Mastering Techniques for LOUD

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

john12ax7

Well-known member
Joined
Oct 15, 2010
Messages
2,463
Location
California, US
Since there a few pro mastering guys wanted to get some input.

I'm not saying masters should be made really loud,  but am curious about what techniques or gear is used to achieve it. I've heard commercial masters at -6 LUFS or even higher that still sound pretty good.

When I've tried to do it things are fine at -10,  at -8 artifacts start to become noticeable,  and at -6 LUFS it's a big mess. I mix with hardware but the final limiting is all plug ins, waves and/or ozone.

Any advice or tips on technique or gear?
 
After many years I conclude a great master comes from a great mix

So a great mix is achieved by lots of subtractive eq carefully tuned so mixing engineer can push levels and make things sound big without sacrificing headroom (sidechaining elements or fader riding always down). It's pretty easy once we understand why we should make things softer to make them louder later. There's no loud beyond louder.

I recently mastered a jazz album. Client wanted it really loud and gave me some references (pop and electronic). Some parallel compression, subtractive MS EQ, adding saturation and distortion (yes, it was jazz) and carefully compensated gain to not fool my ears. After that I chained 2 limiters: first multiband shaving peaks without raising volume; second one for loudness.

I was not satisfied reaching - 8 LUFS on jazz so I invited artist to come to my studio and we compared - 8LUFS master against - 14LUFS level matched. He finally chose dynamic one as he was only releasing on streaming.

Maybe - 6 LUFS sounds pretty good at first listen but it will be inevitably ear fatiguing.
 
It's all about the arrangement, and then the mix.  The big records you hear are for the most part being mixed at those levels or with a dB or two, the mastering girl or guy is not adding significant level.

The old Bob Katz style advice of "Mix for sound and let the mastering engineer take care of level" is horribly out of date, there is absolutely no way to turn a 2005 mix into a 2018 competitive master without destroying it.

The techniques are the same as they ever were :

- EQ
- Clipping
- Limiting

Compression on the master does not help with getting level.

I just spent the morning pushing an absolutely gorgeous record maybe a dB and a half beyond it's comfort zone, artist signed to Major and they were very clear on what they wanted.  I did everything I could to make it work using the above techniques.

You have to ask yourself why you want the master to be that loud.  Sometimes there are good reasons and sometimes not.

The limiter matters, and they are not all the same.  I use five that I have setup in a template to audition quickly per project.  Usually I can tell which will work as soon as I hear the music.

Edit 2 : Extreme low end eats head room but it's almost impossible to make real decisions on what to do at 30 Hz and below without a very serious room and speaker system.  Steep high pass filters are not without side effects, and these days it is very rare for me to use more than a 6 or 12 dB per octave filter.

 
" it is very rare for me to use more than a 6 or 12 dB per octave filter."

Is this just in the case of mastering work?  Which is to say have you worked on material that was mixed with brick walls at 30Hz and found it unruly?
 
boji said:
Is this just in the case of mastering work?  Which is to say have you worked on material that was mixed with brick walls at 30Hz and found it unruly?

I'm not sure I follow your question but let me try.  All of my comments above are about mastering, I haven't done mixing for the last 7 years.

I have worked on every kind of unruly you could imagine, stuff that was wild down low and stuff that was overly filtered and wrong.

My main point was that super low stuff really eats headroom but hearing properly what you are doing down there is really hard and sometimes the phase response change caused by steep filtering can actually cause more problems that it solves (causing more peaks).
 
ruairioflaherty said:
My main point was that super low stuff really eats headroom but hearing properly what you are doing down there is really hard.

For listening to what is going on way down low I use headphones. I don't use them for balance, just for hearing what is going on free of room and speaker artefacts.

Cheers

Ian
 
ruffrecords said:
For listening to what is going on way down low I use headphones. I don't use them for balance, just for hearing what is going on free of room and speaker artefacts.

Cheers

Ian

I did the same for many years until I finally got into a truly great room.  It was the best I could do.  That said there is no way I could make the kinds of decisions I do now on headphones, but I worked with what I had. 

 
I generally agree with what Ruari said but I use some different techniques to get there.

I’m a big fan of multiband compression. Often it’s a discrete event that will prevent a loud master. Like a vocal phrase that sticks out and would distort if you just slapped a limiter on it. Careful multiband compression can solve that pretty transparently.

I only use one limiter.  I do it  after all the EQ and compression. All my preprocessing takes into account what the limiter will do later.

As a general strategy try smashing it through the limiter.  The parts that don’t make it through without damage are the things to concentrate on.  This circles back to arrangement. If things are stepping all over each other you do what you can to create space and clarity before going for level.

That said I don’t do much shiny pop or stuff that has to be stupid loud.  I do some hip hop that does and I always seem to be able to get there.
 
The parts that don’t make it through without damage are the things to concentrate on.

That's good advice. I've used comparisonics with some good result.

phase response change caused by steep filtering can actually cause more problems that it solves
Gotcha. Thank you!

To add context to my vague question, I've used fabfilter's 96db/oct hpf to pull out info below 30Hz sometimes(mixing), and was wondering if I might be doing more damage than good come mastering time.

Edit: Not trying to hijack your thread 12ax!

 
Interesting thread!

I generally support the idea of hazel: Making mixes/masters sound great needs a good foundation to work on. If the mix already sucks regarding equalizing and pitching (especially kick and bass) you won't be able to get the best results.
To my knowledge also 3D & psychoacoustic modalities can work when trying to achieve loudness.
 
boji said:
To add context to my vague question, I've used fabfilter's 96db/oct hpf to pull out info below 30Hz sometimes(mixing), and was wondering if I might be doing more damage than good come mastering time.

Never tried that one.....

I remember liking the BX Cleansweep for some high passing..... Been a while since I've played around though but, I do remember it being at a nice slope.... Maybe not aggressive enough for some  but, for sure a steep filter totally changes stuff....

That old  Engineer's filter plug in really got aggressive but would do the weirdest stuff.....
 
boji said:
The parts that don’t make it through without damage are the things to concentrate on.

That's good advice. I've used comparisonics with some good result.
Gotcha. Thank you!

To add context to my vague question, I've used fabfilter's 96db/oct hpf to pull out info below 30Hz sometimes(mixing), and was wondering if I might be doing more damage than good come mastering time.

Edit: Not trying to hijack your thread 12ax!

fabfilter does some weird stuff in the high frequencies (phase errors etc.) when cutting low-end! i also experienced that with my mixes. i also thought it to be very harsh in the highs (more usable for surgical editing). better off using something like the algorithmix eq or other linear phase. especially for masters. it wouldn't be noticeable on a single mix track though


scott2000 said:
Never tried that one.....

I remember liking the BX Cleansweep for some high passing..... Been a while since I've played around though but, I do remember it being at a nice slope.... Maybe not aggressive enough for some  but, for sure a steep filter totally changes stuff....

That old  Engineer's filter plug in really got aggressive but would do the weirdest stuff.....

cleansweep is really nice and also has a few more slope values if i can remember correctly?
 
weiss said:
fabfilter does some weird stuff in the high frequencies (phase errors etc.) when cutting low-end! i also experienced that with my mixes. i also thought it to be very harsh in the highs (more usable for surgical editing). better off using something like the algorithmix eq or other linear phase. especially for masters. it wouldn't be noticeable on a single mix track though

I don't like the Fabfilter EQ (1 or 2) as much as some other people but the advice to go linear phase is bad, especially for high pass filtering.

I've posted it before but this video gives at least some insight into why

https://www.youtube.com/watch?v=efKabAQQsPQ

If you have good monitoring and very good room acoustics the damage linear phase eqs cause is very obvious.  I use them very rarely and generally because I want to make use of their weak point - the pre ringing to smear HF energy across time.
 
weiss said:
cleansweep is really nice and also has a few more slope values if i can remember correctly?

I don't think so?? I think I must have got the free version.... It's been a while but I just checked and I don't see where you can get different slopes...... ??? If there are, I've never used them...lol


edit....just checked the site.....6db....lol....no wonder I always thought it sounded good.....But it can change stuff way down low......even that gentle......

 
"If you have good monitoring and very good room acoustics the damage linear phase eqs cause is very obvious.  I use them very rarely and generally because I want to make use of their weak point - the pre ringing to smear HF energy across time."

+++1 Thank you!!! Video on lin phase pre-ringing was a super example and very helpful.

I'll be rethinking my tinkering
 
Thanks for all the info so far.

One thing I've noticed is that the snare can often be the biggest peak.  So I've started using a limiter on the drums to catch some stray peaks, then the final limiter is not being triggered as much by then.

But it seems I need to go even further with the actual mix.  To get a loud master you need a loud mix.

Just to clarify,  I'm not necessarily for ultra loud,  but want it in my arsenal for certain genres and instances.

Btw why would you want a 96db HPF? Even if it worked perfect is there an advantage vs a more gentle cut?
 
To get a loud master you need a loud mix.

I recently had the opposite happen. I got plenty of clarity on the vocals and drum machine / samples, (punch included) but when moving over to mastering (they asked for LOUD) I had to compensate, and got rid of other HF in the process. :/ 

I'm trying to learn to 'mix for what it will sound like when mastered'.
 
boji said:
I recently had the opposite happen. I got plenty of clarity on the vocals and drum machine / samples, (punch included) but when moving over to mastering (they asked for LOUD) I had to compensate, and got rid of other HF in the process. :/ 

I'm trying to learn to 'mix for what it will sound like when mastered'.

Re the top commercial tracks that we would all consider loud and punchy regardless of genre; the mixes sound very very close to the master, I mean so close that an untrained listener may have a hard time telling the difference.  The name mixers - Manny, Jaycen Joshua, Costey, Brauer, Ali. Spike Stent, Scheps etc are all mixing into limiters and delivering mixes that are within a dB or two of finished level.

Often the master can sound worse than the mix, but it's just that little bit louder.

So, your mix should sound exactly like what you want the finished record to sound like, if you are chasing a lot of level.  If not then you can go the old fashioned way and have fun mixing.
 
Phase response of a 96 dB per octave high pass filter using Fabfilter Pro Q2

Plug in Doctor displays Radians rather than degrees but the top of the graph is +180 degrees and the bottom -180 degrees.

You can see even with a 30 Hz high pass you have 360 degrees of phase shift at 100 Hz.  On some sound sources that will not be audible (slow moving evolving pads for example), on others it can make a big difference (tight, percussive kick drums).


Edit : lots of typos
 

Attachments

  • Phase_Shift_96dBOct30Hz.png
    Phase_Shift_96dBOct30Hz.png
    265.2 KB · Views: 53

Latest posts

Back
Top