A little something about gain-staging, for those who don't know

[Uros]

-16 dBfs (the average [not absolute] peak level, not RMS) would be 0 VU.

Btw. what meter is in Cubase? I know it's a peak meter, but is it an average peak, or absolute peak meter?

Also, I have one more question on my mind that popped up. If plugins are really calibrated to operate best at 0VU, as the article suggests, then wouldn't it be logical to trim the levels of group busses before going to plugins on those group tracks, or even 2buss?

I guess it sounds stupid, because whole purpose is to build your volume all the way up to the 2nd buss, but this is just regarding this "plugin calibration" thing.

 

[abt]

There was a pretty long thread about this a while back. When I first started doing this I was tracking everything way too hot so changed to tracking at about -18db. At first I thought it was great but after a while it didn't sound as good. Now I've gone back to tracking hot at around -5db then before I do anything else I use the PT trim plugin last in the FX chain and trim it down to -18db. To me the printed hot signal sounds better. Having the hot signal going through all the FX plugs seem to sound better too. Perhaps I've got it wrong. I keep planning on doing more experiments with this but never get time to.

 

[AudioGeekZine]

when I say "average peak" I mean at or around -16. Absolute peak meaning analyzing the file to find the peak level and normalizing it to -16.

then wouldn't it be logical to trim the levels of group busses before going to plugins on those group tracks, or even 2buss?

no because the faders of the tracks feeding the group become the trim in that case.

because whole purpose is to build your volume all the way up to the 2nd buss

I disagree. Mixing isn't about volume at all. It's about balance, and making it sound brilliant.

 

[AudioGeekZine]

There was a pretty long thread about this a while back. When I first started doing this I was tracking everything way too hot so changed to tracking at about -18db. At first I thought it was great but after a while it didn't sound as good. Now I've gone back to tracking hot at around -5db then before I do anything else I use the PT trim plugin last in the FX chain and trim it down to -18db. To me the printed hot signal sounds better. Having the hot signal going through all the FX plugs seem to sound better too. Perhaps I've got it wrong. I keep planning on doing more experiments with this but never get time to.

at least you're not oblivious to it and the choice is made based on what sounds good to you.

Perhaps I've got it wrong.

Perhaps we've all got it wrong!

 

[Uros]

I disagree. Mixing isn't about volume at all. It's about balance, and making it sound brilliant.

no problem, man, I agree with what you say. When I said that, I was thinking in the way that you don't get your loudness (not loudness for the sake of loudness) on the way in, but rather build it though the mix. Ofc that loudness is not the purpose.

 

[AudioGeekZine]

I knew what you meant I just wanted to be a bit of a dick right there. haha The mix bus should be nowhere near clipping though. Personally I shoot for a bit higher overall like -12dB tracking and mixing.

 

[Uros]

haha, and you changed your title accordingly
What do you use to monitor the levels? Built-in DAW meter or something else?

 

[abt]

I think the post a while back was by Skesis? I could be wrong. It had a link to a super long thread on I think Gearslutz that was really good. There's a really dumbed down article that basically says the same thing here (useful for the slow humans like me):
http://www.massivemastering.com/blog/index_files/Proper_Audio_Recording_Levels.php

 

[fitzrangerhoodngarb]

Dropping all your faders down a certain amount not only loses resolution.........etc

I've only skimmed through this but be aware Stav has had his arse backwards in the past when discussing the concept of resolution as it pertains to digital. Not suggesting that comment is attributable to Stav but it reminded me when you mentioned him further on.

Funnilly enough, this used to be a bit of a concern in the analogue world. Carbon track faders can darken up (sometimes quite significantly) as you pull back from unity to infinity. Conductive plastic faders don't have this problem generally.

Digital ain't flawed in that way (anymore).Or more correctly maybe that should read the analogue component integration and decimation process of the conversion aren't. The business about using all the bits is old hat. The problem with the internet is there is not expiry date on information.....

This: http://tinyurl.com/2ut6sfx at PSW is another good thread.

 

[slashvanyoung]

As far as tracking levels go, I just get a decent level in, I'm not looking at numbers.
DI's, the same.

I do the same.

When I used to track with 16 bit systems, I usually aimed at getting the signal as hot as possible without clipping to get the usable signal as far away from the noise floor as possible. Much like Bob Katz always suggested.

Now with 24 bits and DAWs operating at 32 bit floating point internally, you are more flexible in both directions: tracking colder and mixing hotter. In theory at least.

The dreaded noise from the analogue days has quasi become a non-issue, but what people often overlook is that not all their plug-ins in a floating point environment necessarily operate with floating point. So it makes sense to keep the levels that go into the inserts at reasonable levels anyway.

But I'm still not tenaciously watching the meter for peaks not overshooting -18dBFS as in practice I haven't run into troubles nor heard audible differences when peaking anywhere between -9 to -24 dBFS.

In fact, I never really bothered to trim anything that didn't come close or overshot -6dBFS.

 

[gabriel g.]

isnt -18dbfs=0dBu ?

when you track make sure nothin clips in the converter.
when something clips you destroy information which cant be rebuild correctly.

it is most important to check the levels on your converter NOT in your daw when tracking.
with 32bit floating point you have enough headroom in your daw.
when mixing try to have your masterbus something between -18 to -5dBfs .

if some snare or tom hit is killing your headroom correct that hit.

and of the story.
clipping your ind. channels should not happen if you use plugins that cant handle it (no oversampling, fixed bit-depth)

it is always better to lower the other tracks then to rise up one fader in the red area.
Only because you want to export something under 0dBfs as a 24bit or 16bit file.

so the golden rule is: never clip your master-bus!!!

if your daw meters are post faders you can simply lower the masterbus to avoid clipping.

 

[drew_drummer]

I do the same.

When I used to track with 16 bit systems, I usually aimed at getting the signal as hot as possible without clipping to get the usable signal as far away from the noise floor as possible. Much like Bob Katz always suggested.

Now with 24 bits and DAWs operating at 32 bit floating point internally, you are more flexible in both directions: tracking colder and mixing hotter. In theory at least.

The dreaded noise from the analogue days has quasi become a non-issue, but what people often overlook is that not all their plug-ins in a floating point environment necessarily operate with floating point. So it makes sense to keep the levels that go into the inserts at reasonable levels anyway.

But I'm still not tenaciously watching the meter for peaks not overshooting -18dBFS as in practice I haven't run into troubles nor heard audible differences when peaking anywhere between -9 to -24 dBFS.

In fact, I never really bothered to trim anything that didn't come close or overshot -6dBFS.

You might like to know that DAW's like Reaper and Sonar are running a 64-bit mixbus.

 

[Arsenu,]

true words of wisdom.

yet easier said than done when recorcing heavy fucking METAL

 

[slashvanyoung]

isnt -18dbfs=0dBu ?

when you track make sure nothin clips in the converter.
when something clips you destroy information which cant be rebuild correctly.

it is most important to check the levels on your converter NOT in your daw when tracking.
with 32bit floating point you have enough headroom in your daw.
when mixing try to have your masterbus something between -18 to -5dBfs .

if some snare or tom hit is killing your headroom correct that hit.

and of the story.
clipping your ind. channels should not happen if you use plugins that cant handle it (no oversampling, fixed bit-depth)

it is always better to lower the other tracks then to rise up one fader in the red area.
Only because you want to export something under 0dBfs as a 24bit or 16bit file.

so the golden rule is: never clip your master-bus!!!

if your daw meters are post faders you can simply lower the masterbus to avoid clipping.

Good summary and hands-on advice, Gabriel!

You might like to know that DAW's like Reaper and Sonar are running a 64-bit mixbus.

As do Studio One, Ableton, Tracktion and others...

I referenced 32 bit floating point, because that will already give you practically unlimited headroom. Going to 64 bits might result in more precise summing, but when we're talking about headroom issues here, it doesn't really matter.

 

[enc0re666]

This article and this thread has helped me a heap! Thanks OP and thanks guys

+1!

 

[Dollarosa]

This thread has changed my way of mixing. Yesterday I placed a trim plugin on all my tracks (1st insert), making sure input was equal to ouput (around 0db, a little lower), from then on i mixed as usual. Master bus never peaked, everything sounded smoother, rich like. At all times I was making sure no channel or bus was going in the red. When it came to mastering, wow, it was easy to get loud result without ruining the mix.

Thanks guys

 

[MF_Kitten]

i´m ALMOST getting this, but i´m not quite there...

which input are we talking about keeping down here? if i have a firewire interface with a mic plugged into it, recording a snare or a cab, then the mic preamp "gain" knob on the interface should be turned down until the track in the DAW measures it as being between -15 to -18, or what?

could someone exlain it for a totally-not-even-near-pro guy like me? are we talking everything coming from outboard gear only, or are we talking, for example, DFHS or Metal Foundry programmed drums too? in other words: does this apply to things going through an analog-digital converter only, or does it apply to sound created and mixed entirely within the DAW too, like soft synths and sampled drums etc?

an explanation that goes from step to step in the signal chain would help lots.

thanks in advance

 

[AudioGeekZine]

yes.
into the system. In and out of plugins and at your mix bus.
Every step of the chain.

synths and drum samples are WAY too loud to begin with unfortunately.

 

[SocialNumb]

Must try this "trim plugin" thing. Does Logic have one and where is it? I'm not on that comp right now but I don't remember seeing anything called "trim". Could it be the Gain plug? (If I remember right on the name, never did use that Gain plug)

 

[FloridaRolf]

i´m ALMOST getting this, but i´m not quite there...

which input are we talking about keeping down here? if i have a firewire interface with a mic plugged into it, recording a snare or a cab, then the mic preamp "gain" knob on the interface should be turned down until the track in the DAW measures it as being between -15 to -18, or what?

could someone exlain it for a totally-not-even-near-pro guy like me? are we talking everything coming from outboard gear only, or are we talking, for example, DFHS or Metal Foundry programmed drums too? in other words: does this apply to things going through an analog-digital converter only, or does it apply to sound created and mixed entirely within the DAW too, like soft synths and sampled drums etc?

an explanation that goes from step to step in the signal chain would help lots.

thanks in advance

I don't get any of this too, I feel kind of stupid
Sure, clipping tracks are always bad, but I don't get why a track (in the digital domain) should sound better if you just lower its volume. I mean, I don't care why it should sound better, I just want to know how I should apply the things that have been said in the article / this thread