Building A Live Rig, Part 2 – Multiple Keyboards

Using more than one keyboard sounds simple enough. Set the keyboards up, plug them both into a mixer or into separate DIs, and play. But wait, what if you need to use a sustain pedal for parts you play on both keyboards? Now you need two pedals. And what if you have a two-tier stand so one keyboard is on top? Now the pedals need to be beside each other, making it harder to remember which is which. And what if one of your keyboards is really nice to play – a fully-weighted stage piano, for example – while another keyboard is a synth-action workstation that has a lot of sounds but feels cheap? Is there a way to make the workstation keyboard sound while you play the notes on the stage piano instead?

Why Use Multiple Keyboards?

There’s really only two reasons to use more than one keyboard. One is as an alternative to using a keyboard split. A 61-key keyboard won’t give you a lot of space on either side of the split; it’s only 5 octaves, plus you’ll probably end up with some ‘wasted space’ or run out of room unless your split works nicely with the range of keys on your keyboard (i.e. most keyboards range from C to C; if you want a two-octave range and a three-octave range, you can’t do that in G.) If you have two keyboards, you can have each hand on a different keyboard, each playing a different sound. You’ll probably want a tiered stand to do this, as opposed to having two keyboards at right angles to each other, and you’ll probably find it more comfortable to play standing up, though that makes using more than one pedal a bit tricky.

The other reason for multiple keyboards is because one keyboard just doesn’t give you all the sounds you want. I use both a workstation and a stage piano with drawbar-controlled B3 organ because the organ is a big part of my sound and I need the level of control that drawbars give; the few B3 organ presets on my workstation, while quite good, didn’t have the variety. A lot of rock and pop bands would tour with a large number of keyboards, since they wanted to reproduce the studio recording as closely as possible. If you’ve got a small army of roadies, bringing along a number of keyboards that will each be used for one sound on one song might be ok, but for the rest of us, simplicity is better.

Other than these reasons, there’s no good reason to be bringing multiple keyboards, and with the right choices, I’d argue that you only need two period, and if you still need extra sounds, some rackmount sound modules or a laptop are lighter, smaller, and easier to use.

The Master Keyboard

As I mentioned, you could just use two keyboards, completely independently, in your live setup. This complicates things, though; you not only need to remember which keyboard to play when, but you also have a more complicated process for switching presets. Even if you’ve followed my suggestion about putting all your presets in order, it will still take at least two button presses or footswitch presses to change presets on both keyboards. And what if you have more than two keyboards, or you add a rackmount sound module for some more or better sounds?

There are only two ways to effectively deal with multiple keyboards (if you don’t have a keyboard tech to send program changes via MIDI at the right times, of course). If you’re able to put all the presets in performance order on EVERY keyboard you’re using, you could get an independent pedal that sends MIDI program change messages (I use the Tech21 MIDI Mouse for this) and use a MIDI hub or some other type of MIDI THRU unit to send the program change to all keyboards.

The ideal way is to designate one keyboard as your ‘master keyboard’. This need not be the larger keyboard or even the one you play more often (I play as much as possible on my semi-weighted 73-key Roland, not my 61-key synth action Korg, which I use as my master). The only requirement is that the master keyboard can send a number of MIDI program changes (and bank changes, if your gear needs it) to other keyboards. My Korg has a combi mode where you can use eight timbres (individual sounds or voices) at once. Some of those can be external timbres, that send MIDI data to other pieces of gear, including program changes. When preparing a new song, I make a Korg combi for each structural part of the song, and the combi triggers a MIDI program change on my Roland. Other timbres in the combi are used for the sounds I want the Korg to play for that part of the song. Other workstations will have something similar.

One other thing to consider when choosing a master keyboard is clock sync, especially if you’ll be playing sequenced backing tracks, pre-recorded phrases, using an arpeggiator, or even things like tempo-synced delays. One of your keyboards should be sending a MIDI clock signal to the other keyboards to keep them in sync, and the other keyboards should only be receiving clock data and not sending it out. Any decent workstation should let you set a tempo and work as either a master or slave to the tempo. If you’re going to be using anything that’s dependent on keeping time, make sure to have this worked out for your gear.

I mentioned that I also use the MIDI Mouse pedal and that it seems to be used for the less ideal setup. Actually, I don’t use it when I just have two keyboards because the Korg doesn’t send program changes out when it receives an external program change (plugging a footswitch into the Korg and setting its behaviour to ‘Next Program’ does send out the program changes stored in my combi). Instead, when I’m using both keyboards AND my laptop, the laptop acts as the master device, sending the program changes to both keyboards, and I control it with the MIDI Mouse.

Any Sound, Any Keyboard

If you’re going to use multiple keyboards, be sure to arrange your sounds so that what you’re playing at a time is as easy and convenient to play as possible. You can do much more with two keyboards than just having each keyboard only play its own sounds. Maybe one of your keyboards has great B3 organ sounds but it’s an 88-key weighted keyboard; not good for organ slides. If you use a semi-weighted or synth-action keyboard to trigger organ sounds on the weighted keyboard, you get the action you want and the sounds you want. Or maybe you could even layer sounds from both keyboards together – one has great strings and the other’s a synthesizer so you make a warm pad patch and layer it with your strings. Or perhaps you need a piano and a pad, one played with each hand, and the best piano and pad sounds are both on the same keyboard. You can trigger the pad from the other keyboard so you don’t have to do a split on the one keyboard to make two independent parts.

Rule 5: In an ideal keyboard setup, pressing the keys on any keyboard should be able to make any keyboard sound.

If you can follow rule 5, which is really dependent on having the right gear and knowing how to set it up, then you’ll have a really flexible setup that goes a long way to getting the technology out of the way of your musicality. This also lets you make the most use of layering: you can not only combine sounds on a single keyboard, you can mix sounds between keyboards and still play them from one single keyboard.

MIDI Channels

In order to get anywhere with your setup, you need to know a bit about MIDI, starting with MIDI Channels. Think of MIDI messages as if they were pieces of mail. A MIDI channel is like an address on the letter. There are 16 MIDI channels, and nearly every MIDI message is sent on a specific channel. Anything listening to the MIDI data will usually be listening to one or more channels (some devices have a setting to listen to all channels, but you usually don’t want this).

Let’s start with an example. You’ve just unpacked two keyboards and plugged them in, connecting a Korg’s MIDI OUT to a Roland’s MIDI IN (this means that any data going OUT of keyboard A will go IN to keyboard B), and A’s OUT to B’s IN. You pick a string sound on the Korg and a piano sound on the Roland, and you press a key on the Korg. What happens? Probably, you’ll hear both the strings and the piano. Most keyboards, by default, listen and transmit on MIDI Channel 1. So, you pressed the key on the Korg, which did two things: make the Korg’s string sound play, and send a Note On message to the Korg’s MIDI OUT, using Channel 1. Since that MIDI OUT is connected to the Roland, the Roland gets the Note On message. The message is on Channel 1, and the Roland is listening on Channel 1, so the piano sound is triggered as well. If you pressed a key on the Roland, the same thing would happen in reverse and you’d still hear both sounds.

Well, clearly we don’t want both keyboards playing at the same time, so let’s say we go into the settings for the Roland and have it use Channel 2 instead. What happens now? Each keyboard will still make its own sound (it usually takes some work for this not to happen), but since the Roland sends on Channel 2, the Korg will ignore the Note On since the Korg’s only listening on Channel 1. The Korg will also send a note on Channel 1 which the Roland will ignore since it’s only listening on Channel 2. This is actually how we want to have the global channels set up.

Rule 6: Each keyboard should listen and transmit, in general, on a different MIDI channel.

If Rule 6 isn’t followed, there’s no way for a keyboard to do all of: make itself sound, make the other keyboard sound, and make both sound at the same time.

To be most flexible, you need to use at least two different MIDI channels for each keyboard: one for that keyboard’s sounds and one that the keyboard listens to that is triggered by other keyboards. For example:

Roland (playing itself): Channel 1
Roland (playing the Korg): Channel 2
Korg (playing itself): Channel 3
Korg (playing the Roland): Channel 4

In this setup, you’d modify a combi on the Korg so that if the Korg should only play sounds on the Roland, there aren’t any internal sounds being triggered, and you have a voice triggering sounds externally on channel 4, so the Roland would sound. The Roland would of course listen to Channel 4. If the Korg should both play its own sounds and trigger sounds on the Roland, use the one voice for channel 4 plus one or more voices on channel 3. And if the Roland should also play both, use some voices on the Roland on channel 1 plus an external voice on the Roland sending MIDI data on channel 2.

You might wonder why the Roland uses two channels and not just one. Most keyboards always send MIDI data on their global channel, so it’s probably impossible to prevent the Roland from sending MIDI data on Channel 1. When you want to use the Roland to play its own sounds, you’ll be sending MIDI data on Channel 1. If you also use Channel 1 to control the Korg, there’s no way to selectively send MIDI to the Korg when you play the Roland.

Rule 7: No keyboard should listen to another keyboard’s global MIDI channel

With a third keyboard, say a Yamaha, you might want to add in Roland (played by the Yamaha): Channel 5, Korg (played by the Yamaha): Channel 6, and another three channels for playing the Yamaha from each of the three keyboards. I’d suggest not doing this. If you’re trying to recreate a keyboard rig from the 80′s and 90′s, where you’re bringing a small army of synthesizers to get the specific sounds you need, you’ll still probably only use each keyboard for a few specific sounds so you don’t need the complexity (though you’ll probably want a keyboard tech!) – and why not just play each specific keyboard and have it make its own sounds, and not bother with the extra hassle?

On the other hand, decide that your Roland and Korg are the keyboards you want to physically play, while the Yamaha is just there for the sounds (and could possibly be a rackmount sound module instead of a full keyboard). If so, you could also use the list above, but substitute the text “playing the X” for “playing everything else”, like this:

Roland (playing itself): Channel 1
Roland (playing everything else): Channel 2
Korg (playing itself): Channel 3
Korg (playing everything else): Channel 4
Yamaha (triggered by the Roland): Channel 2
Yamaha (triggered by the Korg): Channel 4

In this case, you might think there’s a problem: what if the Roland should play notes on the Korg but not the Yamaha? Easy; in the current preset on each keyboard, the Roland sends MIDI data on channel 2, Korg listens on channel 2, and the Yamaha does not listen on channel 2. Basically, each keyboard that you’re going to physically play should send MIDI data on two channels, and each keyboard that will produce sound should listen on one channel for each keyboard. Set up this way, we reach the goal of Rule 6: you can press a key on any keyboard and have any (or all) keyboards respond to it.

MIDI Connections

So you’ve decided which keyboards you want to physically play and which ones will make sounds. The next step is to actually connect them. With only two keyboards, the MIDI connections are easy: Keyboard A OUT -> Keyboard B IN, and Keyboard B OUT -> Keyboard A IN. With more than this, you’ll need something more complicated. One way is to get a MIDI THRU unit for each keyboard you’ll physically play (such as this or this), and connect as follows:

Keyboard A OUT -> THRU IN
THRU OUT 1 -> Keyboard B IN
THRU OUT 2 -> Keyboard C IN
…

This means that the MIDI data from Keyboard A is sent to keyboards B and C. Keyboard A can produce sounds from its own keys without any MIDI connections, of course.

A setup like this might start to get a bit unwieldy, so you might want to instead consider a MIDI hub, something like this. In addition to being an easy way to connect all of your keyboards to your computer, most/all of the MOTU MIDI hubs have a really nice feature: live keyboard mode. Say you’ve got a Korg, Yamaha, and Roland, and want to use both the keyboards and sounds on all three. You’d connect them to the hub like this, and turn on Live Keyboard Mode.

Korg OUT -> Hub IN 1
Hub OUT 1 -> Korg IN
Roland OUT -> Hub IN 2
Hub OUT 2 -> Roland IN
Yamaha OUT -> Hub IN 3
Hub OUT 3 -> Yamaha IN

What does Live Keyboard Mode do? It means that MIDI data received at an input is sent to all outputs BUT the one with the matching number. So, you play a note on your Roland, which is received at input 2. The note is sent to outputs 1, 3, 4, etc, but not output 2. It seems obvious that this mode means that one keyboard can play all the others, and the reason why you don’t want the keyboard to send MIDI data back to itself is simple. If you play a note on the Roland’s keyboard, the Roland will make a sound and will send out MIDI data. If that MIDI data feeds back in to the Roland, and it’s on a channel the Roland is listening to, this breaks Rule 7. More specifically, the Roland will try to trigger each note twice: once for physically pressing the key and once as a response to incoming MIDI data. Double-triggering a note is an audible effect and clearly not one you’d ever want.

Rule 8: MIDI data should never be fed back to the keyboard that generated it

Pitch Bend, Mod Wheels, and other Physical Controllers

When using physical controllers, it’s generally best to somehow filter them out from being sent or from being received, whichever works best with your gear. As a rule, I suggest using the physical controls on the keyboard you’re playing to modify the sound, regardless of which keyboard actually produces the sound. For example, use the pitch bend wheel on your Roland if you’re playing the Roland, even if the sound is coming from the Yamaha. If you intend to use pitch bend and other controls, you can’t have both hands on the keyboard playing, so if you use the controls on the Roland, you know you’re not adjusting another sound that’s playing. You might, of course, be adjusting another sound in the current preset that isn’t playing yet (for example, using the mod wheel might give you some vibrato on your synth lead but will also add it to the strings you’re about to play), which is when you use MIDI filters. Other than very specific controls, though, you want to filter things out so you don’t have weird effects of adjusting something you didn’t mean to.

Pedals

The two most common uses for pedals are sustain and expression. They need special consideration in a multi-keyboard setup. As an example, you want to layer your Roland’s piano with a Korg synth pad, and with your other hand, on the Yamaha, you want to play some strings. You want to use the expression pedal to control the volume of the Korg pad. What happens with pedals? Well, you’ve obviously got one foot on the expression pedal, but you now have three different parts that need to sustain. What if the sustain pedal is plugged into the Yamaha? Do you need three sustain pedals? The same goes for expression, if you’re using it on more than one keyboard at a time.

You don’t need multiple pedals; you just need to make sure that MIDI data from your pedals is sent to all keyboards on all channels that need it. With a two-keyboard setup, this isn’t that hard: one keyboard will always have easy access to the pedals, since they’ll be plugged into that keyboard, and it needs to send the pedals to the other keyboard on both channels. Here are the channels from before:

Roland (playing itself): Channel 1
Roland (playing the Korg): Channel 2
Korg (playing itself): Channel 3
Korg (playing the Roland): Channel 4

Let’s say the pedals are plugged into the Roland. When the Roland plays itself, the pedals are recognized. When the Korg plays the Roland, the Roland will probably use the pedals for those sounds as well, even though the MIDI channels don’t match (pedal on channel 1, MIDI on channel 4). If the Roland doesn’t do this, you’d try connecting the pedals to the Korg instead. If neither does this, it gets a bit trickier, but for now, assume the Roland behaves this way. When the Roland plays the Korg on channel 2, the pedals should also be sent to the Korg from the Roland, so the only case is when the Korg plays its own notes. To get around this, the Roland should always send pedal data on channel 3 (or, at very least, send pedal data whenever the Korg will be playing its own sounds; I find it easiest to just always send it, since you can make a preset with a voice used to send the pedal on channel 3 and then you never forget. The best choice of where to plug in your pedals is to the master keyboard. You’re already using a voice on that keyboard to send program changes to the other keyboard, right? Why not also use that voice to send MIDI controller data!

If neither keyboard applies pedals to MIDI data that it sounds on another channel, your choices are limited. You could get some other MIDI device that lets you connect pedals and sends the pedal data out; you’d need a MIDI Hub to make this work easily, since the pedal data would then have to go to all keyboards. You could also arrange the MIDI connections so that the master keyboard was looping back to itself, and use something like this to filter out MIDI data generated by the master keyboard, except for the pedals, and change the MIDI channel on the pedals. The looping of MIDI data is fine here since you’re filtering it out before it actually gets back to the keyboard. Some MIDI hubs are programmable as well, so you can do a bit of filtering or change MIDI channels there.

With more than two keyboards, it gets more complicated again. You now have more MIDI channels that need to receive the pedal data, so you may start running out of external voices on your master keyboard (my Korg only allows 8 in total, whether internal or external; newer Korgs allow 16, but generally you won’t see more than 16). Still, with the right combinations of filters or, in the worst case, using something like this between your master keyboard’s output and your MIDI Hub to duplicate your pedal messages on a number of MIDI channels, it’s not impossible to use one sustain and one expression pedal for all your keyboards.

Volumes, a Mixer, and Standard Presets

With multiple keyboards, you probably always want your own mixer on hand. You’ve created your sounds, and especially when there’s sounds from more than one keyboard at the same time, you want to make sure you’ve got the balance right. If your sound engineer is giving you one (or more!) inputs per keyboard, then you can just send the audio back to them and trust them to mix it the way you intended, but if you have fewer inputs than one per keyboard, you’re going to need a mixer if only as a way of getting several signals to the main mixer on one channel.

My preference is to do my own mixing. I play almost entirely at church, and at my church, the sound engineers are not very musical in general; I trust my ear over theirs. The one exception I make for using only one audio channel is for my piano, since it’s my most common sound and often needs to cut over the rest of the band; depending on which gear I have, I’ll sometimes run my piano sound as a separate signal.

The best way to keep your volumes at the same level is to be consistent. Position all the master volume knobs on your keyboards at some set location (half or three-quarter volume) and prepare all your presets that way. When setting up for a gig, you can make sure the keyboard volumes are all set properly, and adjust the mixer as necessary initially (some keyboards have a lower maximum output than others), and then do any fine-tuning as needed: the room you’re playing in probably won’t sound the same as your rehearsal space, and if you work on your presets with headphones to keep the noise level down, there can still be further differences.

I suggest having at least two standard presets that you always put in the same location on your keyboards (perhaps the very first two slots) so that you can easily get to them. One is a Silent preset, where none of the keyboards are producing any sound; you can use this either in an emergency if something gets stuck on or if you’re sitting on stage for a while and don’t want any way of accidentally making sound (this happens more often at a church service than a performance).

The other is what I call a Level Check preset. Make a set of presets, one per keyboard, that cover the major sounds you’ll use and the keyboards you use them on. For me, this is a piano, organ, pad, strings, and some brass. When I created the presets, I adjusted the volumes to the point where each was playing at a more-or-less standard volume for my presets and all were audible. During a sound check, I can call up this special preset to make sure I can still hear each sound as distinctly as I should, and if not, boost or cut volumes on the mixer.

Advertisement