Cites & Insights: Crawford at Large
ISSN 1534-0937
Libraries · Policy · Technology · Media


Selection from Cites & Insights 10, Number 2: February 2010


Perspective

Music, Silence & Metrics

This story—part essay, part casual research—begins with the editorial opening the November 2009 Stereophile. No, it’s not a migration of my My Back Pages grumbling about high-end audio prices (or reviewers who mistake musicality for accuracy) to the main body of Cites & Insights—indeed, it’s an example of why I still subscribe to Stereophile. And there is, believe it or not, a library connection—or at least there should be.

First of several digressions: Core portions of this essay involve a website and movement of sorts called “pleasurizemusic.” I should stress that “pleasurize” is an unfortunate, somewhat Germanic neologism based on “pleasure” and has nothing to do with Nicolai Ivanovich Lobachevsky (Никола́й Ива́нович Лобаче́вский), or Tom Lehrer, or plagiarism. Well, maybe that’s not entirely true—there is a slight Lehrer connection and I don’t know enough about Lobachevskian geometry (or hyperbolic geometry) to be certain there’s no connection at all.

Background: Spaces between the Notes

John Atkinson’s November 2009 editorial (or “As We See It” column) is entitled “The Spaces between the Notes.” It begins when the magazine’s music editor handed Atkinson a Delbert McClinton CD and noted that it was musically good enough to be the issue’s “Recording of the month” but the sound was “a bit funky.” When Atkinson listened to it, he found it so in-your-face and fatiguing that he couldn’t listen all the way through.

I opened the first track with music-editing program Bias Peak Pro 6. As I suspected, the waveform was continually banging its head against the CD’s maximum level—zooming in on vocal sections revealed the squared-off shape of a signal whose peaks have been chopped off, either by being clipped or by being hard-limited.

Unfortunately, says Atkinson, this wasn’t unique.

Heavy-handed compression and even plain old distortion have become ubiquitous in what have become termed “The Loudness Wars,” in which songs are dynamically squashed to the point where they sound uniformly and fatiguingly loud throughout, even when played quietly. As I’ve written many times in the past decade, when all the dynamic contrast is removed, the music is damaged. The notes—the sounds played by the musicians—are not the music, but merely the framework for the music. As Miles Davis said, the music exists in the spaces between the notes. If that is the case, it hardly seems appropriate for recording and mastering engineers to fill up those spaces, even if, in their defense, they’re forced to do so by record-company suits’ incessant demands to “Make it louder.”

Atkinson points to Turn Me Up! (www.turnmeup.org—if you use .com, you wind up at a link parking site), a nonprofit organization “campaigning to give artists back the choice to release more dynamic records.” Quoting from the site:

Today, artists generally feel they have to master their records to be as loud as everybody else’s. This certainly works for many artists. However, there are many other artists who feel their music would be better served by a more dynamic record, but who don’t feel like that option is available to them.

The organization is trying to establish objective measurements for dynamic recordings and has a “Turn Me Up! Certified” label that could go on CDs or related marketing materials. The label includes this wording:

To preserve the excitement, emotion and dynamics of the original performances this record is intentionally quieter than some. For full enjoyment simply Turn Me Up! (www.TurnMeUp.org)

(The two-minute video clip on the organization’s home page, using a Paul McCartney clip, is extremely convincing, at least in my opinion: It shows dramatically how extreme dynamic compression—making everything loud—robs music of intensity and life.)

We’ll get back to Turn Me Up!—but first, back to the editorial that inspired this Perspective. Atkinson mentions “another resource available to those who care about sound quality”—the TT Dynamic Range Meter at www.pleasurizemusic.com, “a foundation that aims to be able to label recordings with a whole-number dynamic value, giving consumers an immediate means of knowing if a record has been over-compressed or not.” Quoting Pleasurizemusic’s Friedemann Tischmeyer:

Music—as an artistic means of expression—should transmit emotions. Nowadays, this is possible only to a limited degree because dynamics—a fundamental part of expressivity—are often missing… Modern mainstream music sounds like a flatly pressed board being rammed through loudspeakers and uses the greatest possible amount of intrusiveness just as advertising does—as a means of constantly trying to get the listener’s attention. In this way, a fundamental aspect of music is lost.

The TT Dynamic Range Meter (henceforth TTDR) is available as a plugin for Bias Peak Pro 6, so Atkinson added it and analyzed a recent CD that he engineered (of Attention Screen, a small jazz combo). “I had had to squash the occasional peaks of the original 24-bit files a little in the mastering to better fit the music within the CD’s 16-bit window. Nevertheless, the meter indicated that the dynamic range was still 10-14dB, which is much wider than a typical electric rock recording.” What about the first and apparently nastiest song on that McClinton CD? 6dB—actually, 5dB in one channel, 6dB in the other. Which is awfully low, if the measured dynamic range means anything at all.

Those are the only test results in the editorial. Atkinson discusses the remastered Beatles CDs and claims by the project coordinator that they were not dynamically compressed “to better match what has, sadly, become the modern norm.” Early reviews suggest that the new CDs are more musical—sound more like live music—and there’s good reason to believe dynamic range would be part of this.

File formats aren’t the issue. No, CDs aren’t “perfect sound forever” and are probably inferior to higher-resolution recordings in some cases, if your hearing and attention are good enough—but CDs do not at all require the kind of dynamic compression being discussed here. (CDs can handle a 96dB dynamic range.) Neither, for that matter, do MP3s. Lossy compression and dynamic compression are two entirely different things. Even 128K MP3 (which even to my aging ears is deficient and tiring) can handle wide dynamic range, and today’s typical 256K MP3s are good enough for reasonably musical sound including full dynamic range.

Second digression—or maybe it’s not. The editorial includes two of the reasons I continue to regard Stereophile as a worthwhile resource, even if I treat many of the columnists and reviews as primarily sources of humor. To wit, the magazine makes a real effort to cover new developments related to musical recording—and John Atkinson, specifically, does use actual measurements as part of the magazine’s toolkits, even though (perhaps appropriately) the measurements don’t drive the review outcomes. (In almost all cases, the review’s written before the measurements and the resulting sidebar are even prepared.) I don’t think I would have heard about TTDR were it not for Stereophile.

Is There Really A Problem?

Yes, I think there is. Admittedly, I don’t listen to a lot of contemporary music, particularly not contemporary rock, so I may not be personally affected by the problem—but, as we’ll see, there are at least some indications that it’s influenced a few CDs that I own.

Should librarians care? I think so. Libraries are, among other things, in the business of collecting, organizing and preserving humanity’s creativity. If music, one of the creative arts, is consistently stored in a manner that degrades that creativity, it’s a loss to both the present and the future. And if, as I (and others) believe, dynamically compressed music is tiring, so that you’re tempted to tune it out and not listen seriously, that’s a loss on several levels. (If you only have stuff on in the background—if music is playing but you don’t really listen to it—then none of this may make much sense.)

The Turn Me Up! site has lots of links to music dynamics in the news. The place most people might start—our friend Wikipedia—isn’t bad: Look up “Loudness War.” But don’t stop there; several other links make good and informative reading. There’s a heartening tale from Bob Ludwig, who mastered the Guns ‘N Roses record Chinese Democracy: When he prepared three versions on a trial disc, one with minimal compression and two with more typical modern compression, the producers (including Axl) chose the least-compressed version: the “quietest” version, in some ways. There’s clearly some public interest in this—as when (some) Metallica fans became aware that Death Magnetic (the CD, compressed all to hell and gone and clipped in the process) doesn’t sound nearly as good as the same tracks in “Guitar Hero” downloads—which aren’t quite as overcompressed. (In this case the band got the CD it apparently wanted.)

There are two slightly different issues here. One is the overall dynamic range of a piece, from softest to loudest—and to a great extent that’s an issue of music type and performer preference. Orchestral music may have the widest dynamic range (although a solo piano has an astonishingly wide dynamic range); a folk singer might seem to have a relatively narrow range; while electronica or metal rock might deliberately have an extremely narrow range, loud all the time.

The other is inner dynamics—the extent to which there’s breathing space within the loudest passages. The software tool I’ll discuss later on, TTDR, apparently only deals with the 20% of a piece that’s loudest overall—then looks at the difference between peak loudness and average loudness within that 20%, to calculate a dynamic range. After all, most instruments aren’t making sound 100% of the time and different instruments should have different dynamics.

Is there a textual analogy? I’m not sure, but here’s a crude attempt. You can express textual emphasis by bolding text or using italics or putting a word or two in ALL CAPS. But how would you feel about a long book that began like this:

MRANDMRSDURSLEYOFNUMBERFOURPRIVETDRIVEWEREPROUDTOSAYTHATTHEYWEREPERFECTLYNORMALTHANKYOUVERYMUCHTHEYWERETHELASTPEOPLEYOUDEXPECTTOBEINVOLVEDINANYTHINGSTRANGEORMYSTERIOUSBECAUSETHEYJUSTDIDNTHOLDWITHSUCHNONSENSE

All the words are there, and it’s readable, but it’s all VERY LOUD and has no interruptions. I don’t know about you, but I’d find that awfully tiring.

Metrics 1

Paul Lamere’s opening paragraph in “The Loudness War Analyzed,” a March 23, 2009 post at Music Machinery (musicmachinery.com), states the problem in a different way:

Recorded music doesn’t sound as good as it used to. Recordings sound muddy, clipped and lack punch. This is due to the ‘loudness war’ that has been taking place in recording studios. To make a track stand out from the rest of the pack, recording engineers have been turning up the volume on recorded music. Louder tracks grab the listener’s attention, and in this crowded music market, attention is important. And thus the loudness war—engineers must turn up the volume on their tracks lest the track sound wimpy when compared to all of the other loud tracks. However, there’s a downside to all this volume. Our music is compressed. The louds are louds and the softs are loud, with little difference. The result is that our music seems strained, there is little emotional range, and listening to loud all the time becomes tedious and tiring.

The post includes some of Lamere’s own experiments. For example, when he plotted Dave Brubeck’s Take Five, he saw a dynamic range of about 18dB—but a track from the new Metallica album had a range of only 3dB, which essentially means ALLTHEMUSICISLOUDALLTHETIME. Impressive for a few seconds; tiresome after a few minutes.

Ah, but Brubeck’s song is a jazz quartet, not heavy metal—so what of rock? Lamere measures Stairway to Heaven by Led Zeppelin: 35dB. But then there’s supermassive black hole by Muse: 4dB. I’ve never head of Muse; Lamere says he likes the group but finds their tracks to “get boring quickly.”

Lamere analyzed “about 15K tracks from the top 1,000 or so most popular artists,” using his own methodology for determining dynamic range. He gives the results in –dB “loudness” numbers; I’ll assume these relate to the average (RMS) level below digital 0 dB, since I’m not sure what else to make of them.

The average across all 15K tracks is about -9.5dB—but a bunch of artists are much louder, -5dB or less. Worst is Venetian Snares (whoever that is) at -1.25dB, essentially unvarying loudness throughout. Oddly enough, Avirl Lavigne (-3.03) is louder than Metallica (-4.14).

Then there are the “quietest artists”—which presumably means those using very little dynamic compression, not those whose engineers just master the CDs at low levels. Heading that list are Brian Eno (-17.52), Leonard Cohen (-16.24) and Norah Jones (-15.75), and the list of those below -13 includes Neil Young, Cat Stevens, the Velvet Underground, Simon & Garfunkel, James Taylor and Paul Simon—but also Pink Floyd, Phish, Phil Collins, the Police and the Grateful Dead.

Lamere’s post also plots loudness against the year of release of a recording—and there’s a fairly clear (if not entirely consistent) track upward, with relatively few “loud” recordings prior to 1990 and relatively few “quiet” ones since 1995.

A caution here: I’m not sure you can directly compare Lamere’s results—fascinating as they are—with the dynamic range results I discuss later. The methodologies differ, although the underlying idea is the same.

Pleasurize

Now we move to the Pleasurize Music Foundation at www.pleasurizemusic.com. Here’s a portion of “Our Aim” from the site, as updated July 18, 2009:

Our aim is to improve the sound quality of music in its various recorded formats—including data compression methods such as MP3—as well as music destined for radio broadcast.

Only music that provides a positive musical listening experience has real market value. The Foundation’s aim is to increase the value of music within the creative production process for the entire music industry.

The objective is to revive the willingness to pay for music and therefore to create a healthier basis for all creative participants within the music industry.

That’s the start of a long page I can’t really summarize. Once again, the aim is to provide a label for recordings indicating that they provide “natural and dynamic sound.” Is such a process plausible or even desirable? I have no idea. Does it make sense to understand what’s going on? I believe so.

If you explore the site—and I encourage you to do so, along with the Turn Me Up! site, if you care about music quality at all—don’t be too surprised if the language sometimes seems a little strained. Pleasurize may be a California foundation but it began in Germany, and much of the site has the feel of English as a second language. By the way, the home page points you to a video—and it’s the same video used by Turn Me Up!

TTDR, the tool provided by Pleasurize (only if you join at $30/year, but free downloads of an older version can be found) measures what the site calls the “inner dynamics of a recording” in whole numbers, as a Dynamic Range or DR. The DR is based on the cumulative difference between peak and RMS loudness, which is already more than I fully understand, and is not the same as “macro dynamic,” the difference between (say) a quiet introduction to a song and the loudest music within the song.

The official short-term goal of the foundation, according to the site, is for recording companies who’ve joined to release albums with a minimum of DR14 beginning July 1, 2010—either by lessening their use of dynamic compression or by lowering average levels for those CDs where there really isn’t 14dB of dynamic range. Thus, most CDs from such labels would be equally loud (or quiet) overall, but would have more natural dynamics.

This doesn’t affect apparent loudness on radios, because most radio stations add their own dynamic compression—in most cases, it’s all going to sound loud regardless. (The website asserts that music that’s already been compressed is more likely to be distorted by radio station compression.)

There’s a chart breaking down minimum suggested DR values for various types of music, calling the ranges “unpleasant,” “transition” and “dynamic and pleasant” respectively:

For largely-synthetic and sample-based music, DR6-7 is transitional and anything DR8 or higher is dynamic and pleasant for this type of music.

For most pop, rock, blues, mainstream “radio music,” the transitional range is DR8-11, with anything DR12 or higher being dynamic and pleasant.

For primarily acoustic music—jazz, folk, country, classical—you need at least DR10 for transitional and DR12 for proper dynamics.

Elsewhere on the site, you get the sense that DR14 is really the level at which “proper dynamics” are at play for most music; I note that, on the colored chart, the acoustic music column has a yellow tint (“transitional”) to the green (“dynamic”) at DR12 and DR13.

I have no idea whether Pleasurize Music Foundation has signed up many significant producers and labels—there are lists on the site (including 71 “record company” signatures and more than 1,200 supposed recording engineers), but they’re not all that meaningful. I suppose we’ll find out more this year. I don’t buy all that many CDs recently (or downloads—so far, none of those), so the fact that I’ve never seen a DR label in the wild isn’t necessarily meaningful. I’m neither touting for the organization nor convinced either that it’s a plausible solution or that it’s hopeless.

It did inspire me to do some of my own tests—recognizing that most of the music I own predates 1995, the point at which heavy dynamic compression seemed to become the norm.

Metrics 2

I located and downloaded a free copy of an earlier version of the offline TTDR (version 1.1)—and was delighted to see that, although it says it only handles .WAV files (regular audio in CD-digital form), it actually handles vanilla MP3 as well, if a bit more slowly. Delighted, because essentially all the non-classical music in my collection (and some of the semiclassical music, that is, acoustic guitar and some piano) is stored in 320K MP3 form. I ran a few crosschecks and satisfied myself that the DR result for a 320K MP3 (the highest quality MP3 you can record) is the same as for the original .WAV.

I looked at just under 1,400 tracks—a little over half of the nonclassical part of my collection, excluding most of the songs I know I’ll never want to mix to CD-Rs or put on a USB player. That collection is mostly from the 50s and 60s and heavy on folk and pop, with many of the same artists continuing into the 70s, 80s and some since. Quite a bit of it is collections, including Sony’s “Essential” 2-CD bargain sets (which typically contain the equivalent of four original CDs of music).

The results show range within this collection—but I have no Metallica, no Stooges, very few contemporary recordings of any sort. Given those caveats, some notes on what I found—first in terms of the actual numbers, followed by a few listening tests. It’s worth noting that I had no idea what I’d find when I did these tests.

Overall Numbers

Of 1,394 tracks measured, the DRs come out like this. Note that 5-7 would be considered overcompressed for any of the music I have, with 8-11 either overcompressed (for folk, country, jazz, classical) or transitional (for rock, pop, blues, etc.):

DR

Tracks

Percentage

Note

5

1

0.07%

Compressed

6

6

0.43%

Compressed

7

17

1.22%

Compressed

8

53

3.80%

Compressed or transitional

9

100

7.17%

Compressed or transitional

10

181

12.98%

Transitional

11

243

17.43%

Transitional

12

313

22.45%

Good

13

242

17.36%

Good

14

140

10.04%

Dynamic

15

57

4.09%

Dynamic

16

26

1.87%

Dynamic

17

6

0.43%

Dynamic

18

7

0.50%

Dynamic

19

2

0.14%

Dynamic

That’s not too bad overall. The worst case (all of the 8 and 9 being folk, jazz or country) shows 12.7% overcompressed, 30% transitional (not great, not terrible), 40% good (dynamic but not up to DR14 standards) and 17% fully dynamic, at or above DR14 standards. (Quite a few of the DR12 and DR13 tracks would meet DR14 standards because their peaks are 1dB or 2dB below digital 0.) The statistically-minded among you have probably already spotted something: The distribution is a classic bell-shaped curve.

I’m guessing, based on everything I’ve read, that a largely-contemporary collection would also be bell-shaped—but the peak of the curve would be considerably to the left. (If the measure used by Paul Lamere was directly comparable to DR, the average for the 15,000 tracks he measured was 9.54, suggesting that the peak of the curve would be centered between 9 and 10.)

Overcompressed Tracks

Let’s look at the tracks that fall into two extremes—those that are almost certainly overcompressed (DR 5-7) and those that are very dynamic (16-19). Are there identifiable characteristics?

For the worst case, DR5, the answer is clearly Yes: It’s Come Together by the Beatles—from The Beatles 1 compilation, which was almost certainly badly compressed for rerelease. Note that two more tracks from the CD have DR7 (Something and Let It Be) and five more have DR8, with three marginal at DR9 and one reasonably good at DR11 (Yesterday—which, notably, would seem to be less dynamic than most Beatles numbers). I didn’t retain most of the cuts on this CD on my PC; I’m guessing most of the rest also have low DR numbers.

DR6 and DR7 are more mixed, but in a number of cases there’s a likely pattern.

Two of the DR6 cuts and five of the DR7 cuts are from Carly Simon’s Bedroom Tapes, a recent (2000) self-produced album produced primarily by Simon herself in a home studio. (One other cut is DR8; I didn’t keep the remaining three tracks.)

One DR6 and two DR7 are by Eric Clapton in his “wall of sound” mode on Clapton Chronicles and Pilgrim (five more cuts are DR8 and three are DR9, with four DR10—none of the cuts that I kept have higher DR numbers). The worst case is Blue Eyes Blue; for a couple of the cuts—(I) Get Lost and River of Tears, both DR7—the narrow dynamic range might be an artistic choice.

The other three DR6 cuts are David Olney’s Round from The Wheel (an a capella round), Randy Newman’s Old Kentucky Home from his Guilty 4-disc collection (a remaster), and The Lovin’ Spoonful’s Did You Ever Have to Make Up Your Mind from a badly-done reissue combining two original Spoonful albums with no access to the master tapes.

Four other DR7 tracks come from recent James Taylor albums—the special Hallmark Christmas Album and the new Covers. Since 11 more tracks from these albums have DR8 and seven have DR9, with one DR10 and none (of those I kept) higher than DR10, it’s extremely likely that both albums suffered from the loudness wars—and, thinking about it, I believe Covers would be a lot more interesting if it had more dynamic range.

Otherwise, three songs from two Sony “The Essential” collections—but those collections aren’t uniformly compressed.

Very Dynamic Tracks

What about tracks with DR16 and above—noting that DR14 and DR15 also have wide dynamic range?

James Taylor may have overcompressed tracks on recent albums—but Never Die Young, recorded in 1988, includes one of the two DR19 tracks, four of the seven DR18, two of the six DR17, and two of the 26 DR16—and two more DR16 tracks are from other pre-1990 James Taylor CDs.

The other DR19 cut is George Winston’s December, solo piano on Windham Hill. The other DR18 tracks include two classical guitar pieces (from two different CDs by different artists) and Gordon Lightfoot’s Triangle—notably, from a compilation CD.

Other DR17 cuts include two solo acoustic guitar tracks (both Christopher Parkening, both classical), Cat Stevens’ If I Laugh, and—oddly—Don’t Worry, Be Happy from a cheapo collection of reggae music.

The rest of DR16? Five more George Winston solo piano pieces, eight more acoustic guitar pieces from five different CDs, two more Gordon Lightfoot songs from the same compilation CD, two by Eddy Raven from a compilation CD—and single cuts by artists as varied as Boz Scaggs (from the Hits! collection), Mannheim Steamroller, Paul Simon (Diamonds on the Soles of Her Shoes, from the very first CD we purchased, 24 years ago) and Randy Newman (Change Your Ways from the Ragtime soundtrack).

That’s the high end. While DR15 is far more varied, there’s still a lot of acoustic guitar (13 tracks), a fair amount of James Taylor (six tracks) and Gordon Lightfoot (five), along with four each from Elton John (two from collections, two not) and Paul Simon (from three CDs), three from Garth Brooks (all from the same CD) and Tom Paxton, two each from Randy Newman, Ry Cooder and Vance Gilbert—and singletons from ten others.

I’ve only touched on the fringes, with 91% of the music in that middle range from DR8 through DR14. Do the very dynamic cuts sound better than the apparently compressed cuts from the same artists? By and large, yes. (Would the differences be more dramatic if I included all the CD tracks, including ones I never expect to listen to again? I’m almost certain they would.)

A Compression Timeline?

It’s clear from looking at tracks within albums, in the overall spreadsheet, that not everybody gets treated the same way. As you’d expect, classical and similar recordings (including those acoustic guitars) aren’t usually compressed very much, and some smaller labels seem to avoid overcompression.

Is there a reasonably clear timeline? Just for fun, I tagged the tracks with the later of either the recording dates or the CD release dates (I couldn’t always find the latter). After doing averages and standard deviations year by year, the patterns were obvious enough that I clustered them into multiyear periods—noting that anything before 1986 almost certainly has a later CD release date. The graph here may be all that needs to be said.

Clear enough? Maybe not, so let’s summarize (noting that the first three time periods include 110-120 tracks each, the periods from 1989 through 2003 include 180 to 238 tracks each, and the 2004-08 period includes 33 tracks from three albums):

Most pre-CD recordings have decent dynamic range—on average, 12.8-13.1 (with standard deviation 1.3-1.7).

Albums from the early CD era (1986-1988) show very well, with an average DR of 13.6 (standard deviation 1.8).

Then things start heading downhill—a little above or below 12 in 1989-1994, dropping sharply to 11 and below after 1995.

The recent sample is too small to mean much (three albums, two from one artist), but it’s appalling: 8.9. If it weren’t for two tracks on Randy Newman’s Harps and Angels with good dynamics, it would be even worse: the two James Taylor albums manage a paltry 8.2 average, with only one cut over 9. Still, that’s better than the astonishing 1.5 to 3dB range of some mangled rock recordings.

Time for another digression—well, let’s make it two in one. First, there’s the question of whether all this has anything to do with Tom Lehrer—and it does, only because I decided to rip the Rhino CD that includes all of Lehrer’s first two albums and four other cuts—this time as .wav files to play it safe. The compilation was released in 1997—and, other than the final four cuts, it’s just one guy at a piano with no fancy recording technology. The DR values? One 9, ten 10 and a dozen 11—but also two 12, two 13 and one 14 (the previously unreleased I Got It from Agnes). Not great results, but also not atrocious. Second, I thought it might be interesting to present the astonishing decline in recording dynamics the way most media would—by using a little chartjunk. Here’s exactly the same chart, but with one modification to make the changes more dramatic:

Wow! Look at that incredible drop! And if I accidentally left the numbers off the vertical axis… Well, as Smoky never said, “Only you can prevent chartjunk.”

The Compilation Situation

Until 1990 and perhaps up until 1993 or 1994, it’s fair to assume that most CD releases of albums that originally appeared on vinyl didn’t undergo dynamic compression, or at least didn’t undergo much of it. The same was probably true of reissued compilations.

Since 1995, compilations of older material have offered an opportunity to make the old stuff louder through compression. Some artists have resisted this (Bob Dylan has been critical of overcompression); many don’t really seem to listen to their recordings (or no longer have the ears to hear critically). Let’s look at some compilations, which make up quite a lot of my CD collection.

Sony’s The Essential series have provided excellent value when you’re trying to get a feel for an artist. Two CDs are filled with songs reasonably carefully chosen from all the albums recorded on Sony-affiliated labels (e.g., Columbia and Epic) and sometimes other labels: many of the CDs come in at more than 70 minutes with up to 20 songs per CD. Target sells these two-CD sets for as little as $11 to $15; I see them on Amazon for $14 to $16. Sony’s still churning these out, with dozens of them—more than 60—covering an astonishing range of artists and groups (including ones that I never thought had any Columbia/Epic affiliation and a fair number of classical artists) and some new “eco-friendly” 3-CD sets. I have sets for Billy Joel, Bob Dylan, Dave Brubeck, Johnny Cash and Willie Nelson (looking at Amazon, I think I should add a few more…) They are, to put it bluntly, a mixed bag sonically:

Billy Joel has a dozen or so at DR8-DR10—but also includes several at DR12-14, including some recent recordings, along with a bunch at DR11.

Bob Dylan has a couple at DR9 and a bunch at DR10, showing probable signs of compression, but also a handful at 12 and above (Don’t Think Twice, It’s All Right comes in at DR14).

Sony did well by Dave Brubeck. The worst track is DR10; most are DR12 or above, with three at DR14.

Johnny Cash doesn’t fare so well—none of the tracks (that I kept, and I kept 26 of them) reach DR12, and fully half are DR7-9, which are marginal for folk & country.

Willie Nelson’s a mixed bag, and here I kept 35 of the cuts. While 15 have DR12 or higher (including five DR13 and one, Good Times, at DR14), another 13 are sonically-compromised DR10-11…and, sigh, seven are down in the overcompressed DR8-9 area, including two that really should have better dynamics (Mendocino County Line and Slow Dancing). It does appear that several with poor dynamics were fairly recent recordings, and that may be indicative of a problem with Sony’s country labels.

What of other compilations? Here are some of them:

Boz Scaggs’ Hits! originated in 1980 and shows it: tracks range from 11 to 16, most of them 12 and above.

I didn’t keep that much of The Best of Eddy Raven, but can’t complain about the dynamics: five tracks, ranging from DR13 to DR16. (The two at DR16—Shine, Shine, Shine and I Got Mexico—are indeed vibrant recordings.)

Elton John’s four-disc 1990 collection To Be Continued… (did Elton really guess he’d still be doing this 19 years later?) does not suffer from too much compression—at least not in most cases. Of the 27 tracks I kept, only two are at DR10-11, and only seven are at the fully-acceptable DR12. The rest are mostly DR13, with half a dozen at DR14 and two at DR15.

You already know that the 1999 Clapton Chronicles is sort of a mess, with tracks ranging from 6 to 10. Since the 1998 Pilgrim has the same range, one can only assume Clapton likes it that way—and while I find some of the songs compelling, it’s true that I only want to listen to two or three at a time.

Gordon Lightfoot’s Gord’s Gold and Gord’s Gold Vol. 2 (1975 and 1988 respectively) are both just fine—the first mostly 12-14, the second entirely 13-18. On the other hand, The United Artists Collection, which is a reissue compilation from 1993 (of much earlier material), doesn’t do so well: of 42 tracks I kept, 11 have DR9-10 and another 11 are at DR11, with only seven above DR12.

Harry Nilsson’s Personal Best came out in 1994 and doesn’t seem to involve much compression—although there’s one DR10, there are a bunch of DR13-15 tracks.

James Taylor’s Greatest Hits is now 34 years old, and it shows few signs of compression: All the tracks I kept, nearly all the album, are DR13 and above. With Taylor, there’s a sharp distinction between pre-1992 albums (almost always DR12 and above and as high as DR19) and those from 1997 and beyond (mostly 9-11 in 1997 and 2002 albums, even worse in 2004 and 2008).

Judy Collins produced her own Forever in 1997. Of the tracks I kept, 20 are DR9-10, only 6 DR11-12, none higher.

Paul Simon’s mid-career Negotiations & Love Songs (1988) is what you’d expect from 1988 or maybe better: the songs I kept are all DR13-15. (That’s true of most of Simon’s CDs—until You’re the One in 2000, mostly 8-11).

Randy Newman’s gargantuan Guilty ranges all over the place, artistically and sonically. Of what I kept (nearly four dozen tracks), I count 16 tracks below DR10 (including two below DR8), but also 16 at DR13-16. (14 are in the middle, with DR11-12.)

Simon & Garfunkel’s Old Friends from 1997 was probably compressed, but not as badly as it might have been. Of 48 tracks I’ve kept (in some cases, I had the song on the original CD and it sounded better there), three are DR9 (all of them songs that should be livelier than they are), ten are DR10 and 13 are DR11—but that leaves 13 at DR12 and eight at DR13, plus one DR14 (Kathy’s Song).

The Beach Boys massive Good Vibrations collection (five CDs) came out in 1993, so you wouldn’t expect much dynamic compression—but that’s not to say these are heavily dynamic (quite a few are partial tracks). I kept 42 of some humongous number of tracks, including six DR8-9, 17 DR10-11, 15 DR12—and five DR13-15 (the DR15? Disney Girls).

OK, I know, The Carpenters made homogenized music—but while Richard Carpenter may not be the world’s greatest arranger, Karen Carpenter was a remarkable vocal talent who died far too young. Richard’s vast ego is ever-present in the four-disc From the Top, released in 1991, but it doesn’t suffer from compression. Of the 35 cuts I kept, there’s one DR10 and eight DR11—but also 14 DR12, nine DR13 and three DR14.

Tom Paxton’s I Can’t Help But Wonder… collection (1999) does pretty well, at least among the 21 tracks I kept: four DR10-11, three DR13, two DR15 and the rest DR12. That range is typical of most of his albums, although one (It Ain’t Easy from 1991) stays down in the DR9-11 range.

Technically, Will the Circle Be Unbroken “by” the Nitty Gritty Dirt Band (but mostly by a group of country’s pioneer musicians) isn’t a compilation—but the 2001 CD reissue doesn’t sound nearly as lifelike as the stunning multi-LP originals, and it’s fairly clear why: five DR8-9, eight DR10, three DR11 and a paltry two DR12 (nothing higher), of the small selection I chose to keep. (One of the two DR12 is a soft acoustic guitar solo that’s also the only non-country song: Both Sides Now.) Too bad; the original was cleanly and simply recorded and almost certainly had more life.

Conclusions

I could discuss individual discs by various artists, but this essay’s more than long enough already. I should mention that Carly Simon doesn’t always cover a narrow range (for example, No Secrets ranges from DR12 to DR14, the latter including You’re So Vain), that soundtrack albums vary all over the place (the 1998 James and the Giant Peach—there’s Randy Newman again—has excellent dynamics but the 2000 O Brother, Where Art Thou, full of first-rate music, is mostly DR9-11), that some country artists got great recordings at least into the early 1990s (Randy Travis is mostly DR12-14, with a few exceptions; Ronnie Milsap consistently 12-15).

Can you hear the difference? I suspect most people who listen to music at least some of the time can—but unless you have uncompressed and overcompressed versions of the same music for comparison, it’s hard to be sure.

I do suspect that, if you have music by artists whose songs you love but you find that you get bored with the music after two or three songs, you may be dealing with dynamic compression. Everything is “loud” (unless you turn it down), but at the expense of natural musical dynamics. If the compression has been done during the mixing stage, there’s not much hope of redemption: As Tom Paxton sings, when it’s gone, it’s gone. But if the big-label music industry could get past the LOUDER IS BETTER obsession or enough artists object, there are probably thousands of recordings out there that could sound better, by restoring the dynamic range in the master tapes (or, these days, the master files). Although I haven’t heard them yet, there’s reason to believe that’s happened with The Beatles; I’d love to see it happen with hundreds of other artists. For the sake of today’s and yesterday’s music that deserves to live on, let’s hope we see progress.

Masthead

Cites & Insights: Crawford at Large, Volume 10, Number 2, Whole Issue 125, ISSN 1534-0937, a journal of libraries, policy, technology and media, is written and produced by Walt Crawford, Editorial Director of the Library Leadership Network.

Opinions expressed in Cites & Insights may not be those of LYRASIS.

Comments should be sent to waltcrawford@gmail.com. Cites & Insights: Crawford at Large is copyright © 2010 by Walt Crawford: Some rights reserved.

All original material in this work is licensed under the Creative Commons Attribution-NonCommercial License. To view a copy of this license, visit creativecommons.org/licenses/by-nc/1.0 or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.

URL: citesandinsights.info/civ10i2.pdf