Long post ahead.
I found this article on Digg:
http://www.cdmasteringservices.com/dynamicrange.htm

Intruiged, I decided to run some experiments. For these experiments, I ripped the songs from the retail CDs at 16 bit 44Khz wav using dbPowerAmp, except for Atomic Opera which I downloaded in the exact same format from Magnatune. The intention here was to see how mastering has changed over time and how mastering differs between some bands/genres. I opened each file in Audacity and compared the waveforms as well as listening to each one. I tried to take at least one loud song and one soft song from each group. These are just my experiments and as with anything else on the internet should be taken with a grain of salt. I found the results to be pretty interesting so I thought I would share them with everybody.

First song was Evil Woman by Electric Light Orchestra. I happen to have two versions of this song, one on an ELO's Greatest Hits CD produced by Jet records and one on the Afterglow compilation from Epic. Both were recorded by CBS records, but the first I believe was released earlier from the second. The date on Afterglow is 1990.
ELO's Greatest Hits:

As you can see the peaks are nowhere near maxed out, and the detail in the music comes out very nicely.

Afterglow:

The loudness is raised slightly, but not to the point of distortion. This version sounds a little bit warmer and more full than the other.

Second song was Mr. Blue Sky also by ELO.
Greatest Hits:

Same story here. Fairly low level recording with excellent detail.

Afterglow:

Again, the loudness is raised a bit, but not maxed out. Sounds a little more full.

I also did Rockaria, in which both versions were practically identical, and Strange Magic, in which the Afterglow version had loudness raised a little. I don't think it's necessary to post their waveforms, since the results are practically identical.

Next set of songs was from an old Bob Marley CD of my dad's that I have for some reason. I'm not sure about the publisher, but it appears to be CSI. First song was Soul Shakedown:

Again, this older recording shows how the range doesn't peak at it's maximum possible value ever. The recording quality on this disc isn't the greatest, so I can't really say much about that, but it's the recording level that I'm looking at here.

The other song I looked at was There She Goes:

Same story as before, fairly low level recording. The quality on this one is a little bit lower than on Soul Shakedown, but I attribute that to this being a compilation CD that took tracks from different recording sessions.

Next set of songs came from the band Sponge. They're a lesser known band, but they were signed by Sony in the 90s, so you can see how the recording level changes. The first two songs come from their 1994 CD, Rotting Pinata. This album is mixed fairly quietly compared to their later CDs.

First song is the title track, Rotting Pinata:

Here we see how the recording level has changed from the songs produced earlier. While not at the maximum level, the song is significantly louder than the songs above. Still, nowhere near maxed out.

Next song was more mellow, Drownin':

The quiet parts are definately quieter, loud parts are pretty constant, but again, nowhere near maximum.

Next I took two songs from Sponge's 1999 CD New Pop Sunday produced by BMG music. Before even doing this test, I could hear that this CD was mixed much more loudly.

First song is the title track, New Pop Sunday:

Wow. Major difference, even though a majority of the song is pretty mellow. Here the "wall of sound" that's so prevalent in modern music makes it's first true appearance in this experiment. At least the whole song isn't mixed at maximum loudness.

Next song is Live Here Without You:

From the waveform, it looks like this song is one continuous sound. For the most part it is, loud from beginning to end. No dynamic range = bland.

Next set I decided to go with a St. Louis band called The Adored that I saw at a local show. They haven't been signed yet. These songs I got from their EP that they sold at their show. I wanted to see if the independent musicians would do anything differently from a band signed by a major label.

First song is Perfect:

Definately much worse. A single loudness throughout the entire song. I guess this is what happens when you can't afford a large amount of studio time and a decent sound engineer.

I took a more mellow song, All That I Wanted, hoping to see a difference:

At least there is some difference between the lowest and highest volumes in this song.


Next was Incubus, one of my favorite bands. Songs are from Make Yourself, the only retail CD of theirs that I have.

First song is Pardon Me:

Here you can see why Incubus tend to sound better than many other bands out there. Sure there are points where the sound maxes out, but that's the whole reason for having those parts, to provide contrast with the quiet points.

Next song is Stellar:

This one is a lot like Pardon Me. Mixed appropriately in the quiet parts, and maxed out when it's appropriate in the music.

For fun I decided to check out some songs from Opeth. Them being a metal band, I expected the entire songs to be recorded at maximum loudness all the way through. This is from their Blackwater Park CD.

First song is Harvest:

This is more or less what I expected, though I'm surprised that my favorite and the most mellow song on this disc didn't have a greater dynamic range.

Second song was Dirge for November:

No surprises here. The acoustic parts are very well mixed. The hard metal parts are maxed as was expected from the style of music Opeth plays.

Almost done. Next I took a popular CD that I haven't thrown into a fire yet (why? oh why?), Linkin Park's Meteora. Crappy CD, but it illustrates the points in the article perfectly.

With the exception of the very quietest parts, pretty much the entire song is one big block of noise. But you've come to expect that from a band like this, haven't you?

Next song was From the Inside. Yeah, another Linkin Park song, but it's to prove a point:

This one is worse than Somewhere I Belong, which almost seems odd as this song's more mellow parts seem quieter than the other one's.

Finally, last couple of songs I took from Atomic Opera's Alpha and Oranges. Again, independent musicians. I don't really know who they've recorded with and what the production company is, but their CD is being distributed through Magnatune.

First song is Rain Parade:

This song is a breath of fresh air. It was released in 2004, so you'd think it would follow the trend in the article, but it seems that they've found a good sound engineer. When listening, this CD does seem quieter than something like Meteora, but the sound is so much more detailed.

Next song is more mellow, Let Go:

Ok, so it's not that much more mellow on the whole, but you can definately tell which parts have mostly backing electric guitar, and which parts are more acoustic. In any case, you can see what seperates these tracks from everything else that's more modern.

I don't know if this will be really useful to anyone, but I found it interesting. For obvious reasons this doesn't hold for contemporary classical or more ambient music. And yes, I know that this isn't the primary reason popular music has gone done the tubes. But you can't deny that with the compression in place in today's music, a lot of sound quality and range is lost. It just sounds like one big block of noise.

It's nice to see that there are still some bands out there who actually care about dynamic range. If I buy more CDs from Magnatune from some of the other artists I really enjoy, I'll be sure to put up those results as well. While I could do it for their 128 kbps mp3s, I think the degradation in sound quality from a wav will throw off the results. Feel free to try for yourselves though.

If the record studios would take time to train good engineers, instead of just those that can compress the signal and make it as loud as possible, I might even go back to listening to the radio. Unfortunately, while the kids who only want to rattle everyone's windows still want to buy loud music, that's all that is going to be produced.

Actually, I went back and converted one song to various bitrate MP3s because I've never checked which bitrate I should use that sounds pretty good to me. I noticed the waveforms, at least the level of the peaks, doesn't really change. So I decided to do two more comparisons, because I found this particularly interesting.

First song is Into the Storm from Blind Guardian from Nightfall in Middle Earth:

You can see the level is very high, but notice that it doesn't actually pass the maximum threshold.

Compare with No One Knows by Queens of the Stone Age:

Completely maxed out except towards the end.

Into the Storm was recorded about 10 years before No One Knows. What's so interesting is how a much more "metal" style song from Blind Guardian isn't compressed as much as No One Knows, which is a completely different style of song that doesn't need to be compressed like it is.

Of course, every song from Blind Guardian's A Night At The Opera album looks pretty much like No One Knows: one flat compressed sound. Needless to say, Into the Storm is a much more interesting song to listen to than, say, Precious Jerusalem.

What's also interesting is that songs like War Pigs by Ozzy Osbourne are recorded at a lower level than something like Speed of Sound by Coldplay. Soundgarden songs are at a much lower level than songs from Audioslave. And so on. But a song like War Pigs, which lends itself to being maxed out due to the style of music, is recorded at a much lower level than Speed of Sound, which doesn't need to be compressed as much, and that really shows what's been happening in the music industry. The older music just sounds more interesting and alive, and less like white noise than today's recordings.

old songs were mixed keeping the old vinyl player in mind. some newer songs were in the tube amp era, and nowadays digital just keeps the clipping to a minimum because you don't need to raise the DB like before. seems like the whole exponential measurements of DB have gotten obsolete since soundtechs are doing a ****ty job on their mixes and just making the overall mix sound louder because we can "turn it down"

bull**** if you ask me.

opeth's blackwater park cd was recorded using one single low gain guitar track and overlaying it until it reached the "hi gain" feel the big rich(even though triads) chords give off when listening to the cd. hence why it has a crisp but blurry sound. (gain through hard clipping)

and you have to take into notice what kind of song is being mastered. if Harvest had a dynamic range like dirge for novembre or drapery falls, the acoustic feel of the song wouldn't have much impact compared to the rest of the cd.

Exactly. And even though we can "turn it down", what the techs are doing is losing detail in the song. When the peaks are clipped, you lose that high point, because the loudness is identical to the rest of the song. What sucks even more is that even in purely digital recordings, which started in the 80's, the level has just been raised higher and higher over time. Case in point: the comparisons between the two Sponge CDs. I looked at Soundgarden's A-Sides and Audioslave's self titled album, and the difference between the two is amazing.



Yeah, you're right. Actually, looking back at it again, the song doesn't actually ever exceed the maximum level. The gain is raised like you said to give more impact. If you compare it to one of the Linkin Park songs or No One Knows, the difference is actually really obvious.

thing is people can't hear those frequencies. sad thing is they don't understand we still -feel- them. just like you can't hear a bass play a low b in the mix, but if it's missing you KNOW something's wrong even if you're not a musician.

but yes, as long as kids will want loud music, cd's will be produced like this. I see the same phenomenon in trying to sell a kickass guitar tube amp compared to the half-priced digital piece of crap that has way too much processing and noise but is apparently popular for the kids.


it's sad that people who don't know **** about music are the ones with the main buying power.

Wait, what? Which frequencies? If you meant below 20 Hz, then yes, of course we can feel them. But good luck finding hardware that is capable of reproducing below 20 Hz. Hell, find speakers or headphones (at a decent price) that can reproduce 20 Hz with the same dB level as 2000 Hz. Most stuff has some pretty deep rolloff.

the hardware won't reproduce it i know, but the environment still implies those. it's called overtones.

I'd be interested in seeing how this sort of thing is affecting genres besides rock, you know? Because the "wall of sound" effect isn't common in all types of music (and it's one of the things that kind of turned me off to most modern rock music).

Yeah, seems like some metal bands want it louder and louder now and thus there's clipping on CD's now. If you open it up like that, peaks the whole song sometimes.

Okay, in order to contribute something I grabbed a couple of songs:

The Herbaliser - Wall Crawling Giant Insect Breaks (Instrumental Hip-Hop)


Insect Breaks has some pretty noticable variation, probably because the song is something of an ADD-driven number that switches from one sampled groove/beat to the next every 30 seconds or so.

Meat Beat Manifesto - Wild (Avant Garde/Nu-Jazz)


There's a little variation, and the obligatory breakdown. I think one of the reasons it's pretty flat is that there's a lot of reverb and ambient noise going on.

Delinquent Habits - House of the Rising Drum (Rap)


This is what I was talking about when I said the flat-block approach made me avoid modern rock so much. Anyway, this is a great song. Predictably, the structure is pretty repetitive though.

Thoughts?

Orrrr maybe this topic is played out. Nevermind :p