The Loudness Wars

[Moon Pix]

http://www.stylusmagazine.com/articl...nd-forever.htm
http://www.youtube.com/watch?v=3Gmex_4hreQ

What do people think about all this? Personally I think its dreadful. When I heard At War With the Mystics a while a go I thought it sounded like mud and I didnt like it. I guess the loudness war is the reason why.

Bastards.

[Paul]

Your Stylus link is busticated.

[bobbydj]

I didn't look at the links and yeah the 1st is broken. But I think I know what this is about. It's about mastering, right? To an lo-fi anorak like me the issue of mastering is really close to my heart. I actually emailed Robert Griffin at Scat to ask whether the geebs' VoT had been mastered, and if so, from what. I guessed they sent him cassette, but he didn't comment on that. What he did say was that he had all his stuff direct metal mastered. Now, I think that actually relates more to manufacturing side than the pre-pressing side. Oops - yeah, I'm talking about vinyl here. Interestingly though, it's apparently best to master differently for cd. This is because the cd manufacturing process can surive much more brutal brick wall limiting, where the dynamic range is attenuated to shit. Whereas with vinyl, the cutting head can actually burn out during the manufacturing process if it gets such constantly hot levels. So that's actually an important point - brick wall "loudness wars" limiting isn't so bad providing you plan to release cds only. But, if you're going to have multi-formats inc. vinyl, you better becareful of how loud your mastering is. If you have too compressed a signal for cutting vinyl, the end result could well sound very bad indeed - there will be a lot of distortion at the peaks. I think this is something that maybe isn't appreciated as much as it could be. Btw, if this all seems like so much technical bullshit, that's only cos I can't explain it very well, plus I don't actually understand it too well either.

But on the more general question of the loudness wars and each mastering engineer's attempt to create THEE loudest track on the radio, I totally agree that it's a total load of bollocks. Various people have taken something of a stand against this over the years, one being albini (I think). The Shellac releases retain a compararitvely high dynamic range - which means the quiet bits are quiet and the loud bits are fucking LOUD. Unlike with some bands - most? - whereby the quiet verse is just as fucking loud as the kicked-in chorus. That's when the loudness wars get REALLY stoopid. It's one thing to have a 2 1/2 minute single - a song like, say, basket case - be loud all the way through, but if your song's get any dynamics, a loudness wars mentality will erradicate that stuff.

[Moon Pix]

http://www.stylusmagazine.com/articles/weekly_article/imperfect-sound-forever.htm

This one works for me.:)

If that doesnt work try this one:
http://www.cdmasteringservices.com/dynamicrange.htm

[Elijah]

http://www.stylusmagazine.com/articles/weekly_article/imperfect-sound-forever.htmThat's a great article. It put what I've been thinking into more words than I could ever be bothered to write or say. Thanks for the link, MP.

I think a fantastic example of what the author is talking about can be found if you compare the vinyl version of Yo La Tengo's "Pass the Hatchet I'm Goodkind" to the 192kbps MP3 that's available on Matador's download page. The vinyl sounds amazing. It makes me want to jump out of my chair. The MP3 makes me tired. I think this is an illuminating comparison, because the vinyl and CD (which I assume the MP3 is sourced from) are separate masters; and the MP3, which to outward appearances sounds exactly the same, produces a drastically different reaction than the LP when actually listened to.

Not that I think MP3's alone are the problem. It's how I listen to most my music, albeit at a marginally higher bitrate (>224kbps), and encoded at the slowest possible speed with a LAME encoder. This means that a 40 minute CD takes about 60 to 90 minutes to encode; but the resulting sound file is well worth the wait. Well-mastered CD's (Pink Floyd's early 90's reissues come to mind) can and do sound excellent when encoded properly. The encoder you use makes a huge difference. At the very least, it's like the difference between using a normal-bias and a high-bias cassette.

In ten years, music coming out today should sound even more dated than a Samantha Fox track sounds now. Or maybe not.
.

[Moon Pix]

You may know this Elijah (Im only just getting hip to all this myself), is this whole loudness issue the reason that some modern records, At War With the Mystics being a prime example, sound muddy and indistinct when listened to through headphones?

[bobbydj]

Ime, brick-wall limited cds don't sound muddier, as such. They might sound distorted. But I tend to think of the notion of "mud" relating to EQ - like, accentuated lows and low-mids, and a lack of definition in the up mids and high frequencies. Maybe overt compression affects frequency responses? Yeah - I think I read that somewhere. But to be honest, I thought it tended to take some of the lower end of the spectrum rather than the top. *shrugs*

[davidleeroth]

Whereas with vinyl, the cutting head can actually burn out during the manufacturing process...

I'll admit that most of the knowledge of making records comes from watching the How It's Made- series but I always thought the reason for "quieter mix" was that you can't physically fit the wider grooves into the record, unless you're fine with 10 minutes of music per LP side. And even if you did, your stylus couldn't track the grooves.

I had read the cdmasteringservices article before and did a little comparison test with JMJ's Oxygene like two months ago. I had the 1976 vinyl and 1997 remastered CD version and borrowed the 1983 and 1991 CDs from library. I ripped and compared some samples with Audacity. The 1983 version was pretty damn close to the vinyl release, even in sound levels. The 1991 version was much louder but with not too much compression. The remastered version had a lot more compression and anybody would hear the difference between it and the first CD release.
I'll have to buy the next version just to see if they can make it clip.

[Moon Pix]

Ime, brick-wall limited cds don't sound muddier, as such. They might sound distorted. But I tend to think of the notion of "mud" relating to EQ - like, accentuated lows and low-mids, and a lack of definition in the up mids and high frequencies. Maybe overt compression affects frequency responses? Yeah - I think I read that somewhere. But to be honest, I thought it tended to take some of the lower end of the spectrum rather than the top. *shrugs*

We might be talking about the same thing jut using different terminology. When I listen to At War With the Mystics through headphones I can't hear any detail or anything, I can't hear inidvidual instruments very well. All I can hear is this horrible noise were everythings all blurry and indistinct, like looking at the world through stained glasses.:(

[tinobeat]

We might be talking about the same thing jut using different terminology. When I listen to At War With the Mystics through headphones I can't hear any detail or anything, I can't hear inidvidual instruments very well. All I can hear is this horrible noise were everythings all blurry and indistinct, like looking at the world through stained glasses.:(


That's because you're listening to the Flaming Lips.

[Moon Pix]

That's because you're listening to the Flaming Lips.

Maybe but not even Loveless is as blurry as that record and MBV were known for that sort of thing.:confused:

Does Matador have any policy regarding compression when it comes to its own releases?:)

[Lukas]

In terms of the Pavement reissues, when I compare the originals to the reissued there's a big difference in terms of loudness. I'm not sure about the compression end of it though...

[Patrick]

Does Matador have any policy regarding compression when it comes to its own releases?

As so often when it comes to Matador, it's 100% up to the bands.

Except in very rare occasions (when we're asked to), we only get involved once a record gets mastered. When we know an album has been recorded in a real recording studio onto tape with minimum processing, we'll recommend true all-analog mastering for the vinyl... most notably the Cat Power album (where some tracks used digital processing and others didn't) and the two Mission Of Burma albums.

The Pavement reissues, I don't know... paging Jesper.

Patrick

[Moon Pix]

Except in very rare occasions (when we're asked to), we only get involved once a record gets mastered. When we know an album has been recorded in a real recording studio onto tape with minimum processing, we'll recommend true all-analog mastering for the vinyl... most notably the Cat Power album (where some tracks used digital processing and others didn't) and the two Mission Of Burma albums.Patrick

Is this the reason that a lot of people prefer vinyl to cd?:)

[Patrick]

Is this the reason that a lot of people prefer vinyl to cd?:)

Well Moon Pix... that question has been argued over for the past 24 years. Vinyl has (potentially) higher bandwidth than CD, has a smooth waveform as opposed to a jagged waveform, may produce even-order distortion as opposed to odd-order distortion (the latter is harsher on human ears). On the other hand, CD sound has improved by leaps and bounds since it was first introduced, reducing jitter and employing upsampling to get around some of these problems. Also, CD does not suffer from surface noise or warps, and inarguably produces tighter (if not richer) bass and finer (some might say etched) high-level detail.

There's also a whole cultural and psychological aspect - playing LPs is a much more involved (or involving) affair - you have to physically lift the needle on, make the mechanical interface yourself, watch the tonearm move across the record, turn the record over after only 20 minutes, look at a full-sized cover with artwork and information - CD you put in a slot (or you put 20 of them in a caddy) and hit shuffle.

In other words there are benefits to each format...

Doubtless preferences for one format over the other involve all of the above.


(MP3 = CD to the tenth in the above comparison.)

Patrick

[Jesper]

Patrick is right, mastering always involve the artists 100%, in most cases they request the mastering engineer as well. As for Pavement reissues sounding louder - CDs have the capacity of being louder today than they did 10 years ago without losing dynamic range. Greg Calbi certainly isn't a volume freak either (Howie Weinberg is however).

[Moon Pix]

I gather from all this that the actual band playing the songs into microphones is only a tiny part of the process of making a record.

[Jesper]

Well, I would maintain that certainly the performance (and perhaps the recording) of an album is the absolute crucial part - anything else, including gear and mixing, is all about enhancing the quality of the already existing performance. My favorite albums have almost all been recorded by equipment and engineers and in studios that no one has ever heard of. Meaning, the quality of the perfomance is what matters to me, and everything else is just bonus. I don't know if there is such a thing as a perfect album, where the best performance, equipment, engineers etc etc have met - I kind of have to say that I hope that it doesn't exist. Many seem to think that if you use an amazing studio and tons of vintage gear your record will automatically be amazing. I think the history of recorded music has proved that idea fundamentally wrong time and time again. To me great music is about amazing accidents, and the opposite of total control and proper use of gear. But I have of course heard convincing opposing arguments too...

this could in turn lead to a neverending thread about 'the best sounding albums of all time' and other related lists... and I would have no idea where to start!

[earl grey]

this could in turn lead to a neverending thread about 'the best sounding albums of all time' and other related lists... and I would have no idea where to start!

hey, that's not a bad idea...

[bobbydj]

CDs have the capacity of being louder today than they did 10 years ago without losing dynamic range.

Everything I've ever read seems to go against this. The cds of 1983 are the same as 2007, in terms of there peak signal. If you put a cd on that was made in 1983, it's peak at nought will be the same 'loudness' as one made today that peaks at nought. But then, I'm certainly no expert in this field.

[the Pawnbroker]

Well, I would maintain that certainly the performance (and perhaps the recording) of an album is the absolute crucial part - anything else, including gear and mixing, is all about enhancing the quality of the already existing performance. My favorite albums have almost all been recorded by equipment and engineers and in studios that no one has ever heard of.

That book about Pavement a few years ago had interesting stuff about recording Crooked Rain, Crooked Rain at Random Falls -- apparently, Pavement had access to some pretty decent vintage gear. I've always thought that album had a very rich, warm, thuddy sound for a low-fi album.

[bobbydj]

So it's not a lo-fi album then. If it was done in a pro studio, the 'fi' is going to be 'mid' at worst, surely?

[tinobeat]

Everything I've ever read seems to go against this. The cds of 1983 are the same as 2007, in terms of there peak signal. If you put a cd on that was made in 1983, it's peak at nought will be the same 'loudness' as one made today that peaks at nought. But then, I'm certainly no expert in this field.

It might not be louder, decibel wise, but what he meant is they can be that volume without sounding like ass. The dynamic range of digital audio has improved exponentially over the decades.

[clell tickle]

This has been discussed in great detail at tapeop.

[bobbydj]

But isn't the loudness wars thing about shrinking dynamic range - not increasing it? I mean, that's what brick wall limiting is about - trying to get *all* the signal to peak as high as possible (so it sounds really loud on the radio, etc. etc.).

[Miss Tasty Princess]

Everything I've ever read seems to go against this. The cds of 1983 are the same as 2007, in terms of there peak signal. If you put a cd on that was made in 1983, it's peak at nought will be the same 'loudness' as one made today that peaks at nought. But then, I'm certainly no expert in this field.I find that, in general, the older a CD is, the higher I have to crank the volume on the victrola. In fact, when making mix CDs for trades on this board and elsewhere, I've had to go with tracks that are roughly equal volume-wise in order to avoid large drops or jumps in volume and age is a big factor. My new computer has software that will allegedly handle this but I'll believe it after I've tried it.

[Elijah]

The cds of 1983 are the same as 2007, in terms of there peak signal. If you put a cd on that was made in 1983, it's peak at nought will be the same 'loudness' as one made today that peaks at nought.This is clearly an assumption on your part, because any attempt to verify this claim with empirical evidence would render it incontrovertibly false.
.
My new computer has software that will allegedly handle this but I'll believe it after I've tried it.iTunes does that. It works pretty well, too.

If you feel like spending $90 (and let's face it--who doesn't?), Roxio Toast gives you the option of tweaking the levels for each individual track, and creating custom cross-fades.
.

[bobbydj]

This is clearly an assumption on your part, because any attempt to verify this claim with empirical evidence would render it incontrovertibly false.


Well, I dunno. Maybe.

But if you wanted to run tests on a random sample of cds from now going back to 1983 you would find that the peak signal on each cd is at 0db. That's the first thing you have to grasp (it can't go any higher! I mean - really though - there is nowhere for the singal to go. Sure, it could distort and break up - but even then it couldn't go 'above' nought. It would just convert into square wave distortion. There's nowhere for it to go).

What's happened *since* 1983 is that more and more mastering jobs have striven to compress mixes, and compress them again and again, crushing the peaks and troughs of the dynamic range until the whole mix looks like a straight line, maxing at 0db.

Now, this will undoubtedly sound louder to the ear. But in one sense it simply is not - because the peak is still the same. It's just that the mix is peaking at that high point for a much more sustained length of time, whereas before it may have just been certain points in the chorus, e.g.

So with all due respect, this isn't just about an assumption. I do burns and mixes of my own, using a domestic 2 drawer burner with clear metering. You can sit and watch how various mixes peak. That's what I do. So that's your empirical evidence right there. With this stuff, the irony is that you have to use your eyes, not just your ears. A slammed to fuck mix will sound massively loud. But its SPL won't be any higher than that of the most dynamic mix you care to cite (providing that mix is normalised, i.e.). This is my current understanding, at any rate. But I'm definitely up for learning more about it, so school away.

[bobbydj]

I find that, in general, the older a CD is, the higher I have to crank the volume on the victrola. In fact, when making mix CDs for trades on this board and elsewhere, I've had to go with tracks that are roughly equal volume-wise in order to avoid large drops or jumps in volume and age is a big factor. My new computer has software that will allegedly handle this but I'll believe it after I've tried it.

Absolutely! I agree totally. In fact, though, this was the same back in the days of vinyl and cassette. The cheap compilation LP version of a hit single (crammed onto the end of one side, after 14 other tracks) would be LOADS quieter than the 7" version you had of the same track. And so when you made your comp tapes, you'd never use that shitty comp LP version cos it was always so lame sounding and even with the recording levels cranked to shit it was much quieter than the tracks either side of it (which had come off well-pressed singles or whatever).

Well anyhoo, plus ca change and all that good shee-ite. Here we are today in the digital era and yeah, the cds from back at the dawn of digi had a much more relaxed approach to mastering. The mastering engineers let the mix breathe, preserving much of the dynamic range and letting the recording sound much more like the final studio mix. But then the late '90s happened and everyone decided they wanted to be the loudest thing on the radio and out went dynamic mixes. So yeah, an older cd will sound significantly quieter than that old stuff. Downer, if you ask me.

[Elijah]

So with all due respect, this isn't just about an assumption. I do burns and mixes of my own, using a domestic 2 drawer burner with clear metering. You can sit and watch how various mixes peak.Hmmm. Well, back when I was making mixes with a tape deck, it was impossible not to notice that starting in the early nineties, compact discs began to have higher peak levels, regardless of how compressed they were. Typically, an older CD would require a higher record level than a newer CD. I've also made mixes with Roxio Jam, which can automatically normalize the peak levels on each track, and then allow you to tweak said sound levels if needed. 99 percent of the time, a track from an older CD peaks lower than a newer one.

I'm not trying to be contentious here, but these are simple, verifiable facts.

[bobbydj]

I'm not trying to be contentious here, but these are simple, verifiable facts.

What are you all of the sudden?? The Karl Popper of audio engineering??

[Miss Tasty Princess]

Here we are today in the digital era and yeah, the cds from back at the dawn of digi had a much more relaxed approach to mastering. The mastering engineers let the mix breathe, preserving much of the dynamic range and letting the recording sound much more like the final studio mix.I'd say it's more that in the early days, the mastering engineers were paid to get things mastered ASAP so the buyers could replace their entire LP collection with CDs ASAP. Mixes didn't breathe then; they were suffocated. Spots in songs where there should have been a big jump in sound (quiet to loud) were smoothed out ("Fight Fire" on Metallica's Ride the Lightning, for example). Plus, clever things were done to fit dbl-LPs onto single CDs like fading songs out early and, in many cases, just leaving songs off with a liner note telling the buyer that he or she won't mind! :rolleyes:

[Moon Pix]

Well, I dunno. Maybe.

But if you wanted to run tests on a random sample of cds from now going back to 1983 you would find that the peak signal on each cd is at 0db. That's the first thing you have to grasp (it can't go any higher! I mean - really though - there is nowhere for the singal to go. Sure, it could distort and break up - but even then it couldn't go 'above' nought. It would just convert into square wave distortion. There's nowhere for it to go).

I dont know if this is relevant to what youre talking about but I saw this DVD recently about the Velvet Underground. One of the scenes in it showed a VU meter reacting to "Sister Ray" and time and time again the little needle hit the +4 mark.

[davidleeroth]

Well, the way I see it, the theoretical peak level was and still is 0dB but back in the good old days the peak level for a recording was set as low as -6dB by choice. Now, when stuff gets cranked up and compressed to hell, you'd most likely find no peaks below the -6dB.

[bobbydj]

I dont know if this is relevant to what youre talking about but I saw this DVD recently about the Velvet Underground. One of the scenes in it showed a VU meter reacting to "Sister Ray" and time and time again the little needle hit the +4 mark.

For sure! That's what you could do with the analogue format, or tape at least. Engineers would - and do - regularly go beyond 0 when using tape. It's a totally different proposition than overloading digital formats, which comes out sounding like the sound of crinkled cellophane or something.

[bobbydj]

I'd say it's more that in the early days, the mastering engineers were paid to get things mastered ASAP so the buyers could replace their entire LP collection with CDs ASAP. Mixes didn't breathe then; they were suffocated. Spots in songs where there should have been a big jump in sound (quiet to loud) were smoothed out ("Fight Fire" on Metallica's Ride the Lightning, for example). Plus, clever things were done to fit dbl-LPs onto single CDs like fading songs out early and, in many cases, just leaving songs off with a liner note telling the buyer that he or she won't mind! :rolleyes:


That's a good point. However, there's no doubt about it that mixes are more compressed today and - consequently - that dynamic range has shrunk hugely. All in the name of getting everything slamming at peak levels (i.e. nought).

[bobbydj]

Well, the way I see it, the theoretical peak level was and still is 0dB but back in the good old days the peak level for a recording was set as low as -6dB by choice. Now, when stuff gets cranked up and compressed to hell, you'd most likely find no peaks below the -6dB.

Yeah, exactly.

[Moon Pix]

http://www.cutesphere.net/data/people/globulator/Hi-Fi/index.php?sid=0

"The irony is that the lower bandwidh MP3s contain less clipping artifacts than the 'better' MP3s and CDs, and therefore are less tiring to listen to."

Is this actually true?

[Patrick]

http://www.cutesphere.net/data/people/globulator/Hi-Fi/index.php?sid=0

"The irony is that the lower bandwidh MP3s contain less clipping artifacts than the 'better' MP3s and CDs, and therefore are less tiring to listen to."

Is this actually true?

Interesting, though highly tendentious, article. How many people actually believe that the reasons the RIAA is losing customers are: compressed music and bad sound quality? Still, it's always entertaining to see people come up with another way to bash the industry - the possibilities are endless!

He does have some good links. I built an Aspen Amplifiers AKSA 100 N+. Great solid state kit amp.