Gain Staging with Gain Utility Made Easy...

[Zarathustra]

First, I'm assuming the reader knows something about Gain Staging (I say "something" because we all, myself especially, could always learn more about everything!) so if you're unfamiliar- do a little research as I had before coming up with this method (Pros, correct me if I'm wrong).

 

So,

 

I know there are fervent debates on the use of the Gain utility vs fader mixing when it comes to Gain Staging, it seems that a lot of people are looking for a workaround because it can be tedious when you're looking at more than 20 tracks.

 

It seems the questions is: "Do I really have to load and edit an instance of Gain on every single channel strip and trim each one independently?" Well, yes-- and no. It seems Logic doesn't really have a "gang settings" function for it's plugins like UA and some other 3 parties- you just use a BUS but with Gain you don't want to structure it that way. You can always "Option+Click" to drag the plug but even that can take longer than you want.

 

Here's a little trick that I do to get around the tedious process of setting the base level (NOTE: this is only when determining headroom and NOT the "end-all" for Gain Staging:

 

First, determine how much you need to trim your tracks. For me, I generally found that I have to start by trimming -10db to -20db depending on how many tracks are being sent to the master (for smaller sessions even -5db works as a base level). So what I did was save custom gain amounts as presets to my Gain plugin (SEE PHOTO #1), -10, -15, -20, -25. Now, you can get more precise if you need to but for me, this is only a BASE level to ensure my tracks and master are not clipping past -10 to -6.

 

Next, create an instance of Gain on your first channel strip and set it to your custom desired preset. For me, in this example, I needed to trim -20db from my tracks (recorded hot and coming in at 60 tracks)- THEN open the drop-down menu and select "Save As Default". After you've done this, any instance of Gain that you've added will automatically be set at -20. Open up your Mixer window and Shift+Select all of your tracks and open "Gain"- now every track is set to your custom db amount! (Note: For projects that I receive or record that were done in different locations and times with varied input levels I will determine the trim amount based on sections, dividing this exact process into something like, Drums at -20, guitars at -15, etc)

 

From there, you can balance, plug, pan, etc etc as you normally would- just remember to reset the Gain default when you've finished. You can even incorporate this process into a template to save even more time!

[Zarathustra]

Forgot to mention: make sure you are fully aware of the Stereo tracks in your project and to use the right plug.

[stardustmedia]

Indeed. I do that too, with the default preset. Changing them to reuse another preset across multiple channels.

 

But you don't have to worry about mono/stereo. The right instance will be placed automatically, depending on what kind of signal is coming in. On mono channels a mono instance, on stereo channels a stereo instance, and on mono channels where a plugin mono>stereo is placed a stereo instance. There are some plugins (like Waves and Fabfilter) that feature different plugins for mono and stereo. In this case, e.g. if you insert a plugin into a stereo channel, the mono channels won't have the plugin.

 

Just select all channels as described, once put a stereo instance into a stereo channel. All stereo channels will get the plugin. Then leave the channel selection, but this time load a mono plugin on a mono channel.

[amiracam]

So I have been doing this by decreasing the gain parameter on the initial regions i.e. when I get files or when I finish tracking, what's the disadvantage to this approach ? Assuming that gain setup is basically one of the very first things done once source files / tracks are available.

 

I thought that the general approach should be to on a per track basis make sure that the levels where coming in at somewhere -18db to -6db. I've read some variances on that. That implies a per track treatment.

 

Also is there a suggested output level that I should be hovering around once I finish gain staging? i.e. if I look at my output level where should I be around ?

 

Finally, tangent once the gain staging is setup, I wonder how many of the pros then go ahead and put a compressor on the output bus i.e. just for hard clipping prevention purposes?

 

thanks

[Zarathustra]

The disadvantage is time consumption, my most recent projects have been insanely large and the idea of going into each track and manipulating the gain back and forth seemed insane to me so I just find where my master is peaking and do the math to make sure I'm staying well under -6db on the master- then I'll go in and tweak my levels on quieter and louder parts.

 

You can compress the Output all you want but don't use any limiters- that's the Mastering Engineer's job.

[amiracam]

interesting, I don't know what that equation is, I have been doing a trial and error thing, reducing what seem to be the greatest offenders etc

 

any articles, resources on gain staging I can be pointed to ?

 

thanks

[amiracam]

btw, I found this article:

 

http://www.soundonsound.com/sos/sep13/articles/level-headed.htm

 

I also put a Waves Meter on my stereo Bus:

 

I'm getting Peaks of 5.1 and RMS of 16.6

 

all of the my tracks are now set to be within -12 to 18db , a couple are at -20db

 

So is this considered a good place ? or still to hot ?

[amiracam]

of course , I mean and should have written -5.1 and -16.6 respectively, thanks

[stardustmedia]

Now, that's fine. I stay at -18 or just a little below. And I put the signal back to -18 after I add a new plugin, to make better A/B comparison.

[amiracam]

so sorry just to be clear let me recap,

 

I should be RMS -ing closer to -18db and after I add a new plugin I should go back to my gains and tweak them accordingly so that my RMS stays approx at -18db or bit lower

 

thanks

[stardustmedia]

so sorry just to be clear let me recap,

 

I should be RMS -ing closer to -18db and after I add a new plugin I should go back to my gains and tweak them accordingly so that my RMS stays approx at -18db or bit lower

 

thanks

 

If I understand you right: Nope

 

Let's say you have a signal on a channel without any plugins. The signal peaks (not RMS, I measure the peaks) at -10dB FS. So I add a gain plugin first at -8dB.

 

Now, when I add an EQ and the processing causes a "new" signal peak at -24dB FS, I'll add +6dB in the EQ-output (if output gain is available/possible) or I'll add another gain plugin with +6dB after the EQ. Then the same with the next plugin. And so on.

 

Let's do an example where you have 4 plugins chained, that each causes a drop of 6dB, in total 24dB. If you wouldn't correct each 6dB drop right after the plugin you'll result in a signal at -42dB. Now if you try to add +24dB in the gain plugin of your signal chain, you'd end up with digital clipping.

 

Or let's say you would like to add a 5th plugin, a compressor. But the -42dB is so low, some compressor wouldn't even start to compress, no matter how low you set your threshold. Imagine a sound that is locut so heavily that it's peak is -30dB lower. The next plugin might have issues with this kind of low signal.

 

You can do the contrary, by driving a compressor hot. So you increase the gain until e.g. -5db FS peak, so the compressor is driven and starts including distortion. Of course you should dial back the gain on the compressor out again to -18dB FS

 

So, I always correct any increase or decrease of gain after each plugin. You don't have to, just be aware of the influences and leverages.

[amiracam]

ok, thanks so much , so the aim is to deliver a consistent level across the signal path from one node to another, some nodes exhibit more character when pushed a hotter but one should account for that and lower the input to the next node etc.

 

Now, how does that work for sends? should I have a gain plugin on the corresponding bus to make sure that the combined signals don't push it over the threshold ?

 

thanks

[stardustmedia]

For effects like reverb, delay, etc. No! (... usually)

Usually there is enough level so the effect works properly. But if you should have only one very low level channel signal sent to an effect, you might add before the effect a gain and lower the effect fader accordingly. Or you could put the effect send "pre fader", but you'd loose the relative relation between fader and send.

 

Aynway, a gain plugin before the effect doesn't make much sense though, if there are other channels sent to the same effect at "normal" levels. The gain plugin before the effect would probably cause distortion and you could anyway rethink about if the low level sends (e.g. less than -40 to -50dB) are really needed? Especially since most channel fader will be lower than 0dB

[amiracam]

and sorry again, at any point in time during the mix process I should have my tracks levels peaking at -18db ish, no matter what plugs or EQs I have added , the last plug's output if available should be reduced in order to attain the -18db-ish level

 

sorry if repeating myself , just seems like a fundamental concept , want to get it right

[Eric Cardenas]

I wouldn't follow this example.

The theory is to keep your peeks <= 0.0 dbFS.

 

The -18dB measurement is used as a calibration level comparing digital audio to analog hardware.

If you want to use this scheme aim for -18dBFS VU or if you don't have a VU work with RMS.

 

It's basically most important when tracking and in monitoring but since you may use other outboard gear and analog emulating plugins it also has a purpose in the box.

 

The easiest way to set this up is to insert a VU plugin. Make sure it's calibrated to -18dB and start recording/playing the track. The. Needle should move around the 0 dBVU mark. It doesn't matter if it gets over that mark unless your peaks are clipping. Most VU metering plug-ins have a gain knob or fader that will let you set the optimal level if the track already is recorded. If you are recording simply use the gain knob on your preamp.

 

Now you can process your audio with an external compressor much in the same way as you would when using an analog console.

Pushing up the gain can now end up in nice saturation and lowering the gain can now help you get a "cleaner" sound.

The same applies to many non-linear analog emulating plug-ins.

[amiracam]

but I thought that whole idea was to provide oneself with enough of a "padding" i.e. headroom, if I'm just making sure nothing is hitting digital clipping i.e. going past unity, then won't I easily by riding a fader up push my stereo out into clipping ? If so then always control that with a limiter on the stereo bus ? Is volume balancing then mostly a practice of riding down faders ? there's also the issue of levels from one plug to the subsequent plug, I believe I have read that for at least some of the UAD plugs going that hot i.e. just pre-clipping may be too hot due the UADs physical modeling , perhaps that the case with other physical model based plugs

[Eric Cardenas]

I added a bunch of stuff to my post to give you an example of why some like to work this way.

[Eric Cardenas]

but I thought that whole idea was to provide oneself with enough of a "padding" i.e. headroom, if I'm just making sure nothing is hitting digital clipping i.e. going past unity, then won't I easily by riding a fader up push my stereo out into clipping ? If so then always control that with a limiter on the stereo bus ? Is volume balancing then mostly a practice of riding down faders ? there's also the issue of levels from one plug to the subsequent plug, I believe I have read that for at least some of the UAD plugs going that hot i.e. just pre-clipping may be too hot due the UADs physical modeling , perhaps that the case with other physical model based plugs

 

Headroom in this case is the difference between the VU level and 0 dBFS.

Keeping your peeks below -18 dBFS will result in you never hitting he "sweet spot" since your "average" level is well beyond the optimal for a variety of hardware and emulations.

[amiracam]

ok, so you would then run the channels hotter than -18db , but are you saying that you would strive to push them as close to 0.0 db via region gain parameter/ gain plugin initially and then via plug output controls if available ? I know ti all "depends" but as first pass setup where I'm trying to get a basic rough balance , with sculpting EQ, and comps, I'm looking for a visual guide across my channels i.e. the peak level readings where I can see that all the channels are falling within an accepted "good" range

 

sorry if I'm being dense, thanks

[Eric Cardenas]

but are you saying that you would strive to push them as close to 0.0 db via region gain parameter/ gain plugin initially

 

No. I never said anything like it.

You can use the Gain parameter and push it 0.0 dBFS or over or under. That's really not the point here.

Why would you want to push it close to 0 dbFS?

[amiracam]

well I don't , just trying to get a heuristic , honestly I don't know what the pros do. The main commandment that I know was not to allow anything to clip particularly when working digitally i.e. "in the box" which is what I do , I don't have any external gear besides an AI i.e. no boards , no external processing.

 

however, I keep hearing about the importance of gain staging prior to getting into the more creative aspects of mixing, so I want to get that right which is why I'm seeking some guidelines as to an acceptable range so that I don't get into trouble at the latter stages of my mix process , as they say "don't want to paint myself into a corner"

 

I accept that I may have some gaps in my understanding

 

thanks

[Eric Cardenas]

I don't know what the pros do.

 

There is no definitive answer to this.

They are many ways to approach this. Another option is just to disregard this altogether since we are working in floating point anyway.

 

not to allow anything to clip

 

That's a great rule to follow.

Many people I know aim for a certain peak level when recording, whether it is - 3, -6 or -10dBFS doesn't really matter.

Once recorded you can safely go over 0.0 dBFS -> because floating point. Just make sure that your output isn't clipping unless you are going for that effect.