issues that I might like among some of my favorite artists (eg, Beatles, XTC, Elvis Costello, Pere Ubu, Marvin Gaye, The Stooges, Crowded House’s Together Alone album, etc)? Similarly, is there an album I might love that offends your research in this area? What does an examination of the soundwaves on a favorite Lou Reed album, say, tell you? I’d really like to know so that I could better assess what it is I may or may not like. Thanks!
It’s late at night here, so i’ll have to address this in the morning, but I’m already analysing some of the CD’s you’ve mentioned there and will write them up. (There’s a couple I already have suspicions about based upon my gut reaction to them).
Just briefly , as for ‘Satisfaction’ – of course it will sound great! Vinyl has a much smaller dynamic range than CD, (i think it’s something like 50dbs vs 90dbs), and is a better natural fit for the human ear, because it can’t reproduce the frequencies that can physically fatigue the ear through the act of listening. CD can and does, especially if it’s handled incorrectly.
The other main difference is between how analog and digital react by going ‘into the red’. Analog just ends up sounding warmer, and you can create some beautiful effects that way. Whereas Digital clips and destroys the impact of the sound.
I’m stating this because it’s entirely possible that if i got my hands on a vinyl release of Carl Newman’s work that I would respond more positively to the music due to the difference in mastering for CD and Vinyl.
More tomorrow.
Yikes. Longish post here.
Well, I’m pleased to see this foray into sound geekery getting support here on RTH. I am one of those Townsfolk who have some interest in production, with just enough knowledge to be both opinionated and erroneous. I would caution, though that there is lots of room for those debates that arise from a lack of clarity in defining terms. For example, Homefront, I’m confused by what you mean about “frequencies that fatigue the ear.” An lp with a range of, say, 20 Hz to 20kHz has plenty of room for all the frequencies the ear is able to hear (especially in us oldsters,) presumably including the most “fatiguing” ones.
The one phenomenon people might think they don’t understand is pretty simple. On a classical recording, if you’re hearing the whole orchestra going full tilt, and you’re listening at a moderate level, and then the instruments stop and you have a single violin plucking a string very softly, it is very difficult to hear that single pluck. The dynamic range, from loud to soft is huge, and is giving you an immense contrast. This kind of usage of extremes of acoustics and so forth is a regular feature of the classical genre. But if you were recording a rock song that went from full tilt to a whisper, you would generally turn the whisper up a lot, so that it wouldn’t be lost in the contrast with the din. The sound waves that look like bricks have essentially no quiet parts. The quietest sigh to the loudest guitars take place within a tiny dynamic range. It is very unlike real life in this way, and is one of the things novice bands get frustrated with when they begin recording, and wonder why they don’t sound like their favorite records.
One thing that happens with recordings I have made is that we often end up unable to get the overall sounds loud enough. They’re always quieter when compared to most commercial recordings. Ultimately it turns out that I was always trying to keep a more natural dynamic range, so that to much of the music was taking place further away from top volume. I went back and compressed the low and low middle frequencies so that they took up less space, and then the whole thing could be turned up farther.
Some good examples of hearable compression are:
Listen to the high hats and the cymbals in general on “I Am the Walrus.” When the cymbal is hit it makes a rasping his that doesn’t really change level, right through until the next hit. It’s a constant sound. Nothing like what happens in the room when you hit a cymbal. The natural decay is jammed right up in volume.
Listen to “Life Is White” by Big Star. The guitars, the harmonica, the whole damn song. If you have it in a mix of various music, it will suddenly jump out like “Crap! These guys aren’t messing around! This shit ROCKs.” (This happens to be true, but it is driven home by the recording and production.) Also, there is a lot of pop, especially things like J-Pop. – Puffy Amiyumi, things like that. And honestly, I don’t even know how they do this. It sounds like, when you listen loud or on headphones, like the guitar sounds, in extreme stereo, are, like pulling the guitar sounds out of the middle of your head! Very, very tiresome sounds to hear, but fascinating as a novelty. I feel like an old geezer working on my old Impala, and suddenly staring at one of these newfangled cars with all kinds of computery jiggery.
This issue first came to my attention with the 80s genre of Metal. I grew up thinking of Sabbath and UFO and so forth as metal. So me and my circle called the 80s metal bands “Alloy.” It was metal, but not heavy. In retrospect, I think some of this was because the ear couldn’t apprehend the heavier bits if there was no quieter bits for contrast to give you a sense of scale.
The earlier uses of compression were using it in such a way that the compression itself wouldn’t be audible. Nelson Riddle’s Sinatra records jump out of the speakers in a semi-natural way due to a large amount of compression, but applied in a fairy subtle way.
Lastly, the phenomenon of clipping and such relates to this: when you hear a sound happening like when your car goes over a rumble strip, or over an open-grate bridge, your tires are making individual sounds which happen fast enough to sound like notes. If you drive over 220 bumps in a second, for example, that is the equivalent of a low “A” and the pitch rises as you pick up speed. So if your song does a digital “clip” 220 times every second, you will hear a note “A” that isn’t a recorded note, but is the result of a sound event happening so frequently as to sound like a note. If this is happening and you get a note in there that’s an “A” and your song is in “B flat” you can see where strange and unpleasant things could happen. A total disaster could be created entirely in post-production after the band has gone home.
Extra dweeb note: This is one way bands used to get that “Rockit” robotic vocal sound, where you set a delay to a certain very fast rate, that is imperceptible as a single repetition, but when the regenerations are increased, it creates a pitch of its own. So you could talk or sing, but every sound you make has this resonant metallic sounding note with it.
Slokie, your post reminds me of a story I read about some music/art dude who almost got a local jurisdiction to install rumble strips according to a precise installation method/layou, such that — as long as you were driving within five mph of the speed limit — the strips would beat out a tuneful tattoo on your tires. In the end, the strips were never installed, but I always thought that was a cool idea.
I’m going to ask our Senior Engineer what his thoughts on this stuff are. Doubtless his answer will be incomprehensible to me. All I know is that the always-loud, all-the-time CDs are irritating after a while — and that I hate and resent the huge volume leaps that occur when an old-school CD mastering job butts up against one of these newfangldy songs you describe, when the iPod is on “shuffle.” (Interesting you should mention J-pop as a specific example; I was shufflin’ the other day, and suddenly a Puffy song BLASTED out of the speakers. It was *so* over-the-top that I stopped what I was doing to marvel at how different it sounded.
Let the sound geekery take flight! Nice opening volley, General.
There were people who thought that the electric guitar was unnatural and that our ears were not designed to hear that kind of sound. There were people who thought that analog distortion was an abomination and that tube compression robbed music of its dynamics. I’m just wondering what hard evidence exists that the digital sound characteristics being criticized here are actually inherently bad, and whether these might be just sounds that some people (i.e., geezers like us) are not yet used to.
Good questions, BigSteve. I’m also wondering if these various waveforms might be better suited to one type of music over another. In other words, is there music that wouldn’t benefit from the sort of dynamics that HFR seems to argue for as a standard?
Granted, I’ve got A LOT to learn about the baseline issues before being able to ask anything more than possibly idiotic questions. I look forward to more informed discussion, even while some of it flies over my head!
Well BigSteve, i suppose it’s possible you could consider it a stylistic choice, but if you take into account how the human ear works it seems counter-productive to create something that can actively drive a listener away.
Ears respond to dynamics. The same tones and frequencies blasted continually cause listening fatigue – our ears grow ‘bored’, and the higher the frequency, the faster is becomes boring and tiring. Clipping mainly creates odd harmonic reponses in the high end.
Ever heard an oscilloscope? How long can you bear it set to a high whine? You wouldn’t listen to it for four minutes for recreation, let alone 70 minutes of a CD.
The other problem with hearing constant frequencies for long stretches of time is the greater likelihood of hearing loss, which is the reason why newer Ipods have a volume cap level.
I honestly think digital dynamics mainly effects our *emotional response* to music. We probably can’t put a finger on why, but albums will seem longer to us, and songs can come and go without holding our attention or being for immediate replay.
Music with a dynamic build / release structure will keep our ears interested. It allows songs to build throughout their length and then peak near the end. If a song starts out near the peak of loudness, the band has nowhere to go.
This leads to horrible songs like ‘When The War Came’ by the Decemberists, (RMS Average -8.03dbs, 307 clips in 5 minutes), that basically hums for 5 minutes.
It feels like it goes nowhere, because without an ability to build and peak, there’s no excitement generated. The end of the song sounds no different to the beginning. It’s wheel-spinning. There’s no reason to pay attention as it plays, and there’s definitely no reason to put it on again.
I think younger listeners have no idea recorded music can generate more excitement than it does, but I have a sneaky suspicion it’s lead to how disposable band fandom seems to be amongst younger listeners, who easily drop or change band loyalties even in the space between two albums.
Sorry hfr, but you’re still arguing by assertion. You have not presented any evidence that the mastering you don’t like “drives listeners away.” In fact recordings mastered that way sell millions of copies.
What you describe as the “ability to build and peak” is certainly one way to make music, but it’s not the only way. People still enjoy harpsichord music (some of us anyway) and that instrument has no dynamic range at all.
And why isn’t analog compression bad? Does it get a pass the way analog distortion does?
If there were recordings mastered the way you like that consistently outsold the recordings you think are guilty of excessive digital compression, you might have an argument. But as long as these new recordings are selling, your case is undermined by the facts.
Maybe the omnipresence of music in today’s world requires different ears than music that was made for a time when listening to music was less common. Just an idea.
I have no clue about most of this stuff, but I for one am very pleased by the fact that one of the screenshots shows one of Andrew Bird’s songs. He’s friggen awesome.
To draw a non-technical point that might help illustrate why modern “loudness”-dominated mastering is a bad thing:
I have no problem with fast food, or salty snacks, or Oreo cookies. In fact, I really like them. *At the right time*. Give me a taste test — i.e., one bite of each — between a McDonald’s cheeseburger and, I dunno, a veggie sandwich on nice whole-wheat bread or something, and, assuming the Mickie D’s cheeseburger doesn’t win (which it might well do if I’m hungry enough), it’d at least be a tie. *At the right time*.
But we all know that as tasty and desirable as a McDonald’s cheeseburger may be, at the right time, more than one or two of them will *always* give you indigestion. It takes less of an over-saturated product — as momentarily yummy as it is — to make you sick and tired.
So, dig: are Mickie D cheeseburgers *bad* or *wrong* or undesirable in some way? Well, as I’ve tried to illustrate: yes and no. My feeling is that modern mastering is a “fast food” approach — designed to get people to say: “yes, PLEASE, may I have another?” — not knowing that another sampling (and another, and another) at the same level of sonic richness is *not* what they need — or even necessarily want.
In our studio, we record a variety of things. For some things (custom music for TV commercials in particular — 30 seconds of IN YOUR FACE music), this “slammed” approach is a very good thing. You want the listener to say “I want MORE!” But we purposely eschew the use of too much of this technology when folks come in to do the album thing. Over that length of time, and for the album-length listening experience, it’s just not as pleasing as something with loudness nuance.
Which is not to say, by the way, that we consider any amount of slamming a bad thing. We frequently toss a bit of it in the mix, if something seems to be just sort of sitting there. That’s arbitrary and somewhat mysterious (i.e., there’s not much real logic behind it), but it sometimes helps.
BigSteve, analog compression gets a pass because it’s analog. It doesn’t clip. You can use digital compression, just leave enough headroom that it doesn’t clip when played at a fun *loud* volume.
There’s a huge step up in volume around the grunge period, 92 or so. The step up to truly obnoxious mastering loudness that results in constant clipping, seems to happen around 1999 or so, co-incidentally around the time of Napster.
The music industry blames illegal downloading for the loss of sales – i have more than a suspicion this is also involved.
I’ll take a random example – i’ve often wondered why Crowded House’s ‘Together Alone’ album was compared so poorly to ‘Woodface’.
‘Woodface’ has some amazing highs, but the 2nd side of the record has a bunch of ‘filler’ tracks that even now i can’t remember anything about from reading the titles, and i listened to it a lot. (Remember, the first version of the album was rejected by the record company as substandard and a few of those tracks remain).
‘Together Alone’ is just one incredibly strong track after another, (minus ‘Skin Feeling’). Most criticisms came down to production, which was odd, because i found it no sonically denser than ‘Siamese Dream’, which was also out at the time.
The difference? No vinyl release for ‘Together Alone’.
‘Woodface’ sits at a nice -15.17 dbs RMS, a level i recognise as usually providing great vocal prescence. A random song, ‘It’s Only Natural’ uses 64.9 dbs of dynamic range.
‘Together Alone’ sits at -10.48 dbs RMS, which is around a level where you’ll start getting clipping at a normal volume. ‘In My Command’ uses 60.5 dbs of dynamic range. Now, i know you’re thinking “it’s only 4.4 dbs!”, but keep in mind that a difference of 6 dbs between two sounds means one is double the volume of the other.
(I’ve seen some belief that it’s 10 dbs to ‘double the volume’, but have seen the 6dbs quoted more often, so will err on that side).
If people it turn up to a level they might have played the previous album at, can you see how criticism of the production is likely, due to the clipping: weak drums, vanishing lead vocals and reduced instrumental clarity?
But wait, following my logic, shouldn’t I hate it? I was behind the times – I had it on Cassette. No clipping.
Out of interest, i looked at ‘Don’t Stop Now’ from their new album again. -7.86 dbs RMS, a dynamic range of, (and i still can’t believe this), 50.1 dbs.
Congratulations, your CD now uses the same range as vinyl, with none of the associated warmth, and the inability to create a sample depth as realistic as a record!
This is technological progression?
haven’t we talked about this before?
i think it’s important to distinguish compression from its (ab)use in the way today’s records are mastered.
alot of the discussion here doesn’t do that, or seems predicated on the notion that they are the same. not so.
also missing from teh discussion, unless i missed it, is that some time in the spring of ’02, the fcc raised the volume ceiling so that cd’s could be a few db louder in general. so now, some cd’s peak at a lower volume than newer cd’s. that DOES NOT NECESSARILY MEAN THAT THEY HAVE BEEN OVERLY COMPRESSED.
however, many (NOT ALL) recently manufactured cd’s, feature an abuse of compression.
the people who are taking old material and remixing it AND remastering it have SEVERAL opportunities to use new digitized forms of compression that are much more detail oriented than what they had available to them the first time around.
my bandmate jeff and i heard a new mix / mastering of fleetwood mac’s “second hand dude” (or whateverthefuckit’scalled) in the car on the way home from the d.c. gig (thanks for comin’ hvb!), and it was so radically different from what our ears were used to hearing that we thought it was a re-recorded version of it. in fact, it was the same performances we had always heard, but it sounded as bouncy as “like a virgin”. it sounded ridiculous.
and of course, people recording new music from scratch are actually using it as a part of their aesthetic.
but that’s a huge can of worms.
people have been using and abusing compression since its invention.
sometimes it’s hard to tell which is which in the heat of a marathon recording session.
it’s a complicated issue…and i could go on and on and on…but my head is exploding with what to say next…so i guess i’d rather not.
this discussion is already going pretty nicely. i hope i’ve added to the discussion.
good luck with the rest of this, folks.
HFR wrote:
I’m thrilled that you brought up these two albums. Together Alone is one of my favorite albums of all time; Woodface is merely the best album Crowded House had made until that point; because it didn’t have that Mitchell Froom production that I so often hate, it was a big improvement for me, but like you said, half of the songs were filler.
Without looking at any meters – as if I could understand them – Together Alone sounds great to my ears. It’s a bit murky and mysterious in all the right ways. The bass is nice and fat. The arrangements aren’t so fussed over. It was the first album of theirs where I didn’t get distracted trying to think what I liked less: the production of most Squeeze albums or the weaknesses of half the songs on any of the first three Crowded House albums.
If you ask me, the main reason Together Alone may have gotten worse reviews than Woodface is the age-old fact that it was one step further removed from the best-loved album, the debut. It had little to do with the 0s and 1s. Rock fans, critics included, often bum out when bands move too far from that “first kiss.”
So, you’ve lost me here, Homefront, except to say that YOU get really bugged by digital clipping. You made some good analogies earlier, as did General Slocum with that point about the continuous A note, but in picking apart one of the most recent albums, to date, to have joined my “All-Time Classics” list – in large part because of my love for the production – I’m highly skeptical.
The whole analog/vinyl thing is an easy sell for me. I lean toward those media, in part because of my fetish for being able to actually touch and see things at work, in part because of my nostalgic geezerism, and in part because I honestly think the hum and crackles of a needle on vinyl and the proper degree of tape hiss help make music more exciting. They add an extra sensory element, like smelling someone or getting distracted/drawn in by the luminescence of their skin as you talk to them in person rather than over the phone.
I know I’m straying from the technical side of the discussion. I really wish I knew what I preferred in this discussion. I do know that most newer (ie, post-1983) albums sound worse, especially albums made during the mid-80s to late-90s. For instance, excluding types of music I simply don’t like or can only like so much, I mark that Talking Heads’ album with “Burning Down the House” as one of the first albums by a band whose albums I’d always liked as being beyond my liking, PRIMARILY FOR ITS COLD, PULSE-FREE SOUND. I used to try listening to that album and crank it up, but nothing ever sounded louder or more exciting. Was this because of the compression? I’m telling you, Homefront, it was more than my dissatisfaction with half of the songs.
In more recent years, it seems more artists working in various styles have found a way to make their albums have more life and, I don’t know how else to explain it, bounce. I’m not just talking about Grand Masters like Nick Lowe’s last few albums, but Beulah’s The Coast Is Never Clear (I think that’s the title) and the Stephen Malkmus album I so love, Face the Truth. This last one, in particular, has a lot of the qualities I like in late-60s albums by bands like The Pretty Things. The Beulah album, I suspect, would make a waveform more like that Smashing Pumpkins recording, but somehow the album manages to sound different depending on whether I listen to it loud or soft, in the car or at home. I’m telling you, Homefront, I couldn’t get that variety out of that Talking Heads album I eventually dumped, and not just because of the slick arrangements. Some albums breathe, in my experience, some don’t. Together Alone breathes just fine for me.
Anyhow, I know this topic can be frustrating if not mind boggling to discuss, but I appreciate people’s coming forward so far, and I appreciate HFR’s efforts at expressing his thoughts on these matters. I don’t discount the value of Townspeople expressing their innermost rock thoughts, even when I disagree with them, am skeptical, or am partially baffled. I can’t tell you how many times I hear a particular song these days and the slightest trace of a point one of you made months earlier, a perspective on the song I’d never considered, creeps into my mind and enriches the listening experience.
“But don’t you understand, it’s clipping” still sounds like “but don’t you understand, it’s distorted” from an earlier era to me. Maybe you should explain exactly why clipping is bad. Actually maybe you should explain more clearly what clipping is for those who don’t record audio.
For example here’s one thing I don’t get. Hfr writes:
The way I understand it, the clipping is in the waveform itself, and wave would be clipping whether you played it back at low or high volume. No?
That’s what I was thinking, BigSteve. I know very little about recording, but if a recording has clipping, you can’t make it go away at any volume. You’re compressing what’s already clipped, right?
I have some friends that have always been vinyl junkies, and I wasn’t an early adopter of CD’s, but after the first gen of players and when the cd’s quit being mastered from fourth gen twice removed vinyl, I thought cd’s definitely sounded better than most records. Not necessarily audiophile pressings, but the average cd was usually a notch above the mass produced record made with 100% recycled vinyl. Mass market major label lp’s got even more complaints than the recent brouhaha over cd’s. They deserved it, too. But some of these friends are pulling out 80’s pressings of early 70’s albums and just because it’s vinyl, anointing the cracks and pops as the “true” sound of the music. Case in point, Steely Dan.
The Dan’s albums got those gold stamps on them in later pressings that said something like “Steely Dan Certified Gold Recordings.” Those lp’s sounded pretty sad compared to the originals. The first Dan cd’s were mastered from those crappy records (I think) and they were horrible. I didn’t get any of their stuff on cd until Citizen Steely Dan came out, and that sounds better than every vinyl release ever. I don’t think there’s even room for any other opinion.
I think I get more fatigue from pristine recordings that never distort or super clean guitars than I do from the lo fi noise I listen to. I’ve only got one New Pornographers album, Electric Version and it sounds really good to me. I haven’t looked at it on a waveform, but I try hard not to let other people or things get in my way of liking something or not. I do find the whole thing pretty fascinating, though.