End the loudness war - back to the roots of full range dynamics and the best music quality

[timaeus222]

The only difference between the before and after tracks that I showed was literally the Master compression. That's it. Don't assume.

[APZX]

The be more clear with your descriptions. It states VERY CLEARLY on the description "Mixing/Mastering". Those two words are not synonymous. I'm not assuming anything other than what your description states.

[timaeus222]

I know they're not synonymous. Clearly, mastering is usually considered a separate process, and it certainly is not merely mixing on the Master track, but I never said I did both. In that case it was Master track compression.

[Possessed]

You need compression on the individual tracks in a recording or it will sound like an indistinct mess.  Especially guitar and bass, drums also need light compression.  Listen to 70's rock music and you can tell even then that the individual tracks have compression on them by how smooth the guitars sound.

Where people mess it up is by then applying an unnecessary and obnoxious extra layer of compression to the mixdown afterward to make it louder.

 

Of course this is only my opinion.

[FastnBulbous]

It is very easy to get carried away with compression.  To avoid this, I try to make a habit of controlling threshold and ratio as diametrically opposed parameters.  If the threshold is low, this will increase the amount of compression applied, this is then countered with easing off the ratio. 

Of course it also depends a lot on the nature of the sound it's being applied to.  A vocal track tends to sound more pleasing (in my opinion) with low threshold and low ratio, where the compression rides the sound  pretty much all the time, maintaining some the loudness variations of the vocalist, keeping things musically interesting while giving the voice a much more intimate sound and giving relative dynamic continuity throughout the piece.

Although a drum track would tend to have a different combined "sweet spot" of ratio and threshold, the technique remains the same.  If you adjust one of the two parameters, it pays to routinely think about about making adjustments to the other.

[Master Mi]

If you use compressors/limiters just as a tool for sound engineering (to create new, interesting sounds) it might be okay.

But just for gaining more (undefined) loudness it doesn't make any sense to use any compressors/limiters if you want to make high quality music - compressors/limiters ALWAYS cut a bit of your sound quality, the natural sound of your VSTI or real instruments.

If you produce music at the basis of EBU R 128 guideline and loudness level there's not even a necessity of using compressors/limiters - these are just loudness tools, weapons of mass destruction, primarily created for the loudness war to destroy the good ol' high quality music with just loudness.

Just listen to my newest version of my Lufia metal remix - it's based on the EBU R 128 loudness and has been produced without any limiter/compressor - it has the full dynamic range and still lots of further headroom (about another 8 dB) for even stronger peaks.

https://soundcloud.com/master-mi/lufia-2-tyrant-breaker-master-mi-remix-version-15


For those who are interested making music at EBU R 128 standards make sure to read my little text I've written about the EBU R128 guideline in the description of my track:
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
"In this version I removed all the compressors/limiters completely and mixed the track again to give the track its full dynamic range without losing a single piece of the sound quality of the instruments.

The standard volume of this track is about 10 dB quieter now.
It's because this track will be my first one which is produced at EBU R 128 standards, a pretty well thought-out technical guideline of the European Broadcasting Union developed by professional sound engineers and other engineers of different broadcasting stations and broadcasting institutes.

With this guideline the main goal of music producing is not to bring the sound peaks as well as the programm loudness of a track close under 0 dB for maximal loudness - because this often results in in a much smaller difference between peaks and programm loudness by using compressors and limiters which cut or really destroy the natural sound of instruments as well as the loudness range/dynamic range and the sound quality in general.
But good music has to breathe as much as it can - while hard limiting or compression is more like gasping or the chronic obstructive pulmonary disease of music.
At EBU R 128 standards music production is not primarily peak-based - it's primarily based on the programm loudness which is also a much better indicator of the perceived loudness of a soundtrack or other kind of programs.
There the target level of the program loudness is - 23 LUFS (dB) +/- 1 dB. So, the program loudness is not allowed to be louder than this limit.
But I allow myself to be much quieter within my track if I want to keep important loudness differences - so, if one of my tracks has an overall loudness difference (loudness range) of about 5 dB the quietest part of the soundtrack will be around - 28 LUFS (dB) and the loudest part of the track will be around - 23 LUFS (db) then.

That means that you'll always have a lot of headroom for the peaks of your music.
At EBU R 128 it's determined that the true peak level shouldn't go over - 1 LUFS (dB).
Based on the maximal program loudness of -23 LUFS (dB) this makes a lot headroom for the peaks (range of at least 22 dB just for the peaks).
So, you don't have to worry about the peaks like a shareholder and there is no basic necessity of using compressors and limiters just to gain the last piece of undefined loudness out of your track anymore.
And if the EBU R 128 loudness becomes the new standard within the next years all those will be fucked who rather produce music at loudness standards with lots of compression.
But those who make totally uncompressed music with the full dynamic range will be blessed.
Because with the EBU R 128 standard the uncompressed high quality soundtracks will sound as loud as the compressed and hypercompressed stuff (so, EBU R 128 would be a great international loudness normalization standard)- and it might sound so much better, clearer - like a musical adventure that tells a big story.   

I'll produce all my further soundtracks, remixes and updates on the basis of this very smart EBU R 128 loudness guideline - so, make sure to turn up the volume a little bit."
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

[timaeus222]

I'll produce all my further soundtracks, remixes and updates on the basis of this very smart EBU R 128 loudness guideline - so, make sure to turn up the volume a little bit."

-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

 

Uh, I had to turn up my volume from 32% to 72% to hear enough detail. That happened to be the near-equivalent of 16 dB... (And I never go higher than 32% on headphones, or 52% on speakers, which is really saying something.)

 

An example of taking advantage of full dynamic range: https://soundcloud.com/biggiantcircles/borderlands-2-tropical-paradise-exploration

[Flexstyle]

Compression does not "destroy" the sound in most use cases. Without compression, we wouldn't have the classic sounds of Pink Floyd, the Beatles, or literally any other major band of the last 70 years. It changes the character of the sound, yes--duh, that's also what EQ, delay, reverb, and literally every other effect out there does. It's incredibly useful as a mix tool--without it, you wouldn't be able to hear half of most vocalists in most music. The instruments would simply drown them out, even if they themselves were uncompressed. To throw away compression completely is asinine. Heck, distortion and overdrive are just extreme compression. Without those, say goodbye to electric guitars, among other things.

Enjoy your trip into theoretical la-la land: without compression, modern music sounds terrible (and 99.99% of listeners will agree). I'll be over here getting paid to do the job correctly, like I have been for years.

[Nabeel Ansari]

You can not mix without compression. Even relatively acoustic music uses compression on the elements to make them stand out more.

 

Compression does not "cut sound quality". It's not signal representation compression (bit/sample reduction), it's peak compression (modifying amplitude per threshold cross). Amplitude modulation doesn't degrade information. You can uncompress a signal if you painstakingly do the math (there aren't many tools that do this because the math is obnoxious). If information is not lost, you are not cutting sound quality.

 

Btw, Timaeus, you asked this earlier; gain reduction corresponds to ratio, they are different representations of the same thing. Ratio is what you set as the ratio you want to drop, gain reduction tells you the actual dB dropped as a result of the application of that ratio (basically just converting that ratio over to a dB value per a specific loudness over the threshold). That's what those meters are (like the big center one in TLs). GR is not a parameter you can adjust directly, it's a description of the end effect, but you can check how much GR your ratio is applying and appropriately adjust your ratio to change the GR.

 

I suppose crudely speaking, if you lowered the threshold, it also increases your GR because the ratio applies to more signal. So in that case, it's actually both; a general descriptor of the "total effect" your compressor is doing.

[lazygecko]

Uh, I had to turn up my volume from 32% to 72% to hear enough detail. That happened to be the near-equivalent of 16 dB... (And I never go higher than 32% on headphones, or 52% on speakers, which is really saying something.)

 

You'll find this is mostly the case with older music as well (that isn't a re-release and in all likelihood remastered). We pretty much have calibrated our amps for heavily compressed music. Some stuff I can barely even crank up above the lowest settings because it gets unbearably loud beyond that. This is also because compression also introduces inharmonic distortion. It manifests all over the frequency spectrum unlike harmonic distortion which just stacks overtones and is thus "musical", so it is something that is much harder to quantify for most people. All you really know is that it sounds like crap. Inharmonic distortion (AKA intermodulation distortion) is something that always exists when you mix more than two tones. It's just a matter of keeping it at negligible levels. Compressing the master will increase this distortion, so if you play it louder you'll notice it more. This also means that uncompressed music can be played at even louder average volume and still sound very pleasant and palatable. This is pretty much how I was taught that the best general benchmark for technical audio quality is how loud you can play something before it gets unbearable. Dynamic music invites you to play it louder. Brickwalled music does the opposite.

 

I think in particular music released in the latter half of the 80's does an excellent job at utilizing full dynamic range, and this is when the overall technical quality peaked in mainstream music. Digital audio was still new enough at that point to be considered novel and this utopian ideal which has finally been realized, so producers made sure to take advantage of its perks. After a while, the situation normalized and people started to regard pristine, digital audio as something mundane, which also played its part in contributing towards the loudness war with very romantisized notions of processing/mastering (also related to mastering engineers needing to justify keeping their jobs when vinyl was being phased out for CDs. Unlike vinyl mastering which is a very specialized and delicate process, getting audio on CD format is just a simple linear transfer of binary information).

[Nabeel Ansari]

You'll find this is mostly the case with older music as well (that isn't a re-release and in all likelihood remastered). We pretty much have calibrated our amps for heavily compressed music. Some stuff I can barely even crank up above the lowest settings because it gets unbearably loud beyond that. This is also because compression also introduces inharmonic distortion. It manifests all over the frequency spectrum unlike harmonic distortion which just stacks overtones and is thus "musical", so it is something that is much harder to quantify for most people. All you really know is that it sounds like crap. Inharmonic distortion (AKA intermodulation distortion) is something that always exists when you mix more than two tones. It's just a matter of keeping it at negligible levels. Compressing the master will increase this distortion, so if you play it louder you'll notice it more. This also means that uncompressed music can be played at even louder average volume and still sound very pleasant and palatable. This is pretty much how I was taught that the best general benchmark for technical audio quality is how loud you can play something before it gets unbearable. Dynamic music invites you to play it louder. Brickwalled music does the opposite.

 

I like all of this.

[timaeus222]

You'll find this is mostly the case with older music as well (that isn't a re-release and in all likelihood remastered). We pretty much have calibrated our amps for heavily compressed music. Some stuff I can barely even crank up above the lowest settings because it gets unbearably loud beyond that. This is also because compression also introduces inharmonic distortion. It manifests all over the frequency spectrum unlike harmonic distortion which just stacks overtones and is thus "musical", so it is something that is much harder to quantify for most people. All you really know is that it sounds like crap. Inharmonic distortion (AKA intermodulation distortion) is something that always exists when you mix more than two tones. It's just a matter of keeping it at negligible levels. Compressing the master will increase this distortion, so if you play it louder you'll notice it more. This also means that uncompressed music can be played at even louder average volume and still sound very pleasant and palatable. This is pretty much how I was taught that the best general benchmark for technical audio quality is how loud you can play something before it gets unbearable. Dynamic music invites you to play it louder. Brickwalled music does the opposite.

 

Exactly. And that bolded part is how I roll.

[Master Mi]

Just watch this and you might understand what I think about the problems of increasing compression for just gaining loudness and the possible solution of the EBU R 128 standards.