Friday 12 December 2008

Loudness war or: How I learned to stop enjoying modern music


The final straw

Just last week, I picked up Hold on now, youngster..., the debut album from Welsh indie pop outfit Los Campesinos! Many reviewers have described them as Belle & Sebastian gone punk. Belle & Sebastian are one of my favourite bands; the last time I picked up an album by an unknown band who were compared to B&S was Neon golden by the Notwist (the love-child of B&S and German electro-pioneers Kraftwerk), and that album paid high dividends for me. So picking up the Los Campesinos! album at the relatively cheap price I saw it at seemed like a no-brainer to me.

Hearing an album by a band you don't know for the first time can be an exciting yet daunting experience. In this case, I had no idea what to expect. Male or female vocalist? Quirky or old-skool punk? The answer to these 2 questions was both and quirky respectively.

But what I wasn't expecting (although maybe I should have, in hindsight) was the wall of noise I was confronted with, within only a few seconds of the opening track. This is the audio waveform of what I heard:


Unfortunately, Los Campesinos! are not alone in the noise stakes. The crazy brick wall you see above is part of a disturbing recording industry trend which has become known as the loudness war.

The loudness war is associated with the perceived need for albums to be mastered louder and louder as each year goes by. The additional loudness is achieved in the mastering process through excessive use of technology known as dynamic range compression.

I have been against the loudness war in all of its forms ever since I become aware of it. But this Los Campesinos! album was the first time that I have ever returned an album for a refund because the CD was too bloody loud to listen to.

Background reading

I'd like to thank Nick Southall, of the sadly departed Stylus Magazine, for his insightful article Imperfect sound forever. This article is essential reading for anyone who even remotely takes their music listening seriously. As an avid reader of Stylus Magazine back in the day, this was the article which enlightened me with information about dynamic range compression and its mis-use in the record industry.

Nick explains compression in a lot of technical detail, better than I could ever do. Nick has been a bit of crusader against the loudness war for many years, and I really respect his opinion on music.

My article will only gloss over the technical details, so please do some background reading if you are interested in further information.

My musical habits

I'm not an audiophile by any stretch of the imagination. Since I purchased my iPod a few years ago, 98% of my music listening happens on the iPod (the other 2% being on CD when I buy an album on the weekend and don't rip it until I come to work on Monday).

My music is ripped at 160kbps CBR AAC. While a lot of music fans would spit at such a low bitrate (and I'll admit that this is far from CD quality), this is good enough for me. Whenever you decide on a bitrate to use, you always need to weigh up quality vs. file size. Back in the day, I only had an 80GB iPod which was disturbingly close to being filled up. I made an executive decision (with invaluable input from MattT) to rip all of my music at 160kps; it was a happy medium between file size and quality.

Down in the music rut at midnight

Over the past few years, I have found myself much less excited about newer music. There are of course some exceptions, but my musical listening tastes have definitely headed towards music from the 80s and earlier (and I'm not just talking about the cheesy fluff here). On my shortlist of potential blog posts was one entitled Down in the music rut at midnight, where I was going to talk about how I was so sick of new music coming out and how old music brings so much more excitement to my ears. But that post will have to wait for another day.

There are many subjective reasons why new music may stop exciting you. After all, there's only so many ways that you can combine vocals, guitar, bass and drums in innovative ways before history will start repeating itself (obviously I'm focusing on rock music here). If it's been done before, it's bound to be much less exciting the 2nd time around; I've certainly alluded to this in previous posts.

But there are also objective reasons why new music may stop exciting you. When I compiled my top 10 albums of 2005, the Go-Betweens' last album Oceans apart made it into the 10th position. At the time, I said the following:

Robert finds a renewed energy in his songs Here comes the city and Born to a family, while Grant reaches the melodic heights that came so naturally to him during their golden years on The statue and Finding you. This album would be higher on the list if it weren't for what appears to be a mastering problem on some of the tracks, where they sound distorted (especially on the iPod). It's both unfortunate and frustrating that a great album has been tarnished by a technical problem.

I wrote this review in December 2005. This was before Nick Southall had published his article, before I knew about dynamic range compression and the loudness war. Back on those days, I thought the only war that was going on was in Iraq. It is only with the benefit of hindsight that I can look back at this post, and see how the loudness war was affecting my enjoyment of music.

From a pure songwriting perspective, Oceans apart was one of the Go-Betweens' finest efforts (only their masterpiece 16 Lovers Lane can beat it). It's only the disgusting production which lets it down. I'm now at the point where, if I feel like listening to a Go-Betweens album, I would choose any album but Oceans apart to listen to.

Excessive compression in music is a little bit like having a dead pixel on your brand new LCD television. You can wallow away in blissful ignorance, enjoying the entertainment and being none-the-wiser. But as soon as you are aware of it, you won't be able to ignore it. So on behalf of the readers of this post who are like me, I humbly apologise for (potentially) destroying your enjoyment of a lot of modern music.

What is dynamic music?

In its simplest form, dynamic music has quiet bits and loud bits. If we listen to an album that has a wide dynamic range, we'll often find that we need to regularly adjust the volume knob while we listen to it. This is because the quiet bits are a little too soft to hear (so we turn it up to hear the subtle details). Then the loud bit comes in, and the new volume we adjusted it to is now too loud so we quickly turn it down again. You may especially notice this while listening to dynamic music in a car, where background noise also interferes with your listening experience.

Let's compare this with music which is generally played on the radio, or through PA systems in shopping centres. Music which is transmitted on the radio is compressed so it fits into the frequency band allocated to the station. Music played on PA systems is hardly an intimate experience, and its purpose is often for background music. Neither of these types of music would generally be considered dynamic.

Let's get (a little bit) technical

Okay, I'm gonna try and keep it fairly simple here. I'm not an expert in sound recording techniques and I'm not going to pretend that I am. There are great articles on the web that could explain this much better than I can. Treat this more as a Dummies guide to dynamic range compression™.

Over the course of the past 15 years or so, the loudness war has lead to albums becoming progressively louder as each year goes by. Why? Well, record company big-wigs seem to think that loud records sound better and will sell more records:

"If my record isn't as loud as your record, it will sound less powerful, less punchy and less impressive. This means it will also get less radio play. And this will result in less record sales."

So what's the solution? We'll have to one-up you and make sure ours is even louder than yours. It's the musical equivalent of the bully in the playground picking on the little guy with glasses, just to make sure that he shows him who is boss.

So what does dynamic range compression have to do with loudness?

Unfortunately, the maximum amplitude of a CD audio waveform is fixed. Trying to increase the sound of the quiet bits will cause the loud bits to exceed the maximum volume. If the waveform exceeds the maximum, it needs to be chopped off to keep it within range. This results in a square wave at the peak, and when a square wave is played back through a CD audio player, it "clips" or sounds distorted. Have you ever tried to play a data CD through a CD player? It's the same sound you get from that, but a bit more subtle as the clipping may only last for a fraction of a second.

To prevent clipping but still maintain the loudness that is desired, the entire sound wave is compressed to a similar volume, and then the compressed signal is adjusted to the new loud volume. The idea is that the whole sound wave is flattened into a homogenous block, which can then easily be adjusted to increase loudness without the fear of distortion or clipping.

In practice, it doesn't always work out. Many modern recordings, despite the "best" intentions of the engineers, still end up suffering from clipping and distortion. This has become disturbingly more prevalent in the past few years.

Compression has proper and tasteful applications during the recording process. For example, individual instruments within a recording can be compressed to make them sound more punchy, or reduce their sustain. This is what producers and engineers do; utilise the studio as another instrument to help make the end result (the album) sound as good as possible.

The problem with compression is that the music industry has been mis-using this technology for the past 15 years or so. Instead of compressing individual elements within a mix, the engineer takes the (already mixed) recording and compresses the entire mix again before it is pressed to CD. This practice completely narrows (and sometimes eliminates) the volume gap between the quiet bits and loud bits, so that everything ends up at a similar volume.

What about data compression?

As I said earlier, I compress all my my music at 160kbps. This is what is known as data compression, and it is completely different to dynamic range compression.

MP3 (and other similar) data compression algorithms are concerned with reducing the overall size of the file so you can fit more information on to your portable listening device. This is achieved by removing the inaudible (to humans at least) portions of the music - mainly the very high frequencies (which only dogs can hear), and very low frequencies.

Data compression also involves other smart techniques. If the compression algorithm detects two waveforms being played simultaneously, one quiet and one loud, it will remove the quiet waveform as it will be drowned out (and thus inaudible) by the loud waveform anyway.

While some data is lost forever in data compression, if you choose a decent enough bitrate, most listeners (except the keenest audiophiles) will find it difficult to tell the difference between a compressed MP3 and the original CD audio.

Most importantly, data compression is designed to preserve the dynamics of a recording.

Musical enjoyment

Dynamic range compression sucks the life out of music. The excitement that you get from a snare hit is gone in a compressed recording, because it is the same volume as the rest of the recording. That funky bassline that seems to come out of nowhere becomes part of the furniture. The beauty of a quiet, whispered vocal is lost because the volume of the vocal is the same as the rest of the mix. It's just not musical.

The disturbing thing is that, to the untrained ear, loud and compressed records do sound better. INITIALLY. It is only with subsequent listens that the listener realises that there is absolutely no depth to the recording. No hidden details, no subtle layers, no instruments buried under the mix.

Why not just lower the excessively loud volume by using the volume knob? Sure you can do that, and it's probably a good idea if you respect your ears. But thinking that this will improve the sound is completely missing the point. The signal has already been compressed, which means the dynamic range has been lost forever. All the volume knob will do is reduce the overall amplitude; the variation (peaks and valleys) have already left the building.

Have you listened to any music from recent years that has given you a headache? It's probably a victim of the loudness war. The human ear can be fairly tolerant to loud music and noises, but its tolerance is inversely proportional to the length of time it is subjected to the noise. When the music has a wide dynamic range, the quiet bits give your ears (and head) a break and a little bit of time to recover. This prepares them for the loud bits when they come.

WHEN THE MUSIC IS MASTERED AT A CONSTANT AND LOUD VOLUME, LISTENING TO IT IS SIMILAR TO READING AN EMAIL OR WEBPAGE WRITTEN ENTIRELY IN CAPITAL LETTERS. IF EVERYTHING IS LOUD AND IN-YOUR-FACE, IT GETS PRETTY TIRING AFTER A WHILE DOESN'T IT? THERE'S NO RELIEF.

But I like my music loud!

So do I! A lot of music sounds much better when played at a loud volume. So if the CD you want to listen has been mastered at a quieter volume to preserve the dynamics of the recording, how do I enjoy it loud?

There's a very simple solution to this problem. Reach for the volume knob and turn it up. This will increase the overall loudness of the entire waveform, but keep the new waveform in proportion with the original, quieter waveform. The peaks and valleys will be maintained and the dynamics will be preserved. Music can be both loud and dynamic, as long as the dynamics haven't been compressed out of the CD.

Some examples

Here is the audio waveform of I'll regret it all in the morning, a tender ballad by Richard & Linda Thompson from their 1975 album Hokey pokey:


And here is the audio waveform of Cast no shadow, a tender ballad by Oasis from their 1995 album (What's the story) morning glory:


Think you can spot the difference? Note that these songs are both [supposed to be] quieter moments on their respective albums.

Look at the peaks and valleys in the Richard & Linda Thompson track. Notice the variance in the wave form. Even if this song isn't to your taste, you get a pretty good feeling from looking at that waveform that it will at least sound pretty good.

Now look at the Oasis waveform. It just looks like a wall of noise. There's very little variance in the track. There's no highs, no lows, and nothing in between. IT IS JUST CONSTANT NOISE. All of the emotional intensity, subtlety and nuance has been sucked out of the music by excessive compression.

Don't get me wrong - that Oasis album is in my top 10 albums of all time. I grew up with this music. But there's no denying that the production is absolute shite.

The brick wall you see above is not what Phil Spector dreamt up when he invented the wall of sound.

Remasters

Remasters are one of the most disturbing casualties of the loudness war. Why? Because they take an album from the past which probably had brilliant (albeit a little bit quiet) sound and "remaster" it for the purposes of selling it off to a new generation of music listeners who may have never heard it before. The record companies would lead you to believe that a remastered recording is superior to the previous one in every way.

Unfortunately, the reality is a lot less pretty than this. In the modern age, remastering is often not a lot more than making the CD a lot louder so it sounds good when listened to next to modern CDs. As described earlier, there are limits to how loud you can make a CD, which is where our evil friend compression comes in. And as described earlier, compression of this form is completely lossy - once a CD is compressed, the dynamic sound is gone forever.

Record companies try to argue this by saying that the majority of music listeners in this day and age listen to music on their portable listening devices, such as iPods. The justification is that if you put your whole music collection on shuffle, you should be able to listen to an old (quietly mastered) Beatles song followed by a new (over-compressed) Muse song without having to get up and adjust the volume.

As I said before, I do practically all of my music listening on the iPod too. And last time I checked, my iPod had a feature on it called Sound check which automatically normalises the volume across subsequent songs so you don't run into this exact problem the record companies are trying to avoid. I'm not sure if other MP3 devices have a similar feature, but I don't care. Maintaining a consistent volume across songs on an iPod is no excuse for over-compressing music on the CD itself.

I'm now at the point where I will actively track down the original CD pressings of old albums, even if it means having to buy them 2nd hand on eBay. This is coming from a guy who only a few years ago sold a few of his original CD pressings to "upgrade" to a remastered version.

One thing to be aware of is that there are always exceptions. Some remasters are done tastefully, and with respect for the music. This is where the power of the Internet at your fingertips is a great thing. Read reviews, and not just from professional critics (who I suspect are often getting kickbacks from record companies). Read user reviews on Amazon and similar websites. You'll quickly find out whether a "remastered edition" is an improvement or butchering of the original album.

The good, the bad and the ugly

Bob Dylan was recently quoted as saying the following:

"I don't know anybody who's made a record that sounds decent in the past 20 years, really," ... "You listen to these modern records, they're atrocious, they have sound all over them. There's no definition of nothing, no vocal, no nothing, just like ... static."

While this is largely true, there are luckily still a few
artists and bands out there who have the balls to stand up against the evils of the loudness war. These musicians have my utmost respect for standing up against this disturbing trend.

Manchester band Elbow have stood up against the loudness war, refusing to over-compress their music like many of their contemporaries, including Coldplay. Their latest album even has the Turn me up! logo on it, a non-profit organisation campaigning for the release of more dynamic records.

Ben Harper's latest album Lifeline proudly boasts that it was recorded entirely on analog equipment, with no use of Pro Tools. And you can hear it on the recording - rich, warm and dynamic. This is how music is supposed to sound; when listening to this album (from 2007) next to other albums from the same year, it sounds like a real musical experience and not a manufactured one.

Other recording artists with integrity, like Tom Waits and Augie March, have also maintained a high degree of dynamics in their recent releases.

I wish I could say the same about Oasis, R.E.M. and You Am I; despite the fact that they have written some great songs on their albums this year, the over-compressed sound doesn't make me want to listen to them again in a hurry. No matter how good the songwriting is, bad production can still ruin it. Even the (usually reliable) Elvis Costello went a bit too loud on his recent effort, although not as bad as some of the others.

Metallica are where it gets ugly. Their latest album Death magnetic is considered by many to be one of the loudest albums ever released. Take a look at this wave-form analysis. It's a monstrosity of music production, and it also suffers from excessive digital distortion or clipping. Which is a real pity, because apparently the songwriting and performance on the album is some of their best in years.

Most people will get a headache after listening to Death magnetic for longer than 10 minutes. You only need to listen to Master of puppets or Metallica (the "Black album") to know that music doesn't have to be compressed to the max to be powerful. Quite the opposite really.

The good news about the botched production job on Death magnetic is that many fans have noticed it and complained. Metallica are a very popular band, and I hope the flood of complaints will lead to some sort of revolution against the loudness war.

It must stop!

Conclusion

Thanks for reading my ramblings. While I am definitely not the first to write about the loudness war and dynamic range compression, I hope you found this article informative. If you find that newer music isn't exciting you as much as it used to, I hope this helps to shed some light on why this may be the case.

Further reading (and viewing)

The Loudness War - A great YouTube clip demonstrating how compression is affecting music
Imperfect Sound Forever - The article that started it all for me
Loudness war - Informative Wikipedia article
Top 10 worst sounding records, 1997-present - Nick Southall's list of poorly mastered albums...
Top 10 best sounding albums, 1997-present - ...and his list of the best sounding albums
Turn Me Up! - Organisation aiming to bring dynamics back to music

9 comments:

  1. Damn fine post. I learn cool stuff every time I read!

    ReplyDelete
  2. Yeah, great post Jiggy. I wonder if there should be two masterings; one for CD's and another for radio?

    As I understand it, one of the benefits of compression is that it transmits more effectively over FM - those lovely peaks and troughs are just totally lost in this medium. There is a reasonable case to use compression there I guess..?

    But I'm with you - when I'm listening to my music I want as much detail as possible please!

    ReplyDelete
  3. @Matt: Your idea about two masterings is an interesting one, but I'm not sure how practical it would be. It would cost more to maintain two streams of production in manufacturing, and since record companies are always about the bottom line, it's unlikely that they would do such a thing.

    I don't understand why the radio stations can't simply compress the dynamic version of the CD before it goes to air. I'm not an expert in radio broadcasting technology, but surely any audio that is transmitted over FM already undergoes some sort of compression already? It seems that compressing the source is a bit redundant in this sense.

    ReplyDelete
    Replies
    1. Radio stations already aggressively apply compression and limiting to all songs before they are played to ensure all volume is the same. Albums that are brickwalled in the studio are further processed for broacast and end up sounding completely terrible, and even quieter than other songs.

      The trick to make your music stand out on the radio is actually for it to be dynamic: http://productionadvice.co.uk/loudness-means-nothing-on-the-radio/

      Delete
  4. Radio stations DO compress music prior to broadcast already, in fact they compress everything - it makes the signal easier to transmit further, louder. In the UK, Radio 1 is noticeably louder than Radio 3, in order to make people stop twiddling the dial when they're searching for a signal.

    ReplyDelete
  5. @Sick Mouthy: Thanks for the further info - it's great to see the author of the article which inspired this blog post commenting on it :-)

    ReplyDelete
  6. Dynamic music. Led Zeppelin. I own a two disc set called "BBC Sessions" and it is quite good! Thanks for the insight.

    Kiminakayakblog.blogspot.com

    ReplyDelete
  7. A handy resource for you to find better masterings.

    http://www.dr.loudness-war.info/

    ReplyDelete

Sing some harmonies here: