Slashdot Mirror
The Loudness Wars May Be Ending
Hugh Pickens writes "Mike Barthel reports on a technique called brick-wall limiting, where songs are engineered to seem louder by bringing the quiet parts to the same level as the loud parts and pushing the volume level of the entire song to the highest point possible. 'Because of the need to stand out on radio and other platforms, there's a strategic advantage to having a new song sound just a little louder than every other song. As a result, for a period, each new release came out a little louder than the last, and the average level of loudness on CDs crept up (YouTube) to such a degree that albums actually sounded distorted, as if they were being played through broken speakers.' But the loudness wars may be coming to an end. Taking advantage of the trend towards listening to music online — via services like Pandora, Spotify, and Apple's forthcoming iCloud — a proposal by audio engineer Thomas Lund, already adopted as a universal standard (PDF) by the International Telecommunications Union, would institute a volume limit on any songs downloaded from the cloud, effectively removing the strategic advantage of loudness. Lund's proposal would do the same thing for any music you could buy. 'Once a piece of music is ingested into this system, there is no longer any value in trying to make a recording louder just to stand out,' says legendary engineer Bob Ludwig, who has been working with Lund. 'There will be nothing to gain from a musical point of view. Louder will no longer be better!'"
Nigel: Exactly. One louder.
Marty: Why don't you just make ten louder and make ten be the top number and make that a little louder?
Nigel: These go to eleven.
Have gnu, will travel.
I hope this is able to transition to broadcast television broadcasts. I'm sick and tired of commercials being substantially louder than the program they're playing within. Every time a commercial break comes around I have to mute the fucking thing, which seems like the complete opposite of what they're supposed to be trying to accomplish.
Proposed solution: following a standard that limits loudness would remove the strategic advantage of loudness.
What will happen: the standard would be ignored.
You mean the triangle ISN'T supposed to be as loud as the canon fire? :)
It doesn't matter to me how loud a song sounds; I can always turn the volume down or use something like ReplayGain to lower the overall level. The real issue is the compression of the dynamic range used to achieve louder sounding music. This proposal doesn't address that: a volume limit isn't going to provide an incentive to expand the dynamic range, since producers are just going to make sure every song bumps right up to the new brick wall.
Dynamic range simply isn't important to most producers and consumers of popular music now.
Isn't that one of the elements of music that we're throwing away? The element of surprise?
Yes, which is why the standard calls for someone to shout "BOO" at 5x the maximum allowed volume at a random point in each song.
I'm a production director of a radio station so I'm constantly working on commercials. Although I don't go so far as to brickwall things, I do use a variety of compressions, limiters, and EQ to balance out the sound of a commercial -- usually to even out a vocal performance or to make it work with music and sound effects better. There's a cookie cutter and hamfisted way to do it and then there is actually using your ears to do it correctly. That said, what was done with the 5.1 remasters of the Genesis catalog were a travesty. Any dynamism was lost because suddenly what was supposed to be a quieter acoustic section was as loud as the full band playing all out. That's not the way it's supposed to sound.
I haven't read the proposed standard (mostly because, not being a sound engineer, I suspect I wouldn't understand a thing), but wouldn't the problem be solved by limiting not the maximum, but the average instead? us Classical fans get our cannon shots just as Tchaikovsky intended, while mainstream Rock music stops sounding like someone fucking your ear with an ice pick, it's win/win. And as a bonus, anyone wanting to have their music louder would have to have more quiet parts to compensate, meaning they'd be encouraged to utilize the full dynamic range instead of pushing everything to the maximum.
No problem is insoluble in all conceivable circumstances.
You can make all the recommendations and standards you want, but you can't force the studio engineers to obey them, nor can you change the studio executives who are demanding the loudness and writing the checks to the studios. There is a great deal of the attitude in the music industry that "I make a lot of money doing this, and you don't, so my way is clearly right!" So, this movement will probably involve a lot of independent artists. We need pop artists on board.
If we can somehow start a campaign to get people to enjoy an expanded dynamic range, maybe we can raise awareness of how much better music can sound. Maybe albums/tracks engineered correctly could have another small logo somewhere indicating such a thing - call it something like "HDR Audio" (High Dynamic Range) that makes people think.. "Ooh, HD, this one is better than the one without it" or "HDR is the popular thing in photography, so it's probably good with audio".
I'm all for more artists and engineers preserving the vitality of their music.
'Because of the need to stand out on radio and other platforms, there's a strategic advantage to having a new song sound just a little louder than every other song.
Wait, what? If they're all doing this, then how is one still louder than the previous song? And what is this talk of the "radio" platform? You mean the NPR/baseball machine in my car can be used for streaming music? How retro!
One more reason not to listen to shitty pop music.
This is a hacked account, for which the owner can not be held responsible.
Fright-core metal is so totally in this year
As a result, this may be a case where too much dynamic range is lost on the listening audience, as the listener just wants to be able to hear everything without having to fiddle with the volume every few seconds.
Dolby Digital on DVD has the option to compress the dynamic range if you are in a noisy environment (or watching the movie at night). I don't see why this could not also be applied to music. Just have a setting on the player to turn on the compression (or even better - adjust how much compression to use).
Such listeners should just change the dynamics themselves, then: The correct point at which to apply dynamic range compression to compensate for a noisy listening environment is within the playback chain for that particular environment.
It's not so hard. My first portable MP3 player had the ability to apply dynamic compression. My not-so-special Pioneer stereos have this ability as well. So does my Droid. So does even the lowly factory CD player in a 1993 Ford van. And my PC. (I'd go on, but why?)
One can always add more compression/limiting ("loudness"), but once applied it's impossible to take away.
Meanwhile, listening environments haven't changed substantially since the first confluence of the walkman, the portable radio ("boombox"), the home hi-fi, and the car stereo: People still listen variously on headphones, or with barely-adequate portable speakers, or in their home on a properly set-up system, or on ruddy computer speakers (not dissimilar from the discount "rack systems" of yesteryear), or in noisy car, with the same variety of background noise that has always existed when listening to recorded music.
All that has really changed in the past 30 years that it's currently very easy to carry a vast amount of high-quality music in a very portable and readily-retrievable fashion, which was previously impossible. I submit that this improved portability has nothing to do with the dynamic content of that music.
Kid-proof tablet..
Any portable device can easily compress out dynamic range on a good recording as needed. However, a crappy pre-compressed recording cannot be re-expanded.
Nigel Tufnel got new amps, you see.
In the future, I would want to not be isolated from my friends in the Space Station.
that's one aspect (the static part).
you miss the compression (active) that they do in order to 'fit' the envelope 'up higher'.
changing the 'higher' point helps but there are other things going on, too.
its a shame, too. cd has about 90db of dyn range (and modern amps and preamps and easily do that, too) and yet they use a fraction of that. you have a spectrum of 'bit space' to use and you use very little. how sad! how wasteful.
shifting volume (replay gain) is doable. undoing compression is not and that is the real issue.
--
"It is now safe to switch off your computer."
Then it wasn't mixed right. That's the whole point of using reference monitors and your ears. You are supposed to mix with loud, quiet, and in-between in your mind, and check your mixes at all of those levels.
At least, that's how I do it.
One day I feel I'm ahead of the wheel / the next it's rolling over me / I can get back on / I can get back on
Effectively the maximum level is set by the format and is generally define as 0 dB. A format also has a dynamic range, which measures how much quieter a sound the format can capture compared to the maximum. For audio CDs this is -96 dB. The loudness wars refer to taking advantage of the fact the volume the human ear perceives is proportional to the mean level and that music would be recorded with the same maximum level but lower levels mapped higher (e.g. a -40 dB sound is compressed to -20 dB) will sound louder.
This proposal, presumably, addresses this by measuring the volume of a track by measuring the something similar to the mean level. It a little more complicated than that, but I think that's the thumbnail.
ReplayGain fixes average differences in volume between different tracks. It doesn't help when a single track was compressed/normalized so that is has no dynamic range. There's really no post-processing that can fix that.
My Other Computer Is A Data General Nova III.
It's the only CD I ever owned where I could hear the engineers turning the mikes down in the middle, because it had just gotten Way Too Loud. :-)
You are right - limits in the levels are missing, but the proposal explicitly states that it is intended as a groundwork for introducing those by defining a baseline algorithm for measurement.
In the long run, this might only limit clipping due to overly aggressive mixing. The true loudness war caused by compression of dynamic range in the mixing process might not go away as a result. And I don't know how that could even be regulated.
http://www.moonlight3d.eu/
...the listener just wants to be able to hear everything without having to fiddle with the volume every few seconds.
Portable music players have more than enough processing power to do that kind of volume leveling automatically. The artists/engineers/producers can make a product that will sound its best in a good listening environment, and leave the rest to playback.
Almost every song, including ones that aren't "loud" are normalized to 0dBFS. The thing is that they have large dynamic range, so their average signal level might be -30dBFS thus making them "quiet" when played back at a given volume. If you limit the shit out of dynamics, it makes the whole thing louder at a given setting on the volume dial.
That is what people mean when they complain about the "loudness wars." Modern music can't force your system to be loud, I can set my receiver to -80dBref and no sound will be louder than 35dB since that is how it works. The song can't override the volume setting. The problem is that they have no dynamic range, and thus don't sound as good.
A song that has dynamic range you actually turn the volume dial up on. As the "ref" part implies my system is calibrated to a reference point, in particular the THX cinema reference of 105dBSPL for mains, 115dBSPL for the sub. So when I set my dial to 0dB, that is the limit. That is what I set it to for movies, and get a theater experience. However I don't blow out my ears since the average level in movies is usually 30-40dB below reference. So despite the limit being 105dBSPL, I am usually listening to things in the 65-75dBSPL range. That dynamic range is what makes it sound good, and is what lets big hits, well, hit.
Music is squashed down, so I have to listen to it at like -30dBref on the dial. ends up being about the same normal volume level, it just means there's no headroom, that everything is the same volume.
The solution is NOT a volume limit, the solution is to have dynamic range in the files themselves, and put a limiter in the playback device. That way if someone wants it limited, they can turn that on, but you can get full range when you wish.
Now..trouble is...if they remaster them properly...I'll have to buy the collection of many things ALL over again....
Oh well...its only money.
Light travels faster than sound. This is why some people appear bright until you hear them speak.........
Of course, the problem is that your average consumer has been trained to like garbage.
Have you ever gotten into a rental car and taken a look at the audio settings? Invariably bass and treble are turned way up. And what's the first thing people in stores do when trying out a sound system? They turn the volume way up. If it's loud it's good, even if the speakers are clipping.
And how much dynamic range does your average pop song have anyway? Not much, it's just a wall of noise. And then if you're listening to stuff like hip hop then you're also dealing with low quality samples.
Wasn't there are article here on Slashdot several months ago about some survey about audio? Researchers found that the majority actually preferred the inferior sound of compressed audio?
So there's no incentive to improve audio quality. The problem is when this sort of crap spills over to good music.
Well, that's why I prefer to buy things in say, lossless formats...good for the good living room stereo, and I can also rip them to lossy formats for portable players in poor environments (gym,car, etc)....maybe they could give 2x versions..one for portable and one for proper home audio system?
Or, does NO one anymore care for quality home sound reproduction? Geez, when I was a kid...we all wanted to eventually have a good sound system...I've been building mine since I was a pre-teen saving money from neighborhood jobs...etc.
Light travels faster than sound. This is why some people appear bright until you hear them speak.........
"In Absentia" was their commercial breakthrough, you could do worse than that. Porcupine Tree are a leading "new prog" band, meaning that their rock music incorporates progressive ideas but is based on a modern sound rather than retro callbacks. Pure Reason Revolution are another great band in this area. Try Muse for something more digestible (Origin of Symmetry for guitar riffs, Absolution for epicness, Black Holes And Revelations for more experimental yet commercial anthemic stuff with synths.)
It's pretty hard to be a new band that sounds like 60s/70s rock without sounding like generic rubbish. There are a few bands that have sort of done it, like Wolfmother and arguably The Darkness. Clutch have pulled out a pretty solid string of blues/rock albums.
Dither helps with converter differential nonlinearlity, and helps make discretization noise become less obvious, but does not help in improving resolution. Filtering does the latter. The CD is sampled at 44kHz. Using ideal brickwall reproduction filter set at, say 20kHz, you gain log2(sqrt(22/20)) = 0.07 bits of resolution. That's it. Dithering on top of it will make you lose resolution.
A successful API design takes a mixture of software design and pedagogy.
It is not all that much smaller. 24 bits is what you need to listen to a 40dB conversation in 16 bits of resolution, duh. So, if you have a dynamic piece, you will appreciate the difference. Symphonic music recorded in 24 bits and played likewise, in a quiet room, sounds beautiful. The CD sounds worse, and I am no audiophile. It's easy to hear once you listen to the 24 bit system -- though the room has to be at least as quiet as a concert hall would be when they pause.
A successful API design takes a mixture of software design and pedagogy.
The auto compression is yet another thing given to us by the RIAA... I can't think of any other industry group that goes to such lengths to destroy their own industry
Screwing around with the volume means music doesn't carry as much emotion so people are less likely to buy the music.
The format doesn't matter when you're discussing loudness, compression, and "brickwall limiting" as is discussed. The audio engineers (and sometimes the musicians) are distorting the signal somewhere along the path to the final master. Thus, what you have in your lossless audio is lossless garbage, and what you have in your lossey audio is lossey garbage.
For example, poke around the Internet for Muse's "Knights of Cydonia". The tracks used in Guitar Hero are cleaner than the mix used for the CD, so someone went through and actually remixed the song from the GH tracks and got a pleasant-sounding result. As a counterpoint, listen to pretty much anything from Foo Fighters' "Wasting Light". They have mp3-quality audio available on their website, and it is not difficult to hear the impact of a good recording/mastering process.
I wish Muse would release a properly-mastered version of Black Holes and Revelations. I think I would even pay for a SACD version of it, just knowing that the intended audience has a higher listening standard. It's sad that the CD just doesn't play well outside of a car stereo, and it paradoxically makes the CD rock less because it's unbearable to crank the volume.