• Audio Mixing Mastering

What is Gain Staging?

Gain staging is one of the most conflated terms used right now. Lots of people love to throw around this buzzword, almost as if they're casting a magickal spell. But to have a meaningful conversation around gain staging, it requires going into a LOT nuance. Hold on to your butts, it's gonna get fun.

At the core, all gain staging actually means is setting your levels appropriately at EACH stage of your signal chain for the best compromise between distortion and SNR. This includes both "in the box" signals and analog - either on the way in to the DAW or routing through outboard gear and back into the box.

General Rules: • In an analog realm, basically record as hot as you can without blowing it up (clipping). • In the box, once audio is all digital 1s and 0s, make sure that each of the various stages (especially plugin inputs) is optimal. More on what "optimal" means, below.

Ideally your volume levels will be peak between -12 dBFS and -6 dBFS before you start balancing with faders.

You want to set "gain" at the appropriate stages. For example, gaining as hot as you can go in the preamp, without blowing up your microphone preamp, as opposed to recording that track where it peaks too low at -40 dBFS and then using plugins to gain it up +30 dB which will potentially sound like crap.

========================= Noise floor / SNR =========================

If you record your signal too low and then just "gain it up" to the level you *should have* set in the first place, it'll work but it potentially amplifies the noise floor as well.

There are two main reasons why your signal may be recorded too low and need amplification in the box later: • Your source was too far from the microphone. • Your gain on the microphone preamp was too low.

---------------------------------------- Your source was too far from the microphone. ----------------------------------------

If your source is too far from the mic, it's going to pick up more of the room. Also your microphone and signal chain have a certain amount of "self-noise" that is always present. This is typically referred to as SNR (signal to noise ratio) and as you gain up the signal in post, you're amplifying both the recorded source AND the self-noise floor.

The whole point in buying quality interfaces and preamps with a high SNR, is to capture as little noise as possible. When you gain the signal +40dB you also gain the noise floor +40dB.

To reach your optimal level, if you only have to gain the track up +6dB, you only bring up the noise floor by +6dB. So it behooves you to get your source as close to the microphone as possible. This goes into recording engineer duties, as opposed to mixing and mastering.

---------------------------------------- Your gain on the microphone preamp was too low. ----------------------------------------

This gets very nuanced. I actually had a discussion with Bob Katz regarding this, because I wasn't too sure myself about the exact details. But some preamps introduce non-linear noise. Which basically means that you may actually have more or less SNR, dependent on where the gain is set on the preamp.

I'm not going to lie, this is way beyond my understanding of how preamps actually work. But I don't really need to understand the reasons why, to know what needs to be done. Record reasonably hot. General accepted levels in the profession audio world is to peak between -12 dBFS and -6dBFS. This is debatable, but this is what is generally accepted as optimal. It also depends on how dynamic the material is. If you're recording an instrument with transient that can jump 10dB, you probably want to leave more room.

---------------------------------------- From the book "Sound for Film and Television 3rd ed." (Tomlinson Holman): ----------------------------------------

"There are many points in an audio chain at which the signal may get to be so large that clipping or distortion occurs:

• In the microphone, especially in those equipped with their own electronics. • In the microphone preamp. • In the ADC conversion for digital recording. • In subsequent signal processing to "improve" the sound quality. • When multiple sources are summed together in a mixing console: One might not be too great a signal, but many added together may be too great. • On any intermediate recording stage. • At the final recording stage to the release medium."

========================= Compression & Limiting =========================

Some people may conflate gain staging with the functions of a compressor or limiter (an aggressive compressor), but it really has nothing to do with compression.

A compressor changes the output level based on settings and how they relate to the input signal (or possibly sidechain signal if it's used). Proper gain staging is setting the correct static levels throughout the signal chain.

The closest area where compression is related, may be something like modulation peaking / distortion from peaking on certain gear like a radio pack transmitter/receiver) or hitting built in limiters in analog gear, but I wouldn't conflate these "peak protection tools" with traditional compressors.

Also, if you're relying on a compressor / limiter heavily to solely prevent peaks or clipping when tracking, you're likely using it wrong. A limiter should be for sparse transients that just happened to be too hot, not the majority of your signal. If your limiter is pegging the red often, your gain is simply too hot - pull it back before it hits the limiter. Unless of course you like the way the limiter sonically sounds. But this is an artistic choice - and one you can't reverse later on if you change your mind.

A compressor is a great tool on the way in, to control peaks and shape the tone, but it should be used as a tone shaping tool, not a crutch because your signal is too hot, due to poor gain staging.

Each of these things comes at a cost, and not typically a sonically pleasant one. Unless "over squashed" is the production decision you're going for. And even if that is the case, you likely should first get it clean signal recorded and then hyper squash it later (in the studio / post) so you have flexibility once you have context of the entire project soundscape. If you nuke it on the way in, you're committed. You get what you get and it had better be perfect, because you can't un-compress / undistort later.

As always, there are exceptions to the rule. Some digital clipping can be restored with tools like RX, but that's far from ideal and is a lot more work for everyone else you work with. It still doesn't give you the same quality you'd have gotten by setting levels right and the worst part - will probably get you fired or not welcomed back in the next project.

========================= "Gain Staging" & DAW internals =========================

Gain staging is hugely complicated by modern tools (DAWs). Back in the day (and still today where analog gear is being primarily used), you had a mix board and for each track / channel, you have very specific and physically-limited workflows. You're literally just transferring / splitting / amplifying electricity from one place to another and there are limitations to what you can do, before things start to either sound bad or gear is ruined.

---------------------------------------- Gain vs. Volume (Fader) ----------------------------------------

First things first, you need to understand the virtual signal flow if your DAW. Most DAWs operate virtually in a very similar to the way a physical mix board works.

There is always room for a DAW to use different terms, but universally - they SHOULD agree on these very specific terms of gain and volume.

"Gain" affects the signal level of the clip/region that is feeding a track. It's the very FIRST part of the DAW track processing.

"Volume", which is typically represented as a fader (most DAWs have a "fader" view of some sort) affects the signal level of the individual track/bus at the very END of track processing, AFTER all your plugins.

I highly recommend you find a flow diagram for your DAW and study it to understand the way your DAW routes the signal. This will greatly improve your skills and understanding.

As such, you need to consider where problems can occur at each stage. The following is a basic example of the various stages and possible problems you can run into.

------------------------------------------ 1. Preamp on the way in ------------------------------------------

The analog realm!

All signals have to start from somewhere physical. Unless of course you're making music that only uses virtual instruments in your DAW. That is the one exception.

Essentially the challenge here is: Are you blowing (clipping) up your preamp / converter (ADC)? ^^^ MOST IMPORTANT. YOU CANNOT RECOVER FROM THIS. PERIOD.

Don't get fired for a very simple mistake.

------------------------------------------ 2. Clip/region "gain" (level of the raw recording) ------------------------------------------

By default, your clips in the DAW will be at and playback at the exact same level your converter got the signal as. You record a signal that peaks at -6 dB, it'll be in the DAW as -6 dB.

Your DAW likely has a "gain" knob for your track.

This adjusts the signal level BEFORE it is routed into any FX in your virtual FX rack!

Clip gain or track gain - is crucial to set correctly, but easily fixable later (no permanent destruction of audio). This is a non-destructive tool within the DAW.

Where some people can get hung up, is on the idea that internally your DAW has infinite head room (technically 700ish dB in 32bit float). This means that internally, you can clip a track to +100dB over 0 dB. It will sound like asshole when it comes out of your interface converter (DAC) and plays back on your monitors, but you can "recover" the signal within the DAW, because your DAW runs at 32bit float or 64bit internally. This means that if you peg a track at +100dB, feed it into a bus that then drops it 100dB before it goes out the master bus, you'll have zero quality loss.

^^^ This assumes you are using NO plugins and are JUST messing with levels in the DAW.

Even though you *CAN* recover signal in this way, you should NEVER rely on this ability for many reasons. This gets nuanced (the whole reason I'm writing this article), but just trust me for now, this is real bad habit to get into, even if you can recover your audio later. Having discipline in your workflows can make your life a lot easier further down a mix and this is one of those things that I'd highly recommend you not make a habit out of, unless you know exactly what you're doing.

One of the exceptions is where all of your tracks are sounding great and not peaking, but after being fed into a bus, they do peak the bus. For example, all your drum tracks feeding your drum bus. You could go and drop all the individual tracks by the same amount, but that is a pain in the ass and also affects other things like your auxiliary sends into other FX busses. So instead, I'd drop the input gain on the bus by probably 6dB to 10 dB and like magic, because my DAW has gobs of headroom (700+ dB), it works like magic. No sonic degradation and all my mix levels stay consistent. This is one of the benefits of working in the digital realm. If this were an analog mixer, the ONLY thing you could do would be to drop the track faders.

^^^ I typically don't like to get into this level of detail, especially when talking to newbies to mixing, because it requires understanding semantically how audio is processed. But I'm having faith that this is explained well enough, that the folks here will understand it. We'll see .

At this stage of the mixing process, your "target" for optimal processing is generally between -12 to -6 average peaks. Reasonably, you can go all the way up to anything that doesn't hit 0.

The main reason you need to be mindful of this stage, is because this will directly affect the signal level that is fed into your plugins, Which *CAN* be VERY important. Read below.

------------------------------------------ 3. Plugins (signal level in and out) ------------------------------------------

Signal processing is typically fairly straight forward in your DAW, but there are some things to know. Your plugins are fed the signal post clip / track gain adjustments. All DAWs have a virtual "plugin rack" where you can stack your plugins and the order they're processed is pretty obvious:

Clips / regions in the track -> plugin 1 -> plugin 2 -> plugin 3 -> yadda yadda -> volume fader -> where ever the track is routed (a bus, master bus, monitors, so forth).

There is an exception to this. Some DAWs have special built in plugins that are processed outside of the virtual FX rack. Examples are things like Cakewalk's pro channel, or Studio One's Fat Channel. These built in FX are typically processed BEFORE your plugin FX rack, but most DAWs let you switch that between pre/post FX rack. It doesn't really matter that much, but you should just be aware of where your DAW is processing it:

Clips / regions in the track -> built in FX -> virtual plugin rack -> track fader -> … vs. Clips / regions in the track -> virtual plugin rack -> built in FX -> track fader -> …

A lot of plugins now a days are emulating some type of analog gear and they have their own optimal levels they operate at. Let's pretend you have a plugin that models a piece of gear that uses tubes. In the actual piece of gear that is being emulated, there is a physical threshold for how much voltage that tube can handle before the device utterly fails. If you're feeding a signal 3 times hotter than the device was meant for, it's obviously going to sound horrible and could actual fry the physical unit.

The same goes for your plugins that emulate these physical boundaries. All plugins (except special mastering plugins that specialize in clipping - don't need to go down this rabbit hole right now), expect your signal to be less than 0dB on the way into their processing. So you need to make sure that you've properly set your optimal level via your "gain" settings on your track / clips.

^^^ Or don't. Just because a manufacture recommends your signal enter the plugin at a certain level, that doesn't mean you don't like the way it sounds! As always, your ears are king.

Now, certain plugins like Slate virtual mix rack, are extra fancy in that they have multiple modules. And you need to think about what's happening in between EACH module. If one of your modules is pushing the signal up too hot into the next module, you've essentially made the same exact mistake as feeding a too hot signal into VMR in the first place.

The same principle goes between each plugin effect in your virtual plugin rack. Each one has an input signal and an output signal that needs to optimally be within the range that the next plugin is expecting. So Fab Filter may not be negatively affected by a signal that hits -1 dB, but maybe the next plugin in the chain is tape emulation plugin that decimates the sound at -1 dB. You have to be mindful of EACH plugin's thresholds.

This optimal threshold is typically: General digital plugins: Anything under around -0.3 dBFS Analog emulating plugins: -18 dBFS to -12 dBFS

^^^ But again. Your ears are king. If you like the way your analog emulating plugin sounds at whatever level you're feeding it, you win. Don't take any of these as unbreakable "rules" if you love what you're hearing. Sometimes, it's fun as hell to push a signal into the depths of hell to see what sound comes out of that shiny plugin!

Funny enough, I recently read the UAD manual for the ATR 102 tape machine and learned that it treats -12 dBFS as unity gain (0dB) for the tape (saturation). These things are pretty important to understand and nothing will teach you more than going through each manual on your own. Also a lot of these analog emulating plugins have a knob that you can adjust which changes their "optimal" level (called "headroom"), because they can work magic when this is all being processed in a computer with no physical limitations in processing. So I can configure that virtual ATR 102 tape machine to treat -6 dBFS as unity for the plugin. The magic of plugins!

STOP ARGUING ABOUT GAIN STAGING NON-SENSE AND READ YOUR DAMN PLUGIN MANUALS !!!

------------------------------------------ 4. Out the track and into busses ------------------------------------------

Your signal level out of your last FX plugin is then completely controlled by your track faders. In most cases this is the last place you can make a level adjustment on the track by itself.

There are other ways to route audio, via things like "sends", but they aren't really important to this conversation. As a bonus, I'll explain pre vs post sends at the end.

Each bus also has "gain" and "volume" (fader) controls. And these can be thought of in the same way we looked at track / clip gain. The bus gain is the first level adjustment of everything that routes into the bus. This starts at unity (0 dB) gain meaning that the level is unchanged. Whatever the sum of the tracks feeding the bus is, is what the bus sees.

And here is where you can use the trick I mentioned way back in section 2.

And your busses may feed other busses.

At each stage, you should NEVER clip. Again, there are reasons why you can recover clipping in the digital realm, but as a best practice, your workflow should attempt to avoid clipping as a standard approach. This is a workflow decision, not a "you're screwed if you don't do this". It's nuanced.

Eventually, all paths lead to the master bus (or at least it should. If it doesn't, I highly recommend you change your workflow. You should only bounce from "master" not the audio interface output channel).

And your master bus also has its own gain and volume (fader).

The master bus is a little special. Since it is the LAST thing before you bounce your audio, I recommend you NEVER EVER EVER EVER EVER change your master fader from unity gain (0dB). It's the only way to get an accurate reading from your mastering plugins, which if you've been paying attention, are processed BEFORE the master fader. What good is it to hit a certainly LUFS reading, if your master fader is changing the level right before the audio is bounced? You've shot yourself in the foot.

So again. Don't. Touch. Your. Master. Fader.

Futz around with the gain knob all you want. All that will do is change the level going into your plugins on your master bus, which might be very beneficial to how your mastering plugins react. (Re-read section 3 if this is confusing).

So… TLDR; the moral of the story about gain staging…..

There are technical nerd reasons that we could argue about all day as to how the audio may actually end up okay even after clipping within the DAW. BUT! You better understand exactly what your doing and choose to make these workflow decisions intentionally. I never recommend a newbie to mixing rely on DAW internal headroom to fix poor workflow habits. There's a lot of potential negatives and not a lot of positives.

Hopefully this clears some things up. A lot of people who throw around the term "gain staging" have no clue wtf they're talking about. They just want to sound cool.

Appendix A: Auxiliary Send Types:

-------- Post-fader send (default) --------

Post fader is the most likely mode you'll use. Once you've set a good balance between your raw source track and the aux bus send, moving your fader up and down will control BOTH levels evenly (your track volume and the level of your send). So you'll still have the same relative amount between the two.

-------- Pre-fader send --------

This is primarily for when you do things like "automating a reverb to make it sound like it's getting closer or further away over time". We can discuss that another time.

The pre-fade send will send the same amount of the source track to your aux send, regardless of the track fader. This means that if your send is at 0dB and you drop your source track fader level to -INF (silent), the full level of the source track is still being sent to the "send".

This gives you fine tuned control over the level sent to your FX bus, but sacrifices complexity. You now no longer can easily bring the level of the source track and the FX up or down evenly, like you could with a post-fader send.

This is also commonly used when making monitor mixes, so track fader adjustments don't change the balance of the monitor mix.

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

21 views0 comments

Recent Posts

See All