Performances

Beside the Voice button is Performance. Hit this one. You are now in Performance mode.
Performance mode can be used to make a multi (up to 16 voices, on different MIDI channels) - but it can also be used to make layers of voices, keyboard splits, or master keyboard-type setups that will allow you to control other synthesisers more efficiently. There are 128 Performance locations, all of them user re-writable.

The EX Architecture

Basic Structure
The EX has two basic play modes, Voice and Performance. Voice is used for a single voice, which can contain up to 4 elements. Performance is used for designing a multi setup to use several voices simultaneously, or else to make voice layers, splits, etc. and provides Master Keyboard functions that allow you to control many synthesisers from the EX front panel.

DSP Limitations

Sampling and AWM elements are not limited. Using just this technology you can access the full polyphony of your machine (64 on EX7, 128 on EX5) without any restriction. However, the insert effects, the VL algorithm, AN algorithm, and FDSP algorithms are all sharing the same area of DSP. When an element of this type is selected, the DSP area is sent the formula to perform that specific function. Since these formulae are of various sizes, it is hard to predict how many AN elements, or insert effects you may be able to use in a certain situation. Your EX will tell you when the DSP area is full, though, with a very irritating message that youll see over and over again when trying to make multi setups on your EX:
DSP Resource Full means you cant turn on any more DSP algorithms without first going back and turning another off. It is up to you to manage the DSP and to decide how best to use it within any song setup. As you might imagine, the VL algorithm is very large, but a chorus effect is comparatively small.
Voice Types and Structure
There are several different TYPES of voices. Put your machine in Voice mode, and hit EDIT/COM/Param. (By the way, we will continue to use this method of describing a page, in which the sequence of keys pressed are separated by a forward slash.) This is one of the Voice Common edit pages.

Performance

MIDI Setup, Controller Setup, Effects Setup Part 16
Part 1 Voice Voice MIDI Channel Key/Velocity Range

Effects

There are two kinds of effects in the EX synthesiser, global effects and insert effects. The global effects are called Rev and Cho, and these are the send/return type. Each element or voice can be sent in different amounts to these, but at any moment there are only the same two Rev and Cho in the machine. Rev and Cho are stored with a voice or performance. The insert effects are called INS1 and INS2, and these can be plugged into a voice and will be saved with this voice. There will be a limit to how many of these you can turn on before you fill the DSP Resource (see the section in the Performance chapter on DSP Resource Full). The Insert effects are perhaps more interesting than the global effects - there are far more variations available within them.

AWM Structure in EX

Part Part Part Part Part Part Part Part Part Part Part Part Part Part Part Part

AWM Voice

Drum Voice Other Voice Types VL, AN, FDSP
Contains up to 4 Elements
sample in Zone sample in Zone sample in Zone Samples in Layers defined by Key. No Wave level.

sample in Zone

Voices
For the following section it is recommended that you Load Synth All from the file on the floppy disk.

Voice Type AWM

4 Elements Effects Setup
Element 1 Pointing to a Wave Key/Velocity Range Filters, EGs, LFOs, etc. Voice Element 2 - 4 same as Element 1
AWM is Yamahas word for PCM synthesis. Basically, this is the playback of sampled waveforms, stored inside the synthesiser, and manipulated with envelopes, filters, LFOs and so on. This technique is used in the majority of contemporary synthesisers. If youve loaded the floppy that came with this book, choose Voice mode, and Voice I1:001: Open Saw. This is basically an open sawtooth wave in a single element. If you look at the page EDIT/OSC/WAVE you will see this:
This is the oscillator page. The PRE means that this wave is from the Preset bank (stored in ROM, you cannot erase these). You can choose RAM or FLS here also, which will then look at any samples you have currently in your RAM or Flash memory (if you have some installed). The next number is the wave number, and in the PRE bank you have a choice of 419 different waveforms. Since there are so many, the Cat column (category) can be used to quickly jump to certain types. For example, the waveform, or synthesiser types of waves are categorised by Wv, for Wave. If you change this to Fx, you will jump to the first wave in the Effect category of waves. The MIX page allows you to define the fine tuning, level, and panning of this element. The ZONE page allows you to define the high and low key of the element, or the high and low velocity of the element. This can be useful for making keyboard or velocity splits. EXIT this page, and lets jump to the Filter (FILT) page. You will see that there are five sub menus in the Filter page called : SCF, DCF, SENS, SCALE, and EG. Go to DCF first.

Probably by now, your Open Saw is starting to get a little less than useful. You might want to consider adding some real-time control to the voice, so that you can adjust the filter cutoffs, resonance, LFO speeds, etc. Lets exit the LFO page, and drop back down into CTRL (Control). There are two sections to the Control page, the first is called Pitch. In pitch you can set the pitch bend range, and the upper and lower ranges can be set seperately. The lower range can be set to a maximum of 4 octaves (48 semitones). This page also lets you set the portamento time, and mode. The two modes are fingered and full time. If you set it to fingered portamento only happens when notes overlap. In full time portamento is always active. Now, lets check the second of the CTRL pages, called SET. SET is basically a controller matrix. You have 16 sets, which is a little like having 16 patch cords. You can assign 16 controller+destination combinations. Okay? You can turn each of these sets on or off for each element (to the left of the screen), although at the moment were working with a single element voice. So lets turn set 1 ON for Element 1, and the turn Knob 1 on, and then down at the bottom set the Destination (DST) to 077: AWM DCF Reso. The DEPTH of this control can be set to the right, lets set it to maximum for now.
Of course, when youre in EDIT mode, the knobs are functioning as data entry for the edits you make, so if you want to check the results of this assignment without leaving the editor, hit the hardware switch called KNOB MODE to test the results of Knob 1. Hit it again to turn it off. Now lets say we want to use the Mod Wheel to control the amount LFO1 modulates the Filter Cutoff. Go up to the top of the screen and increment set1 to set2. Turn it ON for the element. Then turn on MW1. Set the Destination to 053: AWM LFO1 FMD (filter modulation depth), and set the depth again to maximum, like this:
Got it? Okay, lets have a look at the effects. As we mentioned earlier, there are two types of effects: insert effects and send effects. The insert effects can be configured in three ways; parallel, or serial (with either INS1 or INS2 first in line). You can set this on the front page of the EFCT section. This page also lets you set up the send and return levels for each of the send effects (REV and CHO) :
Of course, were not hearing this yet because the effects are currently off. Hit the INS 1 key to enter the first insert effect. Set the effect type to 16: Amp Simulator. Again, you choose which element is going into which insert effect on the left, where each element can be OFF, 1, or 2. Like this:

9. RindMod

10. Seismic
So lets check an example. Choose voice 002 in the I1 bank. Its called FDSP Saw. This voice is using the first FDSP Algorithm called EP Pickup, meaning electric piano pickup. Hit EDIT/COM/FDSP. As with other parameters in the EX, each element can be routed via the left hand of the screen. This voice uses two elements, and both of them are switched ON for the FDSP algorithm. If you turn them off, you can hear that the voice is composed of two, detuned, synth waves. Turn them back on again and start looking through the FDSP parameters.
The three types result in progressively more aggressive settings. The middle integrate type, is most useful for actual pickup simulations. Try adjusting the drive parameter to see what happens. This is like moving the pickup closer to the tine of the electric piano. Youll notice that the higher the drive, the greater the dynamic range of the pickup, meaning that the changes in timbre from low velocity to high velocity are exaggerated.
Lets try some other algorithms. Choose number 3, Water. Scroll to the last of the Water parameters and set the Dry Level to zero, so youre hearing only the effect. Now go back to the top. The pitch parameters refer to pitch of the virtual string or resonator. Key follow refers to how this pitch changes across the keyboard. (Note: With all FDSP key follow parameters, a value of 32 represents an even octave scaling. ie. a filter scaled at 32 will scale such that one octave on the keyboard moves the cutoff frequency one octave. This is useful if youre using high resonance filters, and want to keep them in tune.). Scroll down further and experiment with Resonance, Freq Mod, and Mod Speed. Resonance is that of the filter being modulated. If you scroll further, the Feedback parameter is the resonance amount of the virtual string being triggered by this filter modulation.
Now try algorithm 7, Self FM. Again, first scroll to the end and turn off the dry signal. Essentially, what we have here is a single oscillator FM system. The envelope can control the drive of the modulation. This envelope can be either in attack mode or decay mode (its very simple). In decay mode, it starts at maximum, and you have control over the time until zero. In attack mode, it starts at zero and you have control over the time until maximum. Set it to Decay with a time of about 70. Now go up to drive EG and set to maximum (63), and finally go to the top and start lowering the pitch coarse.

Some parameters are controlled upside down There are several parameters that behave upside down when assigning controllers to them. Most parameters are offset by the controller value (filter cutoff, EG times, etc.), and the controller adds or subtracts from the default value. However, some strange exceptions to this rule exist, and Feedback is one of them. The Feedback default is the MAXIMUM that the knob can take it at full depth. FM Depth also works this way around. Its weird, but true. And we dont know why.

VL Voice

( EX5 and EX5R Only )

Voice Type VL

Element 1 VL Element Pointing at VL model Key/Velocity Range Voice Element 2 - 4 AWM type
Voice I1:004: Moby is a demonstration of the VL algorithm used in the EX5 and EX5R. You dont have access to the actual editing of the physical model within the EX, so instead there is a supply of basic models (256 of them, found in the OSC page). You can then offset parameters using the EX editing system. This isnt necessarily a bad thing. Firstly, the VL model is extraordinarily complex, and editing it through a little screen would be a nightmare. Secondly, the interesting side of VL is not the way it sounds, but the way it moves. You have full access to the controller map, and this is what makes VL interesting. Lets check out Moby. The concept behind Moby is that its a mutated brass type of instrument, much bigger than anything you could blow with your skinny little lips. Thus, when you blow into it, you can barely control the embouchure enough to keep it in one mode at a time. Try this, move Mod Wheel 2 all the way back to zero, and play very lightly in the middle of the keyboard. Just hold the note down, and try not to trigger the aftertouch. The sounding pitch should be quite high. Its a ridiculously high harmonic or overtone. Now slowly bring Mod Wheel 2 up, and youll hear the mode shift down through the overtone series until it finally reaches the fundamental. If you open the controller page, youll see a similar layout to that in the other voice types, but again, notice that the destinations are VL specific.
VL Embouchure is probably the coolest of all the VL destinations. This approximates lip tension in a brass or reed instrument. If you dont care about that, then just remember that its the parameter that can force the VL model into different harmonic modes. In Moby, its assigned to both the Ribbon, Aftertouch, AND Mod Wheel 2. Aftertouch is also adding to the VL Pressure destination, which is the amount of energy being forced into the VL model. If you want to get realistic, this is best done with a breath controller, but Moby is pretty far from realistic. The VL Pressure and VL Embouchure are not fixed discrete paths, as you find in most machines, but influence each other in a non-linear, organic kind of way. For example, if the pressure is higher, the embouchure will tend towards more stable modes. VL is a mysterious creature. Other interesting VL destinations to explore are Tonguing, Scream, Growl, Throat, Damping, and Absorption.

Lets have a quick look at these various sections. The above page (MIX) is basic voice selection, level, pan, output, etc. as well as the effects sends. There are only two global effects for the entire multi, so you have to choose these carefully if youre trying to reproduce a Voice as it sounds in Voice Mode.
Copy Effect When you use a Voice in a Performance, it often doesnt sound quite the same as it did in Voice Mode. This is because the effects may be different. If you want to make sure that the Voice sounds the same as it did in Voice Mode, you must turn on the insert effects and duplicate the send effects (Rev and Cho). The easiest way to copy the send effects is to enter the Effect Editor (EDIT/COM/EFCT/). Choose Rev or Cho and hit COPY (on f-key 5). This will give you the option to copy the effect of the currently selected part into the Performance Effect.
The Insert Effect can be turned on or off for each part, but of course, due to the limited DSP resource, you cant turn on very many at once. The Insert Effect will be that which was stored in the Voice itself.
The next page, LYR (Layer) include the Layer Switch (more on that in moment), and other Layer type parameters, key range, velocity range, note shift, and several switches for the MIDI transmit and Arpeggio settings.
The SOUND page is a collection of offsets to the Voice synthesis parameters themselves, such as filter and envelope settings.
The CTRL (Control) page lets you adjust the pitch bend range, and turn the controllers (knobs, etc.) on or off for each Voice. You cannot re-assign the controllers from within the multi. The PRE (Preset) page lets you setup whether or not the controllers and knobs are transmitted over MIDI. And also their default value which is sent when the Performance is selected. This can be very useful for Master Keyboard or Live Applications in which the EX is used to control other synthesisers. The COM (Common) page lets you set up parameters for the entire performance, such as the arpeggiator and controller numbers and effects settings.
DSP Resource Full ! - DAMMIT!
Yes, we know this message will drive you crazy. But try to be cool and lets sort out whats going on and how to cope with this. First of all, remember all the things that use up our DSP resource: VL Algorithms (EX5 and EX5R only) AN Algorithms FDSP Algorithms Insert Effects Any time any of these are in use, some amount of DSP is being eaten up. In the EX5 you may have 4 insert effects, or else 1 insert effect plus 1 synthesis algorithm. If you decide to turn on the insert effect on the Tube Crunch organ (channel 2), this will be fine. But if you now want to turn on the insert effect on anything else, youll get the dreaded message. However, if you decide you really dont need that AN voice after all, change that voice to a non-DSP voice, and then you can turn on 4 different insert effects. In the EX7 you may have 1 insert effect OR 1 synthesis algorithm. On the Tube Crunch organ (channel 2 of our multi) you cannot turn on the insert effect at all until you select a nonDSP voice on channel 1. In all the EX synths, if you try to turn ON an insert effect and the DSP resource is full, youll get the message and the insert effect will stay OFF. If you try to select a voice that requires DSP and the DSP resource is full, the following non-DSP voice will be selected instead. Theres only so much power, and you must decide how to best distribute it.

Play the keyboard, and you should hear the sample you just made. Will get to editing it in a moment, lets first just check the other methods of getting sample data inside your EX.
Loading a Foreign Sample from Floppy
This one is easy enough. Make sure your floppy disk is DOS format (format it inside the EX if you want to be sure), copy a WAV file onto it. (Theres one on our floppy if you want to try it.) If youre a Mac user you must remember to name the file in an DOS manner. That is, 8 (or less) characters, a dot, and WAV. For example: SAMPLE.WAV Or in the case of an AIFF file, it must be named SAMPLE.AIF On the EX, go to Disk mode, File Load, and choose WAVE. The WAV file on your disk will appear as a WAVE to the EX. Load it. Your done. With an Akai floppy disk, you will see the Akai Program. So make sure that any Akai samples you want are included in the program.
Loading a sample via SMDI
You must have a SCSI board in your EX for this transfer. If you have sample data inside your PC or Mac, you can send this data over SCSI directly to the EX RAM area. SMDI is a standard protocol for sending samples via SCSI like this. Check the SCSI ID of the EX (in Utility/Other), since your computers sample editor will want to know this. Open your sample in the editor, and set it up to send to an SMDI device on the correct SCSI ID. Be careful that you dont duplicate any SCSI IDs with other devices on your SCSI bus. Basically, you just SEND from the computer. If all is correct the EX will show an SMDI Data message while its receiving the data. Be careful not to overwrite existing sample data. SMDI requires a sample position to be designated. If you tell it to send to sample position #1, any existing data in the EX at sample postion #1 will be overwritten.

Sample Editing

So now your sample is inside the EX. (If not, load the LARGE.WAV) from our floppy disk. Enter the SAMPLE Mode, and choose the sample number with our sample in it (very likely, its number 1). So hit EDIT and well enter the Sample Editor. There are three main pages in this section, PARAM, LOOP, and NAME. NAME is quite obvious. Choose PARAM first.
Here you can set some defaults for the playback of this sample. Many of these things can be changed from the Voice using this sample, so its not necessary to make any permanent decisions here. You can choose a default original key (Center), and a default playback method. Lets turn the Loop ON for a moment. (FwdLp = Forward Loop) The bottom three numbers are the sample start, loop start, loop length, and sample end points. But lets hit the LOOP page while we look at these.

You should hear the first couple beats of the drum loop. (This is an ideal sound to hear the difference between Assign single and Assign Multi in the ZONE page, by the way. Hit the key twice quickly, and you can hear whats happening in the multi setting.) Now, this loop is currently incomplete. You probably want to be able to control the length of the loop from the keyboard or MIDI, so we need to change the Recv Note Off parameter to Receive. Do this. Now you can hold the entire loop if you want. You might notice that the release time is rather long, and you cant stop the loop very quickly. This release time remains from the trigger mode that the drum voice defaulted to. We can fix it easily by jumping to the AMP/EG page and reducing the release time. Try this.
Lets make another Key, D1, and ADD the same sample to it. RAM:001:LARGE. Lets set the Amp release to 0, and the Recv Note Off to receive. Come back to the WAVE page:
C#1 and D1 should now sound the same. But lets change the Start Offset of D1 to somewhere around 22000 (use the ten-key pad), so that were triggering the loop at the point where the first snare drum starts. With several copies of a loop like this, you can transform breakbeats into your own new patterns very easily.
Now lets mess it up further. Go out of OSC and into FILT, enter the EG section of the filter page and set up a downward sweep, like this:
Go to the DCF section and choose the BPF (band pass filter) on both filters. Set the resonance quite high. And put the filters in parallel mode. Like this:
Now set the envelope sensitivities so that theyre opposite. This way the sweep will move the two band pass filters in opposite directions.
Its starting to get messed up pretty good now. If you want, go back to the OSC/MIX page, and turn on the insert effect for this layer. Remember that the samples are lost when you power down (unless youve used Flash RAM), but your Drum Voice can be stored in the EX, and will use the same samples if theyre loaded again. The only requirement is that the samples be in the same ORDER next time, since any voice using the RAM samples will do so by its sample position.

This is easy enough to do. If you save WAVE, it will save all the WAVES in memory. And when you Load WAVE you can load ALL of them at once, and they will be in the same positions by default. You can also load a single WAVE from a file of many WAVES if necessary, more on this in the chapter called Data Management.
Floppy Disk Format If you find that the floppy disk formatting is rather slow, use your computer. Its just a standard DOS disk.

The Arpeggiator

One of the great features of the EX synthesisers is the fully programmable, 4 track, arpeggiator. For an extensive explanation of everything the arpeggiator can do, check your manual at the section beginning on page 238. For the purposes of this guide, well show you the basics of making your own arpeggiator pattern, using an imported MIDI file as a template. Lets assume youve got a pattern in your sequencer that you think would make an excellent arpeggio pattern. Save it out as a MIDI file, copy it to a DOS formatted floppy, and import it into the EX. (The example were using is on your floppy disk, and is called Arp.MID). By default this MIDI file will be copied to the EX SONG file. If you then hit Play the onboard sequencer will play this file. Check this to make sure youve got the data. Now choose a voice to test the arpeggiator with. Turn the Arpeggio switch ON, and hit EDIT / COM / ARP. Scroll through the parameter called Type.
The first 50 are preset, you cant change them. At 51 they become User patterns, and you can choose any of these to make your own arpeggio pattern. Choose one of the User patterns, and hit F-3, Arp-Edit. Now you have to copy your MIDI file into this location. Hit JOB (beside the main Edit switch), and then section Job 2 / 3. Get Phrase. Youll now have a choice of what measure and track in the Song to copy into the Arpeggiator. Our example is in Track 1, Measure 001 002. Hit Enter and execute the Copy. Then hit exit a couple of times to get out of the Job mode and back into the Arpeggio editor. If you hold down some notes, it should make some attempt to play an arpeggio, but will probably sound completely ridiculous. Weve still got to set it up. Now it gets interesting. First, change the Arpeggio Length to 1 (since our MIDI file is only 1 measure), and set the tempo to whatever you want.
Now, go to the MODE page, and change track 1 to Search Low. Basically this is one mode that causes the arpeggiator to always play a note for each note of your pattern. There are many different modes here, which are explained in the manual, but for now lets just use Search Low. In our example, there are 7 different notes, even though some are repeated. If I play less than seven keys on the keyboard, Search Low will repeat a note even if it doesnt have seven different notes to work with. If you set the MODE to Non-Search, it will leave gaps when unable to find seven different notes. Now hit EDIT again. You should see something similar to this:

The first column here represents the timing of the pattern. The second column represents the KEY CODE. Now the key code is not the note itself, but is a representation of the position of the notes. For example, the key codes were assigned by the arpeggiator to our original sequence like this:
The lowest note is 1, the next lowest is 2, and so on.
When you hold down keys on the keyboard, it replaces these CODES with the lowest note you play, the second lowest you play, etc. Got it? Now, you could theoretically leave the sequence as is, and use it. But, if we tweak it out further, it can be much more useful, or more musical. The main problem with it at the moment, is that if I put my fist down on the keyboard, the pattern that results doesnt have any of the internal relationships or vibe of the original. Why? Because the arpeggiator called the third note CODE 6, whereas my interpretation is that its actually the first note up an octave. So lets fix it. In the ARP EDIT window, lets go in and change the third event to CODE 2, but in the O column (octave column) lets put +1. Lets also do this for the tenth (second last) event.
Lets do the same for the 8th and 11th events. Lets call these CODE 5 (second note) but with an octave setting of -1. And while were at it, well call the fourth event CODE 4, +1. Basically, Im just removing all excessive key codes, and replacing them with octaves. Now my entire pattern looks like this:
There are only four key codes remaining, 2,3,4, and 5. So now, I only have to hold down four notes on the keyboard to supply enough information to get the arpeggiator to crank out something resembling my original idea. And to tidy it up further, Im going to rename all CODE 2 to CODE 1, rename the CODE 3 to CODE 2, rename CODE 4 to CODE 3, and rename CODE 5 to CODE 4. This just makes it easier to read, and understand what is going on when you hold down 4 notes on the keyboard. So finally, my pattern looks like this:
Of course, you dont have to begin with a MIDI file. If you know what you want you could enter it directly in this editor (using the INS function). You could import any data from a track or pattern on the internal sequencer. Or you could record the basic pattern in real time in the Arp Editor itself. Exit this page to go back to the top of the Arpeggiator section. You can name your arpeggio pattern in the NAME page. And before we leave this, lets have a look at the Play Effects (PFX) page. First of all, turn OFF the FxThru switch on the main ARP page:
These default to FxThru ON (which means bypassing the play effects). Now enter the PFX page itself. Play Effects are offsets to the MIDI data, much like the inspector parameters in Cubase. You can offset velocity, note length, etc. and you can impose quantize grids (called Groove) on the data, without actually changing the data itself. Hold down some keys, and start scrolling through the Groove Type list.

So now, I go back to the PTN page, and I see that the MaxLng display says 1. This displays the length of the longest track in the pattern. Using the Sequencer transport controls, hit Record then Play. The metronome should be ticking out its beat, and you can play something into the pattern. Its just looping the one bar around. Play a simple Kick drum or something into the pattern. Hit Stop, and then hit Play to make sure youve got it. Notice that when you hit the Record button, there are a choice of Record modes, which are equivalent to similar choices you have in your software sequencer. Here are your options.
This can be on or off. If Multi is on, then all tracks are recording at once. This can be useful if you have several MIDI channels coming into the EX from another sequencer, or from live MIDI instruments. When its OFF, you can choose which track to record on the bottom left of the screen. Replace mode. Whatever you record will replace the entire contents of that track. Overdub mode. Whatever you record will be added to the existing data within that track. Step entry. Sometimes this is the quickest way to get what you want, just choose the value of each note and hit it. In it goes. Notice two things in the step entry: 1- the note values can be quickly chosen with the ten-key pad 2 - if you set the velocity to EXT it uses the velocity at which you enter it

RPLC OVER STEP

So well record a kick drum, or something simple into Track 1. Make sure MULTI is OFF, and put yourself in RPLC mode. Hit Record, Play. Play a couple of beats. Stop, play it back, make sure youve got it. Second track. Choose Track 2 with the Program buttons on the right end of the front panel. Its a Rhodes piano sound. Hang on, we want another drum track, but this Performance has a Rhodes on channel 2. No problem, we change the channel of track 2 to 1. Hit TCH page:
Change the Transmit Channel of track 2 to 1 (as Ive done in the screen above). Notice you can also choose which MIDI output to transmit on, and you can disable the internal tone generator (TG) if you want to control only other instruments with the pattern. Now go back to the main PTN page, and Record another drum part into track 2. Some kind of rhythm happening? Alright, lets add a bassline on track 3. I want to make my bassline two bars long, so now I choose Track 3, and hit the PFX page. I choose TR, so that what I do only happens to the selected track, and change the Length parameter to 3.1 (2 bars). Like this:

Again, like Pattern, you have a choice of recording MULTI or not, as well as a choice of recording modes. There is one new mode here, not found in Pattern, called PNCH. Punch recording lets you define an in and an out time for recording. This is useful if you just want to record for a short passage within an otherwise perfect take. The Src parameter you see highlighted above can be switched between normal and arpeggio. If its in arpeggio then the output of the arpeggiator is captured as MIDI data, thus freeing the arpeggiator for use elsewhere.
So lets hit record and play any old stuff into a MIDI track. Done? Does it play back? Okay, hit EDIT.
This is the display of note data, hopefully the notes youve just played. There are two modes to this editor, INS (insert) and CHNG (change). In Change mode, as above, you can alter any of the fields with the data wheel or the ten-key pad. The first field is the time of the event, in bars/beats/ticks (at 480 ticks per quarter note). The next field is the pitch of the note. The third field is the length of the note, again in beats/ticks. (The note-like display to the right of this tries to display the closest musical equivalent of the duration, by the way.) Finally, at the last column, we have the velocity. If you hit INS, you can insert a note or any other MIDI event. Like this:
Lets have a look at the other types of tracks. The Tempo track lets you define tempos at any position in the song, so that you can change speed when necessary. Its usually best to do this with INS, since you cant record this real time anyway. If youve imported a MIDI file with tempo changes in it, you can edit them with CHNG. The Play Effect track (Fx) is a strange one. Here you can record real-time. Try it out. You can adjust the Groove template, change velocity offsets, etc. through the Song. Record some of this and then hit EDIT.
Mad, innit? Now, no-ones suggesting that this is the way you would like to write music, and its probably not your preferred way to edit your MIDI data. However, this sequencer can be useful for several things. For example: Import a MIDI file to see what the Play Effects can do to your grooves There are many Groove templates in here, and you can apply them in varying amounts to the timing, the velocity, etc. Use the Sequencer to play sequences live. Okay, theres only one Song, but it can hold up to 30,000 MIDI events. So you could have an entire set in there, just chained together as one song. Use the Sequencer to capture improvised ideas that occur to you when the computer is off. Use the Sequencer to capture the Arpeggiator output, copy it to a Pattern, and free up the arpeggiator for something else The only way to get your MIDI files into Patterns or Arpeggios is VIA the Sequencer Impress your Mum with a MIDI file of Griegs Piano Concerto while you pretend to play
So you see, its not a complete waste of space.
Sequences disappear when you power down. If you make a good one, save it to disk.

Song PTN

This saves the current Song in Yamahas own Song format. All aspects of the Song are retained (play effects, etc.). Pattern save. This saves all the patterns in memory to a single file. Loading of patterns is heirarchical, you can load just one.
Arpeggio save. This saves all the current arpeggio User patterns to disk. Loading of arpeggios is heirachical, you can load just one.
How do I change the MIDI controllers on the Knobs?
The MIDI controllers are assigned to the Knobs at the factory as CC 16, 17, 18, 19, 20, 21. If you want to change this, you can do it for each Performance. EDIT/COM/CTRL/KNOB
Why is it in Performance? Well, this could be really useful for setting up controls for an entire keyboard rig in a live situation. In one track you may want to control the EX itself, but in another track you might want the knobs to control another synth with a different controller set. Or, in the studio, you could make a Performance in the EX that will control your Bassstation for example, so you dont have to get out of your chair so much.

Effect Bypass

You can choose the effect of Effect Bypass in Utility / Other. You can choose which effects are affected by the Effect Bypass, effectively.

Working with SCSI Drives

To format a SCSI drive, whether it be removable type or a fixed drive, youll probably have to boot the EX while the SCSI drive is ON and connected. If you turn on the drive after the EX has booted, theres a good chance it wont be able to find the drive. The next thing youll need to do is format the drive. This is done in the disk page. Select the correct device first of all. Then you can format. The thing to remember about larger drives is to use directories. Otherwise you may end up with so many files on the root directory that you cant find anything. Theres a function for this called Make Directory.
Try to give your directories names that mean something to you.
Once youve made a couple of these you can navigate during the saving and loading process.
SCSI drives tend to be much faster than floppies also. And if youre working with samples, youll need some space. Id recommend using a SCSI drive of some sort.

Epilogue

Where do we go from here?
If youve followed through most of this book, you should have an idea of the huge potential of this instrument. Not only is the EX providing the raw elements of many synthesis techniques, its also giving you some compositional tools to help generate ideas. Once you understand the basic architecture of the machine, its often useful to forget about the designed purpose of the various sections, and to strike out on your own. For example, the Electric Piano Pickup FDSP algorithm might sound great with a vocal sample going through it. Using the resampling feature, you can sample the results of one DSP, and then apply another to the result. You could import your drum grooves, turn them into arpeggio patterns and manipulate them with the Play Effects, transpose it realtime, etc. Put your breakbeats through the Seismic FDSP and turn them to mush. You could import each section of your track as a seperate pattern, and jam the tune live from a KeyMap. Or make a sixteen part keyboard split as a Performance, set them all to the same MIDI channel, turn the arpeggiator on and stand back. A lot of interesting sounds are creating by merely trying stuff out, experimenting without a fixed purpose. Mistakes can be very cool. Miles Davis said There are no wrong notes. Brian Eno said Make the machinery fail. In other words, go for it, break the rules, and make it squeal like a pig.

1.1 1.2

Introduction
This document summarises a workaround to counter the keyboard layer limitations of the Yamaha EX5. The Yamaha EX5 only allows two voices to be split or layered from the keyboard within a performance. If you attempt to set up more than two voices, you are presented with an error message stating Error Too Many Layers. Refer to Page 165 of the EX5 manual for details of performance layers. This is a rather surprising limitation considering that the EX5 is capable of sixteen part multi-timbral operation over MIDI. I find this a limitation because, although I only have two hands, I would sometimes like to have more than two sounds split on a keyboard for live work to avoid having to select voices during a song. This is important to me given that there is a noticeable delay on the EX5 between selecting a new sound and it being available for playing, which makes quick, on the fly voice selection impractical during a song. In an addendum to the main EX5 manual, Yamaha describes a workaround which involves combining voices together, but this assumes that the total number of elements within the voices being combined is no more than four elements. For example: Four single element voices can be combined into a new split/layered voice. Two dual element voices can be combined into a new split/layered voice. A single element and a triple element voice can be combined into a new split/layered voice.

1.3 1.4 1.5

If you do not already have this addendum, it can be obtained from the Yamaha Manual library at www2.yamaha.co.jp/manual/emi/index_e.html (you will need to register to gain a user ID and password but its well worth it). Do a search for EX5 and download the Additional Performance Setup Tips PDF file (EX7E3.PDF). This technique works, providing you are willing to combine voices in this manner, and providing that they are simple enough with respect to the total number of elements. The rest of this document describes an alternative technique for countering the layer limitation. The technique in outline is to treat the EX5 as two separate units a master keyboard and a tone generator. Before describing this technique, its worth noting that you cannot use it counter the inherent limitations that exist in performance mode, such as the number of effects or that certain voice combinations are not allowed (e.g. AN and VL together); these limitations are due to the available DSP power of the EX5.

1.9 1.10 1.11

MIDI Connection
First, you need to disable local keyboard control of the EX5 tone generator. You do this by selecting the UTILITY button, and selecting the MIDI Setup menu. Within this menu find the Local setting and set it to OFF. The relevant screen is shown on Page 274 of the EX5 manual. With no external MIDI connection youll now find that the EX5 will not be making a sound when you play the keyboard. To get the sound back you need a MIDI connection and a properly configured performance. The MIDI connection can be made in one of two ways. You can either take a single MIDI cable and use it to loop the EX5s MIDI OUT A port back into the MIDI IN A port, or you can make the same connection via a computer or hardware sequencer, providing that your system can automatically echo received MIDI data. Youll have to check your sequencer documents to see if it can do this. I have used the former option of just using a single MIDI lead to prove the concept, because I predominantly use the EX5 for live work when I dont have it connected to a computer, and I needed to ensure that this at least worked. For testing this was also the best option, as I didnt have to worry about where latency was being introduced; if a PC was used to echo data, I wouldnt have known how much latency was induced within the PC. With the single lead, I know that any latency was attributable to the EX5 and the external MIDI flow.

3.1 3.2 3.3 3.4

Performance Setup
Select Performance Mode, and the performance you wish to edit. Pages 165-167 of the EX5 manual describe the editing selections that youll be using to set the parts up. Now consider your performance setup and select the voices for the parts that you wish to split/layer. Normally in a performance, the parts you play from the keyboard must have the keyboard layer switch selected. With this technique all parts are played via MIDI, so ensure that for all parts the Layer SW switch is disabled. Of relevance to this technique, all parts can be independently assigned: A note range the default range is the whole keyboard A velocity range the default range is the whole velocity range A MIDI channel number the default channel number is the same number as the part number Whether or not MIDI data is transmitted for that part on either MIDI Channel A or B.

Thus to spread your selected parts over the keyboard, adjust the note range for each part to restrict it to the area you wish to play it from. You are free to create separate zones, layered zones or a combination of both; this includes layering parts either partially or fully, depending upon your requirements. You may also assign velocity limits if you wish to make things a little more complex (e.g. two layered piano sounds but only one is played if within the selected range). For each part you are using, youll also need to select the Trns MIDI A switch within the LYR page and set it to on (this is described on Page 165 of the manual). You should now hopefully be able to hear your parts in response to playing the keyboard. However, if you have layered parts, you may be using more MIDI channels than is necessary and consequently transmitting more data than is necessary, which may lead to latency problems. To reduce delays induced by the amount of MIDI traffic you can minimise the number of channels used. By default each part within a performance gets its own MIDI channel (e.g. part two gets assigned MIDI channel 2); this channel number is used for both MIDI transmission and reception. You can adjust the channel number for all parts, and several parts can be assigned the same channel number. If you consider four parts split across the keyboard with no layering of parts. All parts can have different MIDI channel numbers, but they can actually share the same MIDI channel with no adverse effects. As the parts are split, no savings on MIDI bandwidth are made, but the unused MIDI channels are made available for other instruments. If you now consider those four parts all layered over the same keyboard range, they are all transmitting and receiving the same note data over four MIDI channels. In other words more data than is necessary is being sent over MIDI. If you set all layered parts to the same MIDI channel number, select one part and enable the Trns MIDI A switch, and disable the Trns MIDI A switch for the remaining parts, you can reduce the number of channels needed and also reduce the amount of MIDI data being sent.

3.8 3.9

3.10 3.11

MIDI Latency

I found in my tests that, dependent upon the playing style and tempo, between four to eight voices could be stacked before latency was noticeable. Fast glissandos on the keyboard really expose the technique when stacking more than four voices. Having said that a four voice stack becomes pretty mushy anyway, and my interest is more in splitting the keyboard, for which the technique works pretty well with no noticeable latency. With sixteen voices stacked together it is quite easy to overflow the MIDI buffer on fast glissandos. As described above, latency can be minimised when stacking by ensuring that the stacked parts are all using the same MIDI channel and that only one of the stacked parts is transmitting the data.

4.2 4.3

5.1 5.2

Conclusion

This document shows how to counter a limitation of the EX5 that I still find hard to believe is there. This technique can be mixed for any combination of split and/layered voices over the keyboard range, and, within the limitations of the tone generator, can be very effective (I just wish Yamaha had allowed me to do this within the machine in the first place!). The workaround is not a cure-all; certain issues such as DSP power and MIDI latency need to be considered when using this technique, but this is true of the EX5 or any workstation in general. The workaround certainly works for me, and I hope that this description will help other EX5 users out there.
Derek Cook 15th June 2000
A Note from Derek and Ski
This mini-tutorial is brought to you in the hope that it will benefit your EX experience. We hope that it will turn into a series of useful documents. If you have a short topic of your own that you think could help others, we strongly encourage you to write it up, even if it's only a page or so. It doesn't have to be a huge tutorial to be useful. We can help a bit with the writing, if necessary, and we'll format it, publish it in.PDF format, and post it for you. Visit us at the EX5Tech discussion boards at http://www.wbmedia.com/ex5, or send email to: Ski ztja05@zdnetmail.com Derek dacook@dera.gov.uk