Computer-based recording can make it all too easy to miss a poorly thought-out signal chain
December 16, 2013, by Daniel Keller
In the pre-DAW era, when we had to rewind heavy two-inch multitrack tapes uphill in the snow, both ways, the concept of gain structure was fairly easy to grasp.
Instead of a virtual studio-in-a-box, each individual component was a link in the audio chain, visibly connected by patch cables, and analog distortion was fairly easy to hear, identify and track down.
In today’s digital world, it’s no longer that simple. Signal paths can be virtual, unconventional, and convoluted, and digital distortion can be subtle, sneaking up on you only after it’s too late to fix.
But while digital audio has fundamentally changed the way we record and mix, proper gain structure is no less critical to a good sounding recording.
The user-friendly, preset-heavy paradigm of computer-based recording can make it all too easy to miss a poorly thought out signal chain, leaving you with some nasty digital artifacts in the final mix, long after your star players have gone home or logged off.
Nothing (And Everything) To Gain
From its initial capture to its place in the final mix, every sound in a typical multitrack recording travels though a multitude of stages and devices. Each one of these stages, whether hardware or software, requires an optimal signal level at its input. Too high an input level can cause clipping and distortion; too low a level can add noise. Keeping an eye on the input and output levels of every stage in the signal chain is critical to ensuring that each device’s output feeds a clean signal to the next device’s input.
With the exception of devices connected via channel inserts, input levels for most plug-ins are controlled via the (hardware or software) mixer’s Aux or FX Send. Most plug-ins also have their own input level, so it’s important to look at both, and match the mixer’s FX send level with the plug-in’s input level for proper gain structure.
Sending too low a level to the effects bus and then turning up the plug-in’s input level to compensate will result in a nosier signal. Conversely, sending too hot an effects send level and then turning down the plug-in’s input level will result in a distorted signal. In most cases, it’s best to start with both levels set to 0 or unity gain.
In fact, with the exception of compressors and other dynamics processors, unity gain is a good rule of thumb – your signal level should ideally be unchanged whether the plug-in is inserted into the chain or not. If your levels are noticeably higher or lower when you bypass a particular device, it’s a good idea to go back and examine the gain structure of that link.
If A Signal Clips In The Forest…
Overloading can be particularly problematic in the digital domain. Raise the input signal to an analog device, and distortion will gradually rise until clipping occurs. Digital circuitry has no such safety zone – a single dB too high will take your signal from squeaky clean to nasty distortion.
Unlike analog distortion, digital clipping can be difficult to hear, particularly when it’s just one element of a complex multitrack mix. If the clipping goes undetected, the digital information for that sound is permanently corrupted, even if the levels are brought back down later in the mix. And digital distortion can have a subtle but undesirable effect on the sonic quality of your track, usually in the form of barely perceptible levels of a brittle, harsh digital sheen that can fatigue your listeners.
Even a relatively minimal amount of gain from certain plug-ins – for example, a high-pass filter – can boost peaks and transients pretty significantly. Don’t depend on your meters to alert you to these, either.
In most DAW setups, plug-in inserts occur pre-fader, so even if you keep the levels of your channel strips below clipping, distortion within a given plug-in may not show up if the level was brought back down further along the signal chain. As always, your ears are your most important tools. Solo each device and listen.
What To Listen For
Needless to say, different types of signal processors will affect overall gain structure differently, and some are easier to work with than others.
Digital reverbs are particularly hazardous – the “soft” nature of some reverb algorithms, combined with reverb’s typical in-the-background role, can mask other artifacts, including noise resulting from too low an input level.
Multiband EQ can be particularly nefarious, especially when it comes to peaks and transients.
With modern multiband EQ plug-ins, it’s not hard to inadvertently overlap a range of frequencies in two different bands, and the cumulative boost can result in clipping.
Compression and dynamics processors, being gain-based concepts, present a whole different set of challenges, and deserve a column of their own (maybe next month).
Briefly, though, it’s important to pay attention to a compressor’s attack and gain settings, as these can have a major impact on gain structure of the signal coming out of your compressor.
Get To Know Your Plug-Ins
Just as every guitar, keyboard, and vintage amp has its own sonic character, so too does every signal processor. This is no less true for software plug-ins than it is for hardware units. Different devices have different ways in which they handle gain and clipping. And getting a good sense of how each of your plug-ins performs in different situations is as important as knowing any other instrument in your arsenal.
In the analog era, engineers would test each new box by running a sine wave through it and looking at the signal on an oscilloscope. They could see where each device would clip at specific frequencies, what kind of distortion would occur, and other characteristics that helped to map out the device’s optimal gain settings and place in the chain.
You can easily do the same thing with your frequently used plug-ins. Open an oscilloscope or frequency display in your DAW, set the plug-in’s input (and output, if it has one) level to unity gain, and send a sine wave through the device. Watch the output as you gradually raise the send level.
Not Exciting, But It’s Real
Of course, the geek factor of testing with sine waves is no substitute for listening. Many tracks in a mix will have multiple plug-ins inserted in their signal path, and it’s important to check each individual track, inserting each plug-in one at a time, and then all together. Things may sound find in the overall mix, but only critical listening will tell you if something’s not quite right.
In the busy hubbub of live recording, particularly if you’re tracking a whole band, picky details like gain structure can be easy to overlook. But it’s a necessary fact of life, and ignoring it is not an option.
The more familiar you are with gain structure in general and your equipment in particular, the less time you’ll have to spend optimizing levels when you’ve got a room full of antsy musicians waiting to lay down tracks.
Daniel Keller, a.k.a., the Studio Curmudgeon, is a musician, engineer and producer. Since 2002 he has been president and CEO of Get It In Writing, a public relations and marketing firm focused on audio and multimedia professionals and their toys. Despite being immersed in professional audio his entire adult life, he still refuses to grow up.