Let's Get Physical: The Complete Guide to Control Surfaces

When it comes to recording, let’s get physical

By Craig Anderton

Until digital recording appeared, every function in analog gear had an associated control: Whether you were tweaking levels, changing the amount of EQ gain, or switching a channel to a particular bus, a physical device controlled that function.

Digital technology changed that, because functions were no longer tied to physical circuits, but virtualized as a string of numbers. This gave several advantages: Controls are more expensive than numbers, so virtualizing multiple parameters and controlling them with fewer controls lowered costs. Virtualization also saved space, because mixers no longer had to have one control per function; they could use a small collection of channel strips—say, eight—that could bank-switch to control eight channels at a time.

But you don’t get something for nothing, and virtualization broke the physical connection between gear and the person operating the gear. While people debate the importance of that physical connection, to me there’s no question that having a direct, physical link between a sound you’re trying to create and the method of creating that sound is vital—for several reasons.

THE ZEN OF CONTROLLERS

If you’re a guitar player, here’s a test: Quick—play an A#7 chord. Okay, now list the notes that make up the chord, lowest pitch to highest.

Chances are you grabbed the A#7 instantly, because your fingers—your “muscle memory”—knew exactly where to go. But you probably had to think, even if only for a second, to name all the notes making up the chord.

Muscle memory is like the DMA (Direct Memory Access) process in computers, where an operation can pull data directly from memory without having to go through the CPU. This saves time, and lets the CPU concentrate on other tasks where it truly is needed. So it is with controllers: When you learn one well enough so that your fingers know where to go and you don’t have to parse a screen, look for a particular control, click it with your mouse, then adjust it, the recording process become faster and more efficient.

IMPROVING DAW WORKFLOW

Would you rather hit a physical button labeled “Record” when it was time to record, or move your mouse around onscreen until you find the transport button and click on it? Yeah, I thought so.

The mouse/keyboard combination was never designed for recording music, but for data entry. For starters, the keyboard is switches-only—no faders. The role of changing a value over a range falls to the mouse, but a mouse can do only one thing at a time—and when recording, you often want to do something like fade one instrument down while you fade up another.

Sure, there are workarounds: You can group channels and offset them, or set up one channel to increase while the other decreases, and bind them to a single mouse motion. But who wants to do that kind of housekeeping when you’re trying to be creative? Wouldn’t you rather just have a bunch of faders in front of you, and control the parameters directly?

Another important consideration is that your ears do not exist in a vacuum; people refer to how we hear as the “ear/brain combination,” and with good reason. Your brain needs to process whatever enters your ears, so the simple act of critical listening requires concentration. Do you really want to squander your brain’s resources trying to figure out workarounds to tasks that would be easy to do if you only had physical control? No, you don’t. But . . .

PROBLEM 1: JUST BECAUSE SOMETHING HAS KNOBS DOESN’T GUARANTEE BETTER WORKFLOW

Some controllers try to squeeze too much functionality into too few controls, and you might actually be better off assigning lots of functions to keyboard shortcuts, learning those shortcuts, then using a mouse to change values. I once used a controller for editing synth parameters (the controller was not intended specifically for synths, which was part of the problem), and it was a nightmare: I’d have to remember that, say, pulse width resided somewhere on page 6, then remember which knob (which of course didn’t have a label) controlled that parameter. It was easier just to grab a parameter with a mouse, and tweak.

On the other hand, a system like Native Instruments’ Kore is designed specifically for controlling plug-ins, and arranges parameters in a logical fashion. As a result, it’s always easy to find the most important parameters, like level or filter cutoff.

PROBLEM 2: IT GETS WORSE BEFORE IT GETS BETTER

So do you just get a controller, plug it in, and attain instant software/hardware nirvana? No. You have to learn hardware controllers, or you’ll get few benefits.

If you haven’t been using a controller, you’ve probably developed certain physical moves that work for you. Once you start using a controller, those all go out the window, and you have to start from scratch. If you’re used to, say, hitting a spacebar to begin playback, it takes some mental acclimation to switch over to a dedicated transport control button.

Which begs the question: So why use the transport control, anyway? Well, odds are the transport controls will have not just play but stop, record, rewind, etc. Once you become familiar with the layout, you’ll be able to bounce around from one transport function to another far more easily than you would with a QWERTY keyboard set up with keyboard shortcuts.

Think of a hardware controller as a musical instrument. Like an instrument, you need to build up some “muscle memory” before you can use it efficiently. I believe that the best way to learn a controller is to go “cold turkey”: Forget you have a mouse and QWERTY keyboard, and use the controller as often as possible. Over time, using it will become second nature, and you’ll wonder how you got along without it. But realistically, that process could take days or even months; think of spending this time as an investment that will pay off later.

DIFFERENT CONTROLLER TYPES

There are not just many different controllers, but different controller product “families.” The following will help you sort out the options, and choose a controller that will aid your workflow rather than hinder it. 

Custom controllers. These are designed to fit specific programs or software like a glove; examples include Ableton's Push controller, Roland’s V-Studio series (including the 700, 100, and 20 controllers), Steinberg’s Cubase-friendly series of CMC controllers, and the like. The text labels are usually program-specific, the knobs and switches have (hopefully) been laid out ergonomically, and the integration between hardware and software is as tight as Tower of Power’s rhythm section. If a control surface was made for a certain piece of software, it’s likely that will be the optimum hardware/software combination.

Ableton's Push controller is an ideal match for Live 9

A different type of controller, Softube's Console 1, is a different type of animal—it has software that emulates an analog channel strip and inserts in a DAW, with a hardware controller that provides a traditional, analog-style one-function-per-control paradigm. The control surface itself provides visual feedback, but if you want more detail, you can also see the parameters on-screen.

Softube's Control 1

General-purpose DAW controllers. While designed to be as general-purpose as possible, these usually include templates for specific programs. They typically include hardware functions that are assumed to be “givens,” like tape transport-style navigation controls, channel level faders, channel pan pots, solo and mute, etc. A controller with tons of knobs/switches and good templates can give very fluid operation. Good examples of this are the Mackie Control Universal Pro (which has become a standard—many programs are designed to work with a Mackie Control and many hardware controllers can emulate the way a Mackie Control works), Avid Euphonix Artist series controllers (shown in the opening of this article), and Behringer BCF2000.

Mackie Control Universal Pro

There are also “single fader” hardware controllers (e.g., PreSonus FaderPort and Frontier Design Group AlphaTrack) which while compact and inexpensive, take care of many of the most important control functions you’ll use.

Digital mixers. For recording, a digital mixer can make a great hands-on controller if both it and your audio interface have a multi-channel digital audio port (e.g., ADAT optical “light pipe”). You route signals out digitally from the DAW, into the mixer, then back into two DAW tracks for recording the stereo mix. Rather than using the digital mixer to control functions within the program, it actually replaces some of those functions (particularly panning, fader-riding, EQ, and channel dynamics). As a bonus, some digital mixers include a layer that converts the faders into MIDI controllers suitable for controlling virtual synths, effects boxes, etc.

Synthesizers/master keyboards. Many keyboards, like the Yamaha Motif series and Korg Kronos, as well as master controllers from M-Audio, Novation, CME, and others build in control surface support. But even those without explicit control functions can sometimes serve as useful controllers, thanks to the wheels, data slider(s), footswitch, sustain switch, note number, and so on. As some sequencers allow controlling functions via MIDI notes, the keyboard can provide those while the knobs control parameters such as level, EQ, etc.

Arturai's KeyLab 49 is part of a family of three keyboard controllers that also serve as control surfaces.

Really inexpensive controllers. Korg's nanoKONTROL2 is a lot of controller for the money; it's basic, with volume, pan, mute, solo, and transport controls, but it's also Mackie-compatible.

But if you're on an even tighter budget, remember that old drum machine sitting in the corner that hasn’t been used in the last decade? Dust it off, find out what MIDI notes the pads generate, and use those notes to control transport functions—maybe even arm record, or mute particular track(s). A drum machine can make a compact little remote if, for example, you like recording guitar far away from the computer monitor.

The “recession special” controller. Most programs offer a way to customize QWERTY keyboard commands, and some can even create macros. While these options aren’t as elegant as using dedicated hardware controllers, tying common functions to key commands can save time and improve work flow.

Overall, the hardware controllers designed for specific software programs will almost certainly be your best bet, followed by those with templates for your favorite software. But there are exceptions: While Yamaha’s Motif XS and XF series keyboards can’t compete with something like a Mackie Control, they serve as fine custom controllers for Cubase AI—which might be ideal if Cubase is your fave DAW.

Now, let’s look at some specific issues involving control surfaces.

MIDI CONTROL BASICS

Most hardware control surfaces use MIDI as their control protocol. Controlling DAWs, soft synths, processors, etc. is very similar to the process of using automation in sequencing programs: In the studio, physical control motions are recorded as MIDI-based automation data, which upon playback, control mixer parameters, soft synths, and signal processors.

If you’re not familiar with continuous controller messages, they’re part of the MIDI spec and alter parameters that respond to continuous control (level, panning, EQ frequency, filter cutoff, etc.). Switch controller messages have two states, and cover functions like mute on/off.

There are 128 numbered controllers per MIDI channel. Some are recommended for specific functions (e.g., controller #7 affects master volume), while others are general-purpose controllers.

Controller data is quantized into 128 steps, which gives reasonably refined control for most parameters. But for something like a highly resonant filter, you might hear a distinct change as a parameter changes from one value to another. Some devices interpolate values for a smoother response.

MAPPING CONTROLS TO PARAMETERS

With MIDI control, the process of assigning hardware controllers to software parameters is called mapping. There are four common methods:

Novation's low-cost Nocturn controller features their Automap protocol, which identifies plug-in parameters, then maps them automatically. In this screen shot, the controls are being mapped to Solid State Logic's Drumstrip processor for drums.

“Transparent” mapping. This happens with controllers dedicated to specific programs or protocols: They’re already set up and ready to go, so you don’t have to do any mapping yourself.

Templates. This is the next easiest option. The software being controlled will have default controller settings (e.g., controller 7 affects volume, 10 controls panning, 72 edits filter cutoff, etc.), and loading a template into the hardware controller maps the controls to particular parameters.

MIDI learn. This is almost as easy, but requires some setup effort. At the software, you select a parameter and enable “MIDI learn” (typically by clicking on a knob or switch—ctrl-click on the Mac, right-click with Windows). Twiddle the knob you want to have control the parameter; the software recognizes what’s sent and maps it.

Fixed assignments. In this case, either the controller generates a fixed set of controllers, and you need to edit the target program to accept this particular set of controllers; or, the target software will have specific assignments it wants to see, and you need to program your controller to send these controllers.

THE “STAIR-STEPPING” ISSUE

Rotating a “virtual front panel” knob in a soft synth may have higher resolution than controlling it externally via MIDI, which is limited to 128 steps of resolution. In practical terms, this means a filter sweep that sounds totally smooth when done within the instrument may sound “stair-stepped” when controlled with an external hardware controller.

While there’s no universal workaround, some synthesizers have a “slew” or “lag” control that rounds off the square edges caused by transitioning from one level to another.

RECONCILING PHYSICAL AND VIRTUAL CONTROLS

Controllers with motorized faders offer the advantage of having the physical control always track what the corresponding virtual control is doing. But with any controller that doesn’t use motorized faders, one of the big issues is punching in when a track already contains control data. If the physical position of the knob matches the value of the existing data, no problem: Punch in, grab the knob, and go. But what happens if the parameter is set to its minimum value, and the knob controlling it is full up? There are several ways to handle this.

Instant jump. Turn the knob, and the parameter jumps immediately to the knob’s value. This can be disconcerting if there’s a sudden and unintended change—particularly live, where you don’t have a chance to re-do the take!

Match-then-change. Nothing happens when you change the physical knob until its value matches the existing parameter value. Once they match, the hardware control takes over. For example, suppose a parameter is at half its maximum value, but the knob controlling the parameter is set to minimum. As you turn up the knob, nothing happens until the knob matches the parameter value. Then as you continue to move the knob, the parameter value follows along. This provides a smooth transition, but there may be a lag between the time you start to change the knob and when it matches the parameter value.

Add/subtract. This technique requires continuous knobs (i.e., data encoder knobs that have no beginning or end, but rotate continuously). When you call up a preset, regardless of the knob position, turning it clockwise adds to the preset value, while turning it counter-clockwise subtracts from the value.

Motorized faders. This requires bi-directional communication between the control surface and software, as the faders move in response to existing automation values—so there’s always a correspondance between physical control settings and parameter values. This is the great: Just grab the fader and punch. The transition will be both smooth and instantaneous.

Parameter nulling. This is becoming less common as motorized faders become more economical. With nulling, there are indicators (typically LEDs) that show whether a controller’s value is above or below the existing value. Once the indicators show that the value matches (e.g., both LEDs light at the same time), punching in will give a smooth transition.

IS THERE A CONTROLLER IN YOUR FUTURE?

Many musicians have been raised with computers, and are perfectly comfortable using a mouse for mixing. However, it’s often the case that when you sit that person down in front of a controller, and they start learning how to actually use it, they can’t go back to the mouse. In some ways, we’re talking about the same kind of difference as there is between a serial and parallel interface: The mouse can only control one parameter at a time, whereas a control surface lets you move groups of controls, essentially turning your mix from a data-entry task into a performance. And I can certainly tell you which one I prefer!

Craig Anderton is Executive Editor of Electronic Musician magazine. He has played on, mixed, or produced over 20 major label releases (as well as mastered over a hundred tracks for various musicians), and written over a thousand articles for magazines like Guitar Player, Keyboard, Sound on Sound (UK), and Sound + Recording (Germany). He has also lectured on technology and the arts in 38 states, 10 countries, and three languages.