It's not just about volume. Tracks mastered with a ton of compression trick your brain into sounding louder than they really are, which is great for a song or two (or if you want pay people to pay attention to your tv commercial), but for listening to a whole album it wears you out. If you have an album you love but somehow never make it all the way through this is probably why. The perception of full volume gets your lizard brain aroused, which is great if you're in the club, but not if you're in the mood to listen to the first three Led Zeppelin records in a row.
On the other hand, older recordings with more dynamic range might sound thin at low volume, but are much richer at higher volumes (you can hear the individual instruments better and feel the space in the sound). If you try comparing older and newer masterings at a good volume the newer mastering usually sounds kinda mushy.
Not only it wears you out, but it can be dangerous, especially in clubs. The longer you listen to loud music (can be just a few minutes!), the more damage is done. Having music with a wide dynamic range helps to have an overall lower volume and preserve ears.
In Sweden, night clubs have a different time period for sound pressure exposure limits than concerts[0] (15 minutes instead of 60), partially due to the difference in dynamic range.
Occasionally that's true, very often it's not. Looking at your second link, there's only one album that has been substantively remastered. (And that better mastering could have been distributed on CD or through iTunes, presumably they just didn't bother to do so.)
For practical listening, I actually prefer modern brickwall mastering techniques to more traditional mastering with a high dynamic range, for one reason: what the author sees as "hijacking the volume control from the listener" I would consider the opposite.
With a high dynamic range, a headphones listener may feel the need to adjust the volume several times in a song to boost the clarity of softer sections or to make louder sections more comfortable to the ears, depending on the listening environment. With a "loud," low dynamic range, however, the listener need only adjust the volume once, as the whole track is roughly the same volume. In other words, the listener is in control of the volume, rather than the engineer.
Just give me the option. When I'm in the car, I like the "wall of sound" since it's a noisy environment, so I have a head unit with a built-in compressor I can control. When I'm at home, I prefer the dynamic range. The problem comes when the engineer takes that option away from me (quite often) clipping and distorting everything to the nth degree. Like the article mentions, this creates ear fatigue and I end up, eventually, just not wanting to listen to the CD at all which colors my perception of the band. So I'm less likely to go to their concerts and give them money. That's just me though.
As an analogy, how about if all text was served to you as .png files where the person writing the text has guessed what size text you want?
This is vaguely equivalent - it is difficult to go from image->text, just as it is difficult (or impossible) to go from low dynamic range -> high dynamic range. The reverse direction is cheap (and possible).
In fact, some compression codecs even have well defined compression curves to use - so your codec isn't just "bitstream in, pcm out" it is "bitstream+listening environment in, pcm out".
Isn't it more like the artist wrote his text using all kinds of font sizes that require the user to zoom in and out all the time to read it comfortably? Compression would mean that the range of font sizes is compressed so that one zoom level works reasonably well for the whole text.
> boost the clarity of softer sections or to make louder sections more comfortable to the ears
You can create these effects during playback with a compressor, which are available in many players[1]. Knowledge about compressors isn't as widely known as it should be; it's an important feature that would help a lot of people, especially anyone dealing with hearing loss.
> headphones
Maybe try better headphones (or a different type, such as enclosed headphones with good room attenuation)? Many cheap[2] headphones have terrible response to different volumes; quieter parts may be overly attenuated, for example. Cheap/bad headphones also tend to have a highly variable frequency response, which can sometimes sound like volume problems. If this is the case, spending some time with a highly configurable EQ might help (possibly in addition to a compressor).
[1] also available in some audio systems (e.g. jack) and as a feature in some drivers/soundcards
[2] and some "name brand", such as older Bose, anything Beats or other brands sold as fashion statements.
edit: re-added sentence that died during a bad copy/paste
edit2: Appendix for anybody that isn't familiar with dynamic range compression.
> need to adjust the volume several times
This is literally what a basic compressor is automating[3] for you. It turns down the output volume when the input level is over a threshold. There are a lot of (often configurable) details like how much to reduce the volume as the input level increases past the threshold, how quickly[4] it responds to a loud transient, and how quickly it returns to normal when the input becomes quiet again. Really nice compressors even smooth[5] the changes near the threshold so the whole effect is less noticeable.
I knew about them and the copy of Android I had installed on my phone included compression.
Like most things in that OS it's very buggy, one day it just dumped full volume noise into my headphones and that really hurt. After that I hardly trust anything in Android to handle audio and don't like plugging headphones into my phone.
There is value to constant volume, but I think a lot of the loudness war came from a time when there implementing a compressor on the user's side was expensive. It's trivially cheap now, basically every device you own that can produce audio now is capable of implementing a decent digital compressor with no additional hardware costs and about 5 minutes of an software engineer's time. If this wasn't hard when the loudness war set in, we'd probably just have a checkbox that defaults to 'on' in all these devises instead of the master being messed up. In other words, the timing of cheap, low power audio DSPs missed the boat by a couple years.
> That's like saying you want low dynamic range on all of your photographs so you can see every detail.
Which is actually a super common problem in amateur and even some pro photography today. "High dynamic range" (HDR) lets you capture an image with a greater dynamic range than the sensor itself can detect, usually by stacking a few photos taken with different settings. It preserves details that would otherwise be lost in shadows and highlights.
In other words, it's compression for images. And it gets over-applied really often, leaving hideous painful to look at photos. They end up too flat and in your face.
There are two parts to this. First is capturing a HDR image, either by stacking photos or using a sensor with a good dynamic range (and capturing in raw). This is a good thing as you are capturing more information.
The compression you talk about is called tone-mapping and is used to make it possible to view HDR images on a non-HDR display. It can either be (over) done intentionally or it can just be an unfortunate consequence of HDR displays not being mainstream yet. Fortunately HDR is becoming a thing now for TV’s so computers are sure to follow.
Now, viewing HDR material on an actual HDR display, that is something different and can look absolutely stunning.
> The compression you talk about is called tone-mapping and is used to make it possible to view HDR images on a non-HDR display.
Yes, and it's exactly how audio compression works. You have a fixed dynamic range and in input signal that doesn't fit in that range. So you attenuate the values near the endpoints (shadows and highlights in the case of imagery) to fit within the range while still leaving values in the midrange with their original spread.
> Now, viewing HDR material on an actual HDR display, that is something different and can look absolutely stunning.
> Yes, and it's exactly how audio compression works. You have a fixed dynamic range and in input signal that doesn't fit in that range.
Not exactly. Audio compression is usually used to compress to a much smaller dynamic range than the output format or playback equipment can support. Some recordings compress to within just a few dB. With HDR pictures it's usually because the display is physically incapable of reproducing the image without compression.
>With a high dynamic range, a headphones listener may feel the need to adjust the volume several times in a song to boost the clarity of softer sections or to make louder sections more comfortable to the ears, depending on the listening environment.
Softer sections are created and meant to sound softer though. Boosting the volume only in certain sections is more like making your own mix...
Listening conditions make a big difference. I have a very dynamic version of Queen's Bohemian Rhapsody (source: MFSL CD) that sounds absolutely great at home but is unlistenable in the car. The average volume is just too far below the peaks, so you strain to hear most of the song with a noisy background.
I have an even simpler wish. All I want is to comfortably listen to my audiobook when running without going deaf when Strava comes on to give me pacenotes. My kingdom for per-app volume control!
That's one benefit of podcasts. They often have lower quality but louder audio so I don't have to pump up the volume when I'm out and about.
The effect you want - roughly the same volume through a whole track, can easily be achieved with tasteful use of compression. Some genres hardly need any to start with.
There are a number of albums that were mastered quite well that have a good dynamic range, yet still sound good in my car, on headphones, or at home on the big stereo.
The heavily compressed stuff just sounds bad all around.
softer sections are an integral part of the music. what you're asking is like forcing movies to have all scenes with a constant luminosity no matter if it's day or night.
softer sections are actually an excellent way to make other sections bang like crazy, just because of contrast. the brickwall mastering techniques are an invitation to neglect composition techniques and overall creativity. the drumcode label comes to mind…
Do you also want explosions in movies to sound just as loud as whispers? Quiet parts of songs are meant to sound quieter, there's even special notation in sheet music for this.
This is pretty funny to me because I mostly listen to (and produce) electronic music.. and there are pretty much no rules when it comes to electronic music and loudness. Stupidly loud music can actually sound pretty dang good [0][1]. The momentary RMS in some Moody Good tracks can actually hit _above_ 0dBFS.
If you have the right source material, you can brickwall the hell out of tracks and not notice the distortion.. or perhaps the distortion will even add pleasant artifacts. One of the more prominent issues with making things stupid loud is intermodulation distortion, but that really only becomes noticeable when you have pure tones or vocals being mashed into the limiter. If the source material is already distorted (think screechy dubstep synths), then it probably don't matter.
But yeah, when you're dealing with more traditional kinds of music, which often times involves vocals or a lot more subtlety to the timbre of the instruments, brickwalling is probably not the best call. It seems that the Search and Destroy "remaster" sounding terribly distorted was intentional.. but IMO it's not very listenable nor does the distortion really bring the grungy character than I think they were going for. It just sounds bad.
I (only half jokingly) think (music like) that Bassnectar track Into The Sun is Lorin Ashton's attempt at hitting The Brown Note when played on a big rig sound system. It's just constant rolling thunder, break down around 2 mins in, followed by more rolling thunder. Sounds like the melody is playing on a different sound system half a kilometer away at the same festival.
Ah, music is weird because even the stuff I like is fun to criticise.
On that note, have you come across Ishkur's Guide to Electronic Music?[1] It hasn't been updated in a long time, but still highly amusing.
Yeah. In fact, if you don't have a LOUDNESS WARS mix on some dance music, it sounds wrong, like rock'n'roll without distortion on the guitar. At that point you need to do it to individual instruments, to achieve the effect people now expect to hear.
>The momentary RMS in some Moody Good tracks can actually hit _above_ 0dBFS.
It's about the dynamics (or lack thereof), not just about the distortion (if it's not clipping to just white noise 0dBFS is not the DAW limit anyways).
Theses mastering techniques have their legitimate use ! Some notes :
* it's a step in musical production where having experience, skills and contact with the artist matters. Not all compressor and limiter are created equal and the default value you use in your media player may not sound as good as what an audio engineer might have done..
* Not everyone have good hardware and a good environment to listen to high dynamic range music like thoses listening to classical music / jazz / Philip Glass, so theses business decision to increase volume for the market made sense at that time I think. Audio engineer simply took profit of having a technically better medium (CD) to make audio sound better (from what I've read theses techniques did not work well on vinyl)
* Loudness wars didn't have an effect on old records since as one can see in this article, we could find the old dynamic ones (and so we actually have the choice of listening to the old untouched record, or the new compressed-for-the-market record, and that is a good thing !)
* Theses music stats (mean RMS, peak RMS, max mean RMS) look at instantaneous dynamic, but a look at the overall dynamic of a song is also very important ! A good article on this topic stating that songs did not lose overall dynamic range that much : ['Dynamic Range' & The Loudness War, 2011] http://www.soundonsound.com/sound-advice/dynamic-range-loudn...
Tangential, but for the longest time I couldn't figure out why VLC always played music / DVDs at such low volume. Setting the system volume to max and overdriving VLC's volume slider was the only way I could actually hear the soft parts.
Recently I found out about the volume compressor, which with a single check box does exactly the right thing. I asked myself "why the heck isn't this box checked by default?" I think the answer is with audio purists wanting to stem the loudness war.
When reading about CD mastering maxing out the volume, It seems like it is the right decision. Most people do want the loudest setting, no mess with the EQ, compressors, etc. Only a tiny population wants to preserve the fidelity of the amplitude.
I think the reason it's not on by default is that so much music is already compressed to insane levels that sounds terrible - double-y compressing things doubles the terribleness.
I love music, but I wouldn't call myself an audiophile (heck, I'm listening out of my lappy's default sound card with cheap headphones) but in this case I totally agree with the audiophile's term "fatigue". Music that is over-compressed doesn't sound bad to my untrained ear, but it's just fatiguing after a while. The silence that finally comes is a pleasure!
But having it double-compress by default is like hearing those people who want to play there car stereo louder than their system can handle and it distorts like crazy: they are obviously still enjoying it - but it's not the song the artists recorded. (Ok, maybe an exaggeration... but I still don't think we should encourage it!)
Movies generally have a wide dynamic range so that loud action scenes are very loud. You don't want your normal dialogue to be at the same volume as your explosions, unless you're watching at home with the volume down.
Much to the chagrin of people like my wife, who, during scenes with just dialog will remark, "I can't hear it, turn it up!" and then when the action starts, "Too loud!" I end up adjusting the volume up and down over and over throughout a 2-1/2 hour film. As much as I appreciate good dynamic range in recordings, I wish my sound system had a "constant loudness" function that I could just engage and put the remote down.
That function exists generally in audio, and is called dynamic compression or just compression. I've seen some TVs with that functionality and I know Apple TV can do it, so I would guess you might have that option lurking somewhere in your system. When you turn that on, you're taking a wide dynamic range and squishing it, which essentially just pulls up quieter content, so you have to consider how pure and accurate you like your content playback to be when doing that. Well-crafted films in particular occasionally play with audio level for effect (Interstellar being a dramatic example of that), so you might subtly suck a bit of air from a director's intention which might matter to you.
If you have a nice 5.1 setup, try boosting your center channel, too. That might help.
That is almost definitely related to the center channel being either too low, not existing, or some element of the sound setup mixing the center into the left and right.
I have the same problem and I have not fixed it yet because I have yet to purchase a sound system with a center channel.
Most good quality receivers sold in the past decade have a volume limit setting. It also sounds like you might not have the gain set right on your center channel. Two things to check which might help!
> I wish my sound system had a "constant loudness" function that I could just engage and put the remote down.
Any half-decent A/V receiver has this option. My Marantz receiver for example can scale the compression automatically depending on volume. The more you turn down the volume, the more it will compress, if you turn it up to max there will be no compression.
Somehow I suspect that the same people who like overcompressed music (to drown out everything else) love a big dynamic range in movies (gut-shaking explosions when dialog is at comfortable listening volume) and that those who like music with silent parts really would not mind being able to leave a cinema without a minor tinnitus. It's a confusing world.
I have this problem too. I'm assuming it's because I don't have a center channel, just to big stereo speakers. Even with the receiver set to know that there isn't a center channel, it doesn't seem able to compensate. I guess movies are just mastered assuming 5.1 now and if you only rock stereo, you get compromised sound.
> unless you're watching at home with the volume down.
Is this not the most common scenario, though?
Super high dynamic range is great in a cinema but terrible if you're watching at home and want to hear the dialogue but also not wake the neighbourhood when it flicks to an action scene.
My upstair neighbour stopped banging on the floor when I started using reclock's dynamic audio compressor in MPC-HC.
It sometimes end up slightly unnatural, but it beats having to constantly change the volume throughout a movie - something I didn't do too well anyway, evidently.
I still prefer the media to have better dynamics and have control over how much compression I want to apply.
Since Youtube and most streaming services started to automatically balance tracks based on their average perceived loudness (not all of them use the same metric, but the purpose is the same), loudness war is almost dead. If you brickwall your song, it will not be played louder than competition anymore.
Fortunately many audio providers now have been fighting back against this. For instance YouTube will punish video uploads that are louder than -13 LUFS by attenuating the level. This will provide a somewhat level playing field and encourage people to upload with a reasonable dynamic range.
Ideally compression could be left for the end device. CyanogenMod has for a long time had compression control in settings. I can understand why some compression is useful, but it is a shame for the streaming provider to do this since information gets lost in compression.
This is something I have battled with for a long time, personally I master to make the music sound good, sometimes that is loud (new dance music) sometimes not so much (old African recordings remasters).
Coupled with the data from this page [1] there is no point in going too loud anyways, that's why you have gain / volume control. I'm not sure how I feel about streaming services implementing extra processing tbh. Spotify is the worst culprit adding limiting which can significantly change the sound of a recording.
I just wish other engineers would have more pragmatism in this industry, way too much overcooked and distorted music around.
Mastering is in a way comparable to capitalism. You can certainly try and be nice about it but pushing harder will usually get you further, because most people care more about one aspect than all the others aspects combined. In capitalism it is getting a great product for a great price. In mastering, it is loudness.
Loudness is a bastard. There is a reason, why all the pros are usually very, very careful about level matching when doing any sort of audio comparisons. Even when you know that louder can easily fool you into thinking something is better (which most listeners don't), you're still susceptible, if you don't counter act it. Wanna convince a recording artist in the studio it's great? Turn up those big speakers. Instant gratification.
When it comes to music consumption I like to think this is not really a problem: The sound of compression and distortion is the sound of current music and there is nothing inherently bad about it. Older generations will tend to oppose any new musical trend for various reasons, which all end up being subjective. The younger generations that grow up on this new sound do not care about brickwall limiting, because there is nothing to fucking care about.
Music production has been and forever will a mix of mostly people copying other people and flowing with the stylistic currents while adding a little something themselves. Sometimes something radical will happen. Mostly not. If you wanna stay relevant you go with the former and keep reaching for the later. Pretty much the same, as with coding or design.
There is research that shows that on a casual listen, people believe louder music sounds better. Similarly, people tend to find brighter pictures more appealing at a glance, even though both of these can be fatiguing and cause discomfort. That's why stereo stores (in so much as they still exist) crank up the music to very loud volumes and all the TVs at Best Buy are grossly mis-calibrated.
Unfortunately, most musicians are just hoping you will notice their song when it comes up on the radio, pandora, a friends iPhone and so everyone is incentivized to crank their song to the max.
The thing that bums me out is that there were really good records released in the 2000's that are mastered terribly, and we may never hear a better version. It's one thing when a stooges album is fucked up on re-release, I can always grab the original, there may never be another version of "Is This It".
The thing that bums me out is that there were really good records released in the 2000's that are mastered terribly, and we may never hear a better version.
There's an excellent example of this occurring in the recent remastered re-release of Oasis's Be Here Now which had horribly thick and overdone production but has now been opened up somewhat. It's still loud, but it's Oasis after all.
I don't have an example of the before, but the 'after' is good: https://www.youtube.com/watch?v=jyJU2136ym4 .. interestingly, this is also an example of remastering and re-editing an old music video too.
Because it's consumerism. You want quick and deep stimulus. Not subtlety. When you're "educated" [1] you start to look the other way around. I remember understanding the value of a good DAC chip on a smartphone, even with a crappy mp3, I could rediscover songs that I've listened to a thousand time on bad DACs.
Same goes for photography, poetry etc etc.. mass market doesn't mesh well with that.
> Unfortunately, most musicians are just hoping you will notice their song when it comes up on the radio, pandora, a friends iPhone and so everyone is incentivized to crank their song to the max.
Most people also just aren't listening through good equipment that emphasizes good dynamics. Tons of people are happy listening to compressed songs through the earbuds they got for free with their phone.
I found myself pondering this in relation to the resurgence of LPs.
I wonder if the lower dynamic range of LPs restricts music released on them from being overly loud, and thus not as fatiguing to listen to as more modern formats.
It is technically the case that vinyls have less frequency response (basically a low pass filter) than good (really effectively perfect) reproduction formats like CD. That certainly means that the maximum loudness value is lower, but dynamic range is not required for loudness, gain and frequency response are.
If anything, having less dynamic range would encourage people to make louder records.
A CD is limited to 22khz, whereas vinyl is analogue and doesn't have this limit, digital files obviously can go up to ~96khz due to the 192khz sample rate achievable by some sound cards.
> A CD is limited to 22khz, whereas vinyl is analogue and doesn't have this limit, digital files obviously can go up to ~96khz due to the 192khz sample rate achievable by some sound cards.
The only reason to sample at 96/192k is to make cheaper analog filter hardware. While technically vinyl can reproduce frequencies in excess of 50kHz, it has limits on amplitudes at high (audible) frequencies because of the physics of the needle and groove, especially on less-than-excellent cartridges. Nobody wants to listen to something with that much energy in the highest audible frequencies (16-20kHz) anyway, so it's a moot point.
There is absolutely no point in representing frequencies in excess of 22.05kHz. Truly exceptional (never documented) human ears might technically be capable of faintly hearing 23 or 24kHz tones, but that ability is likely to manifest and deteriorate over the course of (at most) a handful of contiguous years in their life and never return.
There is no difference between discrete signals sampled at 44.1kHz, and continuous signals, that has any bearing on human hearing. In addition, 14 bits is about enough to represent all discernible dynamic range in human hearing, 16 bits is more than enough. Vinyl has considerably less dynamic range than this.
Any vinyl on any record player has poor performance for reproducing audible signals when compared to an about-average CD player. LP vinyl is an excellent final form of the phonograph, but a mere intermediate to poor form in all of audio reproduction, especially considering that CD is essentially perfect.
> Truly exceptional (never documented) human ears might technically be capable of faintly hearing 23 or 24kHz tones, but that ability is likely to manifest and deteriorate over the course of (at most) a handful of contiguous years in their life and never return.
When I was in my twenties, I could easily identify sounds in the 23-26k range (tested using professional gear more than once). When I was younger it was quite painful to hear as my ears were so sensitive to it (as a young teen, I would wince in pain when a monitor was left on without an input signal even in a room of people yelling noisily at each other). Even now in my late 30s I can still readily hear in the low 20k hz easily.
We could have by having EBU R128/ITU-R BS.1770 published before CD, and having those metadata built in to the CD standard, and perhaps requiring it.
The loudness war started before CD, but if we had this stuff back then it might have been possible to avoid these problems. As a consumer, I add BS.1770-based replaygain tags to all of my music, and set a constant gain offset for all applications which don't have loudness normalization.
I'm most annoyed by albums which have huge loudness gradients for completely different movements or independent tracks, it forces me to manually remove album replaygain tags in favour of the track values.
Notable also, this page makes no mention of loudness monitoring, which I think is outrageous. To complain about this problem without mentioning the existing solution is pretty bogus.
I'm a self-described audioweenie who has a dedicated listening room with a decent setup. It does make me sad that much new music is mastered so loud because information is lost when you "brickwall master". For older music, a lot of audiophiles will seek out original non-remastered releases of music made before the "Loudness Wars" (early 90s and before) where, as I understand it, it was generally considered taboo to have a "digital over" in a mastering. However some "Loud" masterings still sound quite good. It depends a lot on the type of music and how the compression was applied. I mean, someone overdriving a tube guitar amp into soft clipping is essentially compressing the sound in the analogue domain - that's what makes guitars sound so awesome.
I wish to make one main observation about the vinyl resurgence. Vinyl (which I enjoy, mostly for nostalgia because I'm old enough to remember when it was the main format) is on the rise for the wrong reasons. One reason is that it has become trendy - and I have no problem with this, but it's a real thing. Another more frustrating reason is the perception that vinyl masters cannot have the same amount of compression as digital masters, hence the perception is that vinyl version of a modern master will less compressed. Many, many audiophiles believe this. However I can tell you that the vast majority of modern vinyl releases are the same exact mastering as the digital version. The digital tracks have already been "squashed" and that mastering is fed to the cutting head after applying the RIAA curve.
However, the "dynamic range database" (results from a piece of free software that applies an algorithm to digital music and assigns a number related to the ratio between peak and RMS energy in the music) will regularly indicate vinyl versions of music (recorded and digitized by someone on their home setup) with more dynamic range. The problem is that this "extra" dynamic range arises from the inability to reproduce square waves (those flat-topped 0 dBFS regions in the article) in the analogue domain... there are overshoots that "add" peak energy that didn't exist on the squashed digital master. So you have people who think squashed music that has gone through all of the processing that is required to make a vinyl record magically comes out on the other end with more dynamic range, sounding better, when it's added a bunch of additional distortion. It's part of what makes the audio hobby so much.... fun.
On the other hand, older recordings with more dynamic range might sound thin at low volume, but are much richer at higher volumes (you can hear the individual instruments better and feel the space in the sound). If you try comparing older and newer masterings at a good volume the newer mastering usually sounds kinda mushy.