Gain-Staging is one the most important aspects of audio signal flow, even though many people are unfamiliar with the concept. It’s the practice of insuring that audio signals are at the optimum level at all times, as they pass through the various processors and mixer stages. Of course, professional engineers have always known how critical this is in the analog world, to guarantee a clean signal at the end, free of excess noise or overload distortion. But given the extremely high headroom of modern DAWs (thanks to their internal 32-bit floating-point resolution), engineers and mixers often assume that they no longer have to worry about these concerns. However, the practice of careful gain-staging is still an important part of any recording or mix session—if it’s not handled properly, even with the expanded headroom of a modern DAW, problems can occur. Here are five basic gain-staging mistakes you’ll want to avoid.
Most plug-ins have input and output gain controls that are provided to allow the user to adjust the signal level coming into the processor, and the level exiting the plug-in. By default, these will always be set at Unity Gain, and they don’t necessarily have to be changed—if the incoming signal is at a suitable level, and the processing doesn’t drastically change that level of the outgoing signal, then it’s best to just leave them alone. But often, the incoming signal won’t be at the best level, and the plug-in’s processing may alter that level to a significant degree, and that’s when you’ll want to utilize these level adjustments.
There are various reasons why an incoming signal might be at an inappropriate level. The level might have been changed at a previous stage in the processing chain, or the audio file may have been recorded at too high, or too low, a level to begin with. An especially hot signal may be more than a certain plug-in is expecting, and reducing the input gain would yield better processing results—a good example would be a compressor, or any dynamics processor. If you’re used to dialing up, say, certain Threshold settings that you know should work well, a signal that’s abnormally hot or low in level may not react as expected, making the job of getting just the right degree of effect that much more difficult. When the incoming level is a problem, you’ll want to fix that at the first step in the plug-in chain, utilizing any Input Gain control the plug-in provides (although some DAWs also have a Region/Clip Gain feature, which allows non-destructive adjustment of the audio file’s level on a per-region/clip basis, before it hits the first insert slot—another way to properly gain-stage the signal). Once the optimum input level has been set, you’ll want to maintain that level from insert to insert, so any time you remove or bypass an effect, the level of that track in the mix remains consistent.
Another good example of the need for this involves the use of EQ. When you set a large boost or cut, not only does the tone change, but the overall level is altered as well. Ideally, you’ll want to be able to use the Bypass button to compare the original and the processed versions of the signal, but a big level difference would make that kind of direct A/B’ing ineffective (the louder version always seems to sound better, at first blush). Not to mention, if the level of the track in the mix has already been set, then a noticeable level change from drastic EQ will ruin that balance. You could tweak the level at the Channel Fader, but then if you decide to momentarily bypass, or remove the effect for any reason, the mix is again out of whack. Using any Output Gain control the plug-in offers will adjust for the level change incurred within that processor at the proper place in the chain.
That’s what gain-staging is, in practice—optimizing the level at the input and output of each step in the processing chain (real or virtual). In the digital domain, you may not necessarily suffer noise and distortion issues if you fail to do this, but it’s good practice, and will ensure a smoother flow as a session or mix progresses.
To continue with the theme of using a plug-in’s Input and Output level knobs, you’ll want to be aware of the difference between those kind of general-purpose gain-staging controls, and more specific gain controls that may be part of a plug-in’s processing section itself. The most obvious example of this would be guitar amp simulators. Modeled after real amps, they’ll usually have an input level control, typically labeled Gain, or Drive, and an output level control, often labeled Master Volume. But these two gain knobs are part of the effect—they’re designed not to be used for gain-staging, but to push the signal into the virtual overdrive sections of the model (the simulated preamp and poweramp tubes, respectively).
In fact, to get the desired, and expected, response (one that effectively simulates the range of effect of the original amp), you’ll want to adjust the incoming signal before it hits the model’s gain controls, to insure that the plug-in sees an incoming signal at the typical level a guitar signal would be at in the real world (many amp sim makers, especially the ones whose plug-ins also include separate Input and Output controls for gain-staging purposes, will provide recommendations for proper input levels in their documentation, to ensure that users will get the best sound out of the modeled amps).
This one may seem pretty obvious, but not always to newbies. I’ve often observed novice mixers set up a standard Send & Return hookup, with several tracks feeding the same Reverb, in an Aux channel (Return), via Sends. Then, the mixer notices that, say, there’s too much reverb on the drums, so, he reaches for the Aux fader—the Return—and pulls it down a bit, forgetting that this will alter the Reverb level for all the other tracks that are feeding it as well!
Like I said, to experienced mixers, it’s an obvious mistake, but for someone new to the art, it may take some time to get used to the need to reach for an individual Send, to adjust the level of Reverb for that one instrument in the Return (Aux). When teaching recording and mixing, it’s clear the setup and use of Send & Return hookups is one of the most difficult concepts for novices to get a handle on, so it’s definitely worth the mention. Unless you’re dialing up a spatial special effect, it’s always better to get used to leaving the Aux fader (for a Return) at Unity Gain, and making any & all level adjustments elsewhere.
Just as with Aux faders, the Master Fader (Mix Bus, Stereo Output, it goes by many names) should also be left at Unity Gain. It can sometimes be tempting to use it as a convenient volume control to adjust listening level, but you should avoid this studiously! Not only will you no longer be looking at a useful readout of the overall level of the mix, it’s too easy to forget and accidentally bounce the track at an attenuated level, which may go unnoticed until some inconvenient moment when the mistake proves embarrassing (like handing it off to a client expecting a hot track).
Fig 4 The Master fader should always be left at Unity Gain
Really, the only time you should adjust the Master Fader is to perform a fade, and even that is sometimes done with a subgroup channel fader. And if you’re prepping a 16-bit master for CD duplication/distribution (yes, people still buy CDs!), with a Dither plug-in correctly added in the last insert slot, you’ll want to be especially careful not to deviate from Unity Gain with the Master Fader, which would negate the benefits of the Dither.
Finally, a word about those overload indicators that may get to flashing from time to time. As I mentioned earlier, many people know that the internal 32-bit floating-point processing resolution of (all) modern DAWs pretty much precludes the possibility of digital overloads, in most situations, and they’ve gotten in the habit of ignoring the red overload indicators on channel strips and in plug-ins. But that’s not good practice—even though you may be able to get away with it 90% of the time, you should still employ proper gain-staging, and try to avoid overload indications at every stage of the processing chain. There are a couple of good reasons for this...
First off, you always want to avoid red overload lights going off in the Master Fader/Main Stereo Output channel strip. Even though you won’t hear any digital clipping while you’re still in the (32-bit world of the) DAW, when you do the final Bounce to Disk, that floating-point headroom vanishes, and the resulting audio file will be clipped, probably necessitating a redo. The occasional overage light in a Channel Strip can probably be safely ignored, if the Master Output is clean, but it’s still a good idea to keep an eye on the overload indicators for the ins and outs of each plug-in in the processing chain. Again, as I touched on above, many (a great many!) virtual processors emulate the warmth and presence of analog gear, and to enjoy the best they’re capable of, you need to feed them clean signals, that have been gain-staged to properly emulate the expected levels of real analog signals arriving at these processors’ inputs. Often, the code that emulates the desired analog drive will also emulate the unpleasant distortion those devices would generate if fed with an analog signal at the kind of excessive level that digital signals could otherwise get away with. To get the best out of these plug-ins, you’ll want to gain-stage as carefully as you’d need to if you were actually working in the analog domain. And for those who actually do occasionally exit the workstation (via i/o plug-ins), to run signals though an actual physical device (during an otherwise virtual mix), observing proper gain-staging is the only way to get the audio in and out of the DAW cleanly.
So, despite the temptation to ignore gain-staging in the forgiving world of the modern DAW, observing best practices when it comes to level—analog or digital—is the way to ensure the best results at every stage of production.