The end of the Loudness Wars

Discussion in 'Mixing and Mastering' started by statik, Sep 18, 2017.

  1. statik

    statik Audiosexual

    Joined:
    Jul 3, 2014
    Messages:
    1,516
    Likes Received:
    659
    Location:
    under your bed
    got this in my mailbox last week, thought i'd post it here, quite interesting

    SEPTEMBER 13th, 2017

    “The debilitating loudness war has finally been won.” – Bob Katz (Grammy winning mastering engineer)

    Loudness is a very misunderstood beast.

    All of us have had that moment – you know exactly what I’m talking about. Sitting in our studios, referencing a track we’ve slaved for hours and hours over, only to find that it doesn’t quite stack up to the professional tracks in your library. You think, “man, this is pretty well balanced and the mix sounds pretty tight – but for some reason, my track is WAY quieter than the reference.”

    A few years ago, it became common practice to suggest tons of steps to help generate a louder track, but today? Not so much. It turns out that loudness is NOT as big of a deal as we originally thought. In this post, we are going to dig deep into the common misunderstandings and misgivings about loudness that has permeated the internet’s forums, blogs and Youtube comments over the past few years.

    By the end of this article, you will gain a much greater sense of how loudness is measured, as well as how we can reference properly, generate a healthy amount of loudness and not get overly frustrated in the process.

    1. Death of the Loudness War

    The loudness war is basically over. Loudness has finally been defeated.

    Sure, people are still making very loud tracks – and it’s still a very valid goal for many producers out there to mix and master their music as loud as possible. Truth is, I do tons of mixing and mastering work every week, and to date, getting solicitations to make a mix as loud as possible is still, hands down, THE most popular request I get. However, with the boom of streaming services like YouTube, Spotify and Apple Music, for the average listener using streaming services, loudness simply doesn’t affect them anymore.

    So, why is the loudness war over?

    Well, all the streaming platforms have individually decided to match the loudness of the music on their platform to one singular loudness level. So, whether you make super loud dubstep or very light orchestral music, the average loudness for all listeners on a given platform will be exactly the same.

    The platforms in question have adopted using a target playback level with LUFS (Loudness Unit Full Scale). More on LUFS later, but basically here is the level your music will now be streaming at based on the platform you use…

    • -13 LUFS
    • -16 LUFS
    • -14 LUFS
    So, if you make a track that is -6 LUFS, Spotify is going to lower the gain of that track until it’s 8 LUFS quieter and streams at -14 LUFS. This essentially eliminates the need to master your music at extreme levels of loudness. At the end of the day, it all gets turned down anyway.

    The pros of this?

    Well, we can now have more dynamics in our music – which translates to added emotion and an enhanced musical journey for the listener.

    In fact, music that has a greater dynamic range and a lower LUFS value can actually now sound louder when streamed through one of these services, as opposed to a track that’s been over-limited with a small dynamic range. Here’s an image to help wrap your head around this concept…

    [​IMG]

    Pretty crazy, right? The world of music production is changing fast – and we stand by the fact that this is a very positive shift for engineers, producers and artists alike.

    2. Understanding loudness (LUFS vs. RMS)
    So first off, what the hell are LUFS?

    If you haven’t heard of RMS or Root Mean Square, it is basically a measurement in dBs of the average loudness or volume of a song. But with advanced metering technology, LUFS has evolved as a more accurate measurement of average loudness.

    The main difference between the two is actually really simple: RMS measures the total signal, and it takes in all frequencies. LUFS, however, takes out all outlier signals (both frequency outliers and peak dB outliers). This means that LUFS meters are ignoring the moments or instances in a track that has extremely short increases in volume (like an intense drum transient) or extremely loud sub information.

    And what does that mean for us? LUFS is a much more accurate measurement of perceived volume because our metering tools aren’t getting distracted by outliers. This becomes really important in referencing and matching your mix and master to a reference track.

    FREE VIDEO WORKSHOP: Check out this 60-min webinar covering 8 lesser-known steps to creating a louder mix.

    3. Proper referencing
    [​IMG]
    So, how does this help us with referencing?

    Before LUFS, when you brought in a reference track to your DAW you had to master your song and compare to the mastered reference track. This is fine for the final mastering stage, but can be difficult to use for when you’re referencing your pre-master mix.

    Because a mix is much quieter and more dynamic than a mastered waveform, it’s often hard to determine mix problems that are preventing your track from sounding like your reference. Obviously, the best way to fix this issue is to lower the volume of the mastered reference files to be more in line with your mix. Well guess what? LUFS is a tool that can accurately tell you how much you need to lower your reference tracks by.

    Unlike RMS or your ears, LUFS will accurately take into account the perceived level of the audio making gain matching simple and achievable. If we were to do this with just RMS, the meter can get tricked by those outliers we spoke about earlier.

    In order to do this, simply measure the LUFS value your track is hitting in a drop, verse or break section, and then measure the LUFS of the reference. Next, lower the reference to match in LUFS value – it should now be much easier to hear where your mix and the reference track are truly different in both frequency and dynamics.

    Here’s an example below using Logic’s built in Loudness Meter

    [​IMG]

    Taking this even further, we can also check what our songs are going to sound like when streaming through services like Spotify or Apple Music, because we can anticipate the final result. This is amazing stuff – we can then theoretically make mixing and mastering decisions knowing the final playback level our track. You also will get less intimidated when you reference a really loud song, and instead, can focus on the clarity and fullness of our mix.

    4. Don’t obsess!
    [​IMG]
    All this talk about loudness, and I almost forgot the most important point – do NOT get super stressed out by loudness. This can easily become an obsession and start to hinder your production because you, as the producer, are focused more on generating appropriate loudness levels than doing the things that really matter – like composing a beautiful melody, creating smooth transitions, or programming a memorable drum groove. With the rise in popularity of streaming services, loudness just simply shouldn’t be a large focus of your production.

    And look, I get it. I’m not immune to the loudness war. Hell, I’ve written tons of music where I almost mentally despised a track, simply because I couldn’t get the final result to be as loud as my reference. This sort of mentality has got to go.

    Every song is different and sometimes, you have just happened to choose sounds that weren’t capable of being as loud or as clear as a louder reference. To deal with this, simply use multiple reference tracks. Think of your multiple reference tracks as more of a sliding scale – shoot to have your mix & master somewhere in between your references, rather than just obsessing over one individual song. Oftentimes, the songs that we love are dramatically different from one another in terms of loudness and clarity.

    5. Okay I get it…but how loud should we be?
    With all this being said, there still is a respectable level of loudness your productions need to be, but the tools needed to achieve this are free or cheap or easily accessible, and something that doesn’t need to a burden in your music production.

    Now the first caveat here is something I touched upon above: production and sound choice are crucial in obtaining the end result you are looking for.

    Let’s pretend for a minute that your drop has big saw chords as a main element, but you’re referencing to a track with just a bassline and drums in the drop. It might literally be impossible to get your track as loud as the reference, even with the proper knowledge and tools at your fingertips. You can always get somewhere close, but I highly recommend not getting hyper focused on matching tracks to one another unless they are almost identical in sound design and style.

    So basically, all I’m saying is, don’t freak out.

    Get your tracks to a reasonable level of loudness (-10 to -7 LUFS). Even though streaming services cut off even quitter than this, we still want our productions to be relatively loud for when they aren’t on services that match LUFS levels (Soundcloud, downloads, etc.). Between -10 and -7 LUFS is where most commercial electronic music sits.

    Now this would most likely be at the drop or chorus of your track. As long as this part is clear and relatively loud, you can experiment with lowering other sections of your track in volume to bring back dynamic variations across the whole track.

    Encoding
    It’s also important to remember that encoding and converting your file to an MP3 will also add some gain and potential clipping to your song. This is why so many people have different opinions on what is the best ceiling value to set your limiter. Many engineers suggest setting your ceiling at -1dB, which will allow the encoding to do its thing without clipping. Worth noting: this does sacrifice some potential volume (for the sake of retaining the cleanest signal possible).

    Personally, I set my ceiling anywhere from -.6 to -.2 dB. This means there is a chance that when converted to an MP3 there might be some minimal clipping, but the amount is usually not enough for me to hear anything that I would deem as an ugly artifact or even digital distortion.



    6. Tips for Loudness/Mixing for Loudness
    Let’s talk about what we can do to increase loudness in subtle ways to make sure we hit the modest LUFS values we spoke about above. At this stage, we should be comfortable with referencing and understanding the limitations of our track. We should also know that we might need several reference tracks to get our production sitting where it belongs.

    Here is a list of the 4 best ways to increase the LUFS value of your track without introducing unpleasant distortion or the destruction of dynamics…

    1. Saturation throughout the production process
    2. Bus compression in layers
    3. Oxford Inflator
    4. Chaining Limiters
    Saturation Throughout
    [​IMG]

    One of the easiest ways to get your track to sound not only louder, but fuller too, is to saturate your sounds. I’ve done loads of tutorials and blog posts on saturation, watch a 47-minute live workshop I hosted about saturation or check out 20 of my favorite saturation plugins.

    In layman’s terms, saturation adds harmonic content to any sound you place it on. It essentially makes a sound richer and perceptually louder. By doing this throughout the mixing process in a subtle fashion (and maybe even in the mastering process too), we can get a much louder track without the need for over-limiting or over-compressing.

    My favorite saturation plugins for these techniques are the SoundToys Decapitator, the Fielding DSP Reviver and the built-in Ableton Saturator.

    Bus Compression
    Bus compression can be a powerful tool in making your mix feel more glued together. It makes all the different elements in your mix get slightly reigned in dynamically so they sit in a more “glued” or similar place in the mix (i.e. not having one sound be significantly louder than another).

    One method you can use to increase the perceived volume of your track is to chain together 2-3 different bus compressors in the mastering chain. Now the key, here, is to be subtle. Shoot for only 1 to 2 dBs of gain reduction on your compressors. The trick is to really is about having as little compression going on as possible but to still get a less dynamic final result. Try starting your attack setting at approximately 30ms and your release setting at approximately 50ms. A ratio of 2:1 or maybe as high as 4:1 should be fine as long we are achieving 1 to 2 dBs of gain reduction based on the threshold.

    Obviously, compressors vary by settings and style, but if you chain a couple of your favorite compressors together with similar settings to those above you can achieve a nice increase in perceived loudness in your master. My favorite compressors to use are the Waves SSL Comp, Fab Filter’s Pro-C2, the built-in Logic compressors, and the Plugin Alliance Vertigo VSC-2.

    Pro Tip
    Try using the sidechain feature of your compressors if possible. Sidechain is more than the pumping dance music effect we all know and love. Try using sidechain as a method to filter out of parts of the original signal, making the compressor react differently. The built in Logic compressor is really good at this. The idea here is to filter out the low-end signal with a high pass filter so that the compressor only reacts to the upper frequencies. The whole signal will still be compressed, but it will be much less dramatic, only reacting to the mid and upper frequencies (instead of the loud sub information). Check the image below…
    [​IMG]

    Oxford Inflator
    A secret weapon many producers are using is the Oxford Inflator. I won’t get into details about what it is (read the manual if you’re interested), but it basically increases the perceived volume of a sound without using saturation, compression, EQ, or limiting. Basically, some crazy coding voodoo magic

    [​IMG]

    You have to just trust me here and give it a shot. Simply turning the effect knob up makes almost any sound you run through it sound fuller and more alive. Definitely worth every cent of the asking price.

    Chaining Limiters
    Very similar to chaining bus compressors, you can chain limiters. The idea here, is that several limiters can share the work of what just one limiter is doing. The Fab Filter Pro-L limiter is far and away my favorite limiter, and it’s a super transparent, go-to professional limiter. I’m sure I’ll get some haters for this, but – you basically don’t even need hardware limiting if you use this limiter.

    [​IMG]

    If you can’t get to where you want to be with one limiter, try putting two or three together to slowly limit the sound instead of dramatically lowering the peaks with one limiter. A big trick here is to set the output of the first limiter to something much lower than you’d want for the end result. Limit into a ceiling of -2 to -1 dB first, and then into your final ceiling of -.5 to -.2dB. This way your first limiter is focused on limiting the signal a lot less aggressively and your next limiter is taking care of the final level increase.

    FREE VIDEO WORKSHOP: Still struggling with loudness in your productions? Here is 60-minute workshop covering 8 lesser-known steps to creating a louder mix.

    Summary
    At the end of the day, loudness will always be a necessary component to creating commercially relevant music. But with the influx in popularity of streaming services, it’s safe to say that the loudness war is officially over.

    Not only that, but you will make your music audibly worse if you end of focusing on generating a loud track instead of focusing on achieving a relatively loud mix with proper dynamics.

    So, next time you’re entering the final phases of a track, perhaps the question you should really be asking yourself – how much quieter can I get away with making this track?
     
    • Like x 10
    • Love it! x 4
    • Useful x 3
    • Interesting x 2
    • Creative x 1
    • List
  2.  
  3. digitaldragon

    digitaldragon Audiosexual

    Joined:
    Apr 27, 2016
    Messages:
    1,256
    Likes Received:
    1,063
    A nice read, and some good tips. I may have to give Inflator a test run to see what's up with it.
    The timing of this is nice as it's what I'm currently working on.
     
  4. Von_Steyr

    Von_Steyr Guest

    Great post.
    Too bad, this means the end of career for @foster911 :)
     
    • Funny Funny x 10
    • Like Like x 1
    • List
  5. mild pump milk

    mild pump milk Russian Milk Drunkard

    Joined:
    Mar 29, 2014
    Messages:
    2,683
    Likes Received:
    2,262
    Location:
    Russia
    Loudness war...
    So, why do developers still create many brand-new hi-tech limiters and clippers with a lot of adjustments, lookahead, ISP, oversampling etc. if my mixed tracks without them are louder than those -14 dBFS, not oversquashed and peaking much lower than 0dB etc.?
    So limiters and clippers are garbage, because these days CD are almost useless, and it is the time of Internet distribution and special loudness standards...
    So we don't need more "extremely transparent but very loud limiters and clippers" anymore, because our reference IS classical music, enigmatic, ambient, and it doesn't matter whether you produce and record hard metal, dubstep, drumandbass, trap, hip-hop, pop, EDM or porngrind. Otherwise with your limiters and clippers your overaggressive track will be oversquashed, flat but quiter than Beethoven's romantic piano masterpiece. So Bach, Vivaldi and others will be LOUDER. Imagine these guys louder than you?)))
    Uninstall all your limiters and clippers, let's download underquietest gain-downers, volume lowers, quietmakers and antinormalizers.
    It all started with this...First limiters, clippers...then they will start to fight with compression, oversquashness to be as dynamic as classical music...then sidechain wars...and we will be so anticool, quiet, not so groovy... Huh
    When you start to think about this and think about future, sometimes it makes me feel confused...humour or really war. To be as Beethoven..
     
  6. No Avenger

    No Avenger Moderator Staff Member

    Joined:
    Jul 19, 2017
    Messages:
    8,915
    Likes Received:
    6,112
    Location:
    Europe
    Oh, oh,

    tons of steps? Really? Never tried it that way.

    That would be totally distorted RMS of 0dBFS. Never tried that either (but urgently have to, could be an interesting effect).

    YouTube is measuring every single piece of music?? RESPECT!!

    Ehm,...nope!

    Most positive effect is for the listeners!

    What a great idea, I'm so thankful.

    Ehm, nope, this doesn't take too much time.

    Ha, now I got it, I never had that. Call me selfish but whenever I could determine the loudness of a song, I stopped when I recognized that it's going to lose too much punch because I was so eager to get it in.

    Why should I do this??? That's no reference.

    So that's too loud to sound good. Thanks for the hint.

    So, you set it too high because there are amps that respond to that.

    I'm gonna stop here altough I could go on and on but I'm getting bored and have to do something useful like, ehm, anything else.
     
    Last edited: Sep 18, 2017
  7. Who Me

    Who Me Producer

    Joined:
    Dec 21, 2015
    Messages:
    169
    Likes Received:
    84
    What these techniques/recommendations don't take into consideration is not all music being listened to is being played from a streaming site.

    People still buy music to own (mp3, wav, flac, cd, vinyl etc...). I really can't see there being a massive shift to mastering everything much quieter in those release formats in future just to match what the streaming sites are doing.
     
  8. Von_Steyr

    Von_Steyr Guest

    You seem to be new here, though i suspect you dont know what the nasty rating is for do you buddy? If you dont understand internal humor, cool, but with such an attitude you aint gonna get far in this community.

    EDIT:
    All solved, wasnt intentional, no harm feelings.
     
    Last edited by a moderator: Sep 18, 2017
    • Agree Agree x 1
    • Useful Useful x 1
    • List
  9. dragonhill

    dragonhill Guest

    some think it's this:

    1986? wow
     
  10. No Avenger

    No Avenger Moderator Staff Member

    Joined:
    Jul 19, 2017
    Messages:
    8,915
    Likes Received:
    6,112
    Location:
    Europe
    Yep, you should really test it because it could produce a hell lot of aliasing
     
    • Agree Agree x 1
    • Interesting Interesting x 1
    • List
  11. foster911

    foster911 Guest

    @statik do you think music industry's major issue is loudness? I don't think so. It's just a slight hitch or a brain teaser.:bleh:

    The main bogey is the good music and the lack thereof. Let's first fashion a good one then be sure I'll submit all ones to yourself for mastering.:rofl:
     
    Last edited by a moderator: Sep 18, 2017
    • Interesting Interesting x 1
    • List
  12. saltwater

    saltwater Guest

    great read, thanks for sharing.
    never heard about this Inflator, interesting!
     
  13. digitaldragon

    digitaldragon Audiosexual

    Joined:
    Apr 27, 2016
    Messages:
    1,256
    Likes Received:
    1,063
    Thanks for the tip, I'll keep an "ear" out for that while testing. Wonder if it has oversampling which should help.
    I don't know exactly what it does, so will have to do some learning first as well.

    As for saturation to add harmonics and such, I think I'm doing that as a by-product of running through the types of compressors, strips, and eq's I use as well as an outboard tube preamp.
     
  14. Olaf

    Olaf Platinum Record

    Joined:
    Jun 5, 2011
    Messages:
    549
    Likes Received:
    231
    Who is the sender? A Nigerian prince?

    Fake news? This quote is not from 2017, but from 2013 and is actually "The loudness war has finally been won! iTunes Radio solved it!" to advertise a mastering service or the like for exactly this iTunes Radio.

    Your source is obviously "Hyperbits" aka some Serik nobody ever heard of. You can follow him on Instagram if you like, but maybe he is not the right person to stop the loudness war... :unsure:

    -> Be sure to check your sources. :guru:
     
    • Interesting Interesting x 2
    • List
  15. Backtired

    Backtired Audiosexual

    Joined:
    Jan 15, 2016
    Messages:
    958
    Likes Received:
    644
    Shouldn't you have at least two versions of your tracks?
    One for "online streaming", and one for those who wanna own it?
     
  16. No Avenger

    No Avenger Moderator Staff Member

    Joined:
    Jul 19, 2017
    Messages:
    8,915
    Likes Received:
    6,112
    Location:
    Europe
    I found these in another forum made with a single sine wave at 1kHz


    inflator%201.JPG


    inflator%203.JPG


    inflator%202.JPG
     
    Last edited: Sep 18, 2017
  17. SineWave

    SineWave Audiosexual

    Joined:
    Sep 4, 2011
    Messages:
    4,273
    Likes Received:
    3,371
    Location:
    Where the sun doesn't shine.
    I used to use Inflator years ago, but never more than 50% of "effect". I think this thing has been kinda obsolete for years now. It doesn't have oversampling and all it does is add harmonics [saturation] AFAIK. It's better to use Klanghelm, Soundtoys and Fabfilter Saturn these days IMO. But what everybody seems to forget are Voxengo saturators. They are top of the line! :wink:
     
  18. digitaldragon

    digitaldragon Audiosexual

    Joined:
    Apr 27, 2016
    Messages:
    1,256
    Likes Received:
    1,063
    @No Avenger, thanks for the research. That's crazy, the amount of additional stuff being added to the original sound. Seems like some type of harmonic excitation from the looks of the spectrum analyzers.
     
  19. No Avenger

    No Avenger Moderator Staff Member

    Joined:
    Jul 19, 2017
    Messages:
    8,915
    Likes Received:
    6,112
    Location:
    Europe
    @digitaldragon, yeah, the big ones are odd harmonics added on purpose and the hundreds of smalls are (unwanted) aliasing. The interresting thing is full effect and full curve result in least aliasing (as long as the input doesn't cross 0dB) and loudest harmonics. But last one could be too much so you may try it as a send effect.
     
    • Agree Agree x 1
    • Useful Useful x 1
    • List
  20. Von_Steyr

    Von_Steyr Guest

    The music industry released a statement to the world. " We are sorry for this meaningless war, it was the money (coke + tits) that made us do it.
     
  21. statik

    statik Audiosexual

    Joined:
    Jul 3, 2014
    Messages:
    1,516
    Likes Received:
    659
    Location:
    under your bed
    these are not my thought, it's something i received in my mailbox. imo one of the main problems is the difficulty in getting things published, it's all friendpolitics and having to know the right people, which explains the amount of crap that's being published
    you seem to have answered this question yourself so why ask?

    i didnt say it was the absolute truth, only that it was an interesting read. sept 13th was the date in the article put there as publishing date not date of that quote. so you might wanna check your assumptions, i will however say i'm glad to see people not taking 'published' articles as facts and using their brains.

    i could go into other comments but i am far to busy getting my work at the now mandatory youtube. spotify and itunes levels
     
Loading...
Similar Threads - Loudness Wars Forum Date
Check your loudness penalty for streaming on your master Mixing and Mastering Jan 9, 2024
BUTE Loudness Normaliser | This Plugin is MAGIC for Audio Broadcast Standards Software Reviews and Tutorials Jun 14, 2023
Increase loudness without increase the peak level Mixing and Mastering Jun 10, 2023
Adjust loudness and compression in bulk? Working with Sound Jun 2, 2023
Best saturation plugin for the sake of perceived loudness Mixing and Mastering May 16, 2023
Loading...