How to get to particular places, what leads where and so on. The easier you can find your way around, the quicker and more rewarding your programming will be.
Exercise #1:Do you know how to get to the following screens?
You can find these out by reading any synthesis books, music magazines, and experimenting on 'easier' synths such as analogue mono synths.
Note: this text is not a primer on synthesis fundamentals - search elsewhere for that!
Exercise #2:Do you understand each of these terms?
Exercise #3:Do you understand each of these terms, including how they relate to each other?
Nothing kills the creative flow faster than an hour-long data-entry spree around modern synths. I prefer to just sit down with the WS and a pair of headphones, with no other distractions, and perhaps get one or two new sounds done each session. That way, the next music session you do, you will have a few new, original sounds to work with, which can often be a good source of inspiration too.
"You gotta keep 'um separated".
It's very tempting to layer patches after patches of sounds together, and the Wavestation makes this fairly easy. However, these sounds a difficult to fit into a mix without swamping it, and without programming in 'space' into the sounds, are likely to sound too cluttered (especially with wave sequences). It has has a detrimental effect on the available polyphony - for example, two four-osc wave cross-fading wavesequence patches use 16 oscillators, giving you 2-voice polyphony!!
This maximises the gain stage of the WS and ensures the highest signal to noise ratio for the best quality output.
This is because 2-osc patches are louder than 4-osc ones. Think about it: a 4-osc patch will output each oscillator at 25 per cent of its full output, in order that the D/A converters can output all four oscillators mixed together.
The really nice sounds tend to be the ones where there is some level of real-time control, and the WS is pretty good at this. It has extensive modulation routings, and controller 'inputs'.
You have at your disposal:
I don't find editing the WS from it's front panel too difficult - I do most of my editing there rather than with a computer editor, which avoids sysex transmissions etc.
However, some form of librarian system to store and organise my sounds I find ESSENTIAL. Especially when dealing with the dependancy management. These programs allow you to move a patch around while automatically updating and performances that use that patch to reference the new location. It makes organising sounds a breeze.
On screen editors also provide nice functions, like being able to see whole lists of the PCM Waves etc, which can be cool.
Exercise #4:Load a bank of performances, then a completely separate bank of patches (ie patches that are used for a different performance bank than the one you just loaded), then a completely seperate bank of wavesequences. In fact two banks of each. Or three.
Now you will have banks of performances that use the wrong patches, and banks of patches that use the wrong wave sequences.
Go through the performances and give them a try. Some will be useless, some probably won't make a noise at all, and most will probably require some editing to get the most from them. But you can stumble across some really cool sounds this way - without buying any new ones, or programming your own!
If you're lacking in inspiration, try this:
Exercise #5:Load a full bank of sounds. Select a pad sound, or one that has a decent sustain, and play and hold a note or two.
Now using the INC/DEC buttons, change performances. The notes you are holding will sustain, but the effects will change to the new sound's settings. This is a quick way of auditioning different effects and can also produce some way cool sounds. When you find something you like, use the FX copy facility to paste the effects settings into the original sound.
If you are layering two or there patches in a performance, use the panning and buss settings of the PART DETAIL page to space the sounds across the stereo spectrum, to widen them out and give them more room to breathe.
For instance, you may have a main pad suond set to 50/50, right in the middle of you head, with two other layers set to 30/70 and 60/40, which is left-ish and right-ish.
It's not usually a good idea to put things right smack left and right ("A" and "B") unless it's for a specific effect - the sounds right at the extremes of the panorama tend to disappear when mono'd down.
It's possible to get a delay effect without using an FX processor up. Simply put two copies of the patch into a performance, and adjust the 'Delay' parameter on the Part Detail screen for the second part to something in time with your music. Try around 1500 or so. This does however reduce the available polyphony.
Keep an initialised bank of WS sounds around, so instead of resetting the WS to clear its contents, just reload the initialised data via sysex. This is a good way to start programming from scratch.
The WS get a bit senile sometimes, especially after much sysex transfers. It's good practice to save the RAM contents every once in a while (say, every month or two) and reset the WS using the special screen functions. This cures it's somewhat addled brain and the WS will behave again.
This is also worth trying if you get lots of sysex checksum errors when sending data to the WS.
Don't forget, while your sounds are in memory THEY ARE NOT SAFE!
They are ONLY safe if you have backups on a computer disk somewhere that can be reloaded.
If your WS crashes, it can sometimes be necessary to reset and re-initialise the machine. If you didn't have backups of your sounds, you've lost 'em for good.
Come on, stop using the presets (which is after all what this article is all about, surely...) to play, and use them to investigate programming.
There is some very good programming in those sounds, and you can learn a lot from how these things are done. If you hear something in a sound and don't know how its done, try to find out by delving in to the editing screens and taking a look.
As mentioned earlier, it's good programming practice to try to make a sound that you know in advance.
Perhaps you heard a great sound on a CD or something. Could you recreate it using your WS? Have a go. You'll need to identify the major parts of the sound, the tonal characteristics, the envelopes, LFO's, additional things such as pitch bends and filter sweeps and so on. Take each section in turn, and take it bit by bit, referencing the original sound from time to time to make sure you're not wandering off track.
In the beginning, even if you don't get quite what you're after, you will still be gaining programming experience and coming up with useable sounds in the meantime.
Use the WS Jump/Mark facility to access pages that you find yourself going to a lot, or ones that you find difficult to find again.
The WS interface is pretty good at letting you know where you most of the time. Get into the habit of looking at the top line of each screen, which gives the page name and sometimes other details, and also other things - for instance, the Part Detail page of the performance edit screens shows the current part you're changing. You need to know you are changing the right one...
Other times, check that you are editing the correct oscillator within a page, or the correct wave sequence. This things usually display the relevant information at or near the top of the screen.
Well okay, an exaggeration, but you are hardly being original... The WS is a great source of original sounds, but these really are demo material only...
(Before I get flamed, this is all IMHO... :)
Take a two or four oscillator patch, with differing waves on each oscillator. Now, when you move the joystick around, the timbre should change as the relative waveform level balance changes.
This movement can be automated by the vector mix envelope, so program in a basic looping envelope that changes the timbre by cycling between the waves.
Now, this is all well and good, but the vector envelope retriggers for every note. How can we get it going without this retriggering, so that when you play staccato notes, the timbre changes for each note. This is one of my fave tricks, and I really like the effect.
The trick is to route an LFO or two to the vector envelope modulation sources. The LFO is set to free-running mode so that it isn't retriggered every note. This will make the vector envelope to be effectively free running, even when notes are not being played (this may also be useful for WS A/D users processing external audio through the analogue inputs).
Okay, let's do this. You need to do the following:
Edit a performance, and select a four oscillator patch.
PATCH --> WAVES --> MIXEV softkeys, which will take you to the Edit Mix Envelope page. Enter a little looping vector envelope:
Level Time Point 1: 25 40 Loop: 0 <--> 3 Point 2: 25 40 Repts: INF Point 3: 25 40 Point 4: 25 40Now press the MIXMOD softkey, and select LFO 2 for X Source 1 and Y Source 1, with amounts 127 and -127 respectively (the exact amounts won't really matter, as long as you can hear the LFO working. Have a play with the amounts until you get something you like).
Now prss the LFO 2 softkey and enter the following parameters:
Rate: 20 Initial Amount: 127 Shape: TRIANGLE Sync: OFF Delay: 0 Fade-in: 0 Depth + Rate mod both 0.You may want to experiment with the exact LFO rates.
Now if you play the keyboard, you will notice the vector changes continously, rather than retiggered every note.
See Tip #23
Yup, wave sequences can be modulated. basically, you can define where you are going in the list of waves.
The parameters you want are in the Wave Sequence Utilities page. Select a patch within a performance that uses a RAM wave sequence, and go to the Edit Wave Sequence page, and press the UTIL softkey.
There are mod source, mod amount and start step parameters. If you for example set the mod wheel as the mod source, and a corresponding positive amount (the exact amount will depend upon the amount of steps in the wave sequence), then moving the mod wheel will sweep the oscillator though the wave list in real time.
Selecting velocity as the mod source will result in higher velocities starting further up the wave table. You can use this to give some very nice velocity dependant sounds.
If you want extra volume for a particular patch, route each oscillator to all four FX busses, and change the patch routing within the performance to PATCH for the settings to take effect.
The first is done from the FX-BUS softkey from the Edit Patch screen. You'll see the (up to) four oscillators and the respective assignment for each. Change all settings to "ON".
Then exit back to the edit performance screen, and go to the Part Detail page, to change the FX-Bus setting for that patch to "PATCH".
Take care of your overall levels - the WS is quite easy to drive into clipping which doesn't sound very nice. Take special care when modifying some of the level/gain stages on some effects. (For more info on gain staging, see Gain Staging (or How to avoid clipping)).
In this first lesson I want to focus on the effect section. I've downloaded and checked out the Techno Volume that was created. There are some really good ideas here that are missing that "that something extra" and usually I found that it is due to not using all the tools in the effect section.
For this lesson I'm going to walk through a Performance from this Techno Volume. This means that you need download this bank of sounds and have them loaded into your Wavestation of choice before proceeding -If you don't have them, the comments might still help you just won't be following along with the same sound I am walking through.
Program #13 "Reflex *"is a good Performance for us to look at.
Hit the Edit Button. Notice that there is a single patch that is using the effects in Serial with a Stereo Chorus (FX1) followed by a Flanger into a mono Delay (FX2). Major Tip: Mastering the effect section in any Korg product is a big part of getting the most out of these instruments. Since we are in the digital domain these effects can have very dramatic results with out increasing the noise floor (very much). if you conceptually see the effects as another element to the SOUND making process and not just a last minuet addition, your sonic output will change dramatically.
Our first lesson will be in swapping FX1 for FX 2 and FX 2 for FX 1 - after doing this we will then add some low and high eq to give the sound more definition.
If you want to swap FX from the same performance in a Wavestation, you first have to copy the performance to another location since there is no "Swap FX" option. Hit the "Write" button and choose #14 to keep things easy. If you want to rename the first letter in #14 to something else it will make the next step much easier.
From with-in the FX section (hit "Effects" or Page up 2 times on a WSSR) there is a Copy but this copies both effects which is not what we want. Inside of each effect (Press FX 1 on a WS or WSAD to see the individual "COPY" option. On a WSSR hit "Edit" when the screen says "FX1:StereoChorus" and page up two or three times).
Do this slowly to not mess up.
From within FX 1:
From within FX 2:
If you now play the sound and are hooked up in Stereo, you will notice a lot more dimension to the sound since the Stereo Chorus is not the last stage before the outputs instead of the Flanger>Delay (this is a mono effect). Hit the compare (WSSR hit the "Write/compare" button 2 times) button to check out the difference a simple effect swap can do.
Next we want to give this sound a little extra definition using the shelving EQ in the stereo chorus - there are also shelving eq in the Flanger>Delay that can be used for more sonic shaping but be careful, the Wavestations have a point where they start distorting quite badly internally if levels are too high so be aware of this fact when working with the EQ.
Edit effect 2:Stereo Chorus - I found that Low EQ +5 High EQ +7 sounded better. I also liked slowing down the chorus rate to .15 so that it evolved at a rate similar to the Flanger.
Edit effect 1:Flanger>Delay - to get more definition out of the bottom half of the keyboard, setting Low EQ to +5 while bringing up the LFO Depth to 65 fit the bill.
Now hit compare to see how we've changed from a nice mono program to a stereo program with added definition by only using the effects.
A few more tips:
The issue is related to something you might be familiar with in mixers: gain structure. Basically, all systems (analog and digital) will have some point of maximum gain, after which they begin to clip (analog systems may do this gradually; in digital systems, there's usually no middle ground). The irony is that, right up until that clip point, the louder the signal is, the greater the S/N ratio and thus the better the sound.
So, the trick is to get it as loud as possible before clipping. I'm sure that most of you have heard that before.
In a digital system, it's the same thing. However, in a digital *synthesizer*, it becomes difficult to realize the optimum automatically. This is because we have more than a single signal (say that ten times fast) to worry about; we have 16 or 32 or 64 separate signals, which then may be summed into a single stream. More complex, those voices each have their own varying levels. So, how do we deal with achieving an optimum gain structure?
The ultimate in conservatism would be to make sure that nothing clips, ever. This is pretty easy to achieve; you just assume that each of your voices is playing a full-scale sine wave, and scale the volumes back so that the summed result is just under clipping.
The result of this technique will be that your voices are pretty low-level, in which case S/N ratio suffers and overall output level is low. Play an original EPS keyboard for an example of this.
Most of the time, the voices will be considerably softer than a full-scale sine wave. For instance, they may have a loud transient, and then a softer sustaining tone. In this case, each individual voice can be considerably louder before clipping than is assumed in the conservative gain-scaling scheme above.
So, we compromise. Scale the level back so that, in the general case, things won't clip. The grey area is handled by voicing staff; when a sound clips, they scale back the Part levels, just as David Holt recommends above.
The result is that some human discretion is necessary, but that the overall output level is hotter and S/N is better.
The LFO's in the Wavestation are quite capable, and they include a 'random' waveform for general warbling. However, you can increase the range of random effects by cross-modulating the LFO's, where LFO is generating it's waveform and also being modulated by LFO2. This is achieved by selecting LFO2 as a mod source in LFO1. (You could also do the reverse as well for some really wacky effects).
The effects section is the key to all the effects/output routing, and there are a number of ways to route different things in different ways.
The FX section has 4 inputs ("busses"), named A, B, C and D. The soundgenerating parts of the WS arrive at one or more of these busses before going through the FX processors and on to the outputs.
The FX section has four outputs, named "1", "2", "3" and "4", and these directly correspond to the four output jacks on the rear of the WS. How sounds are routed from the input busses to the outputs are determined by the effects configurations.
Note: For the time being I will ignore the FX MIX parameters (more on those later).
Now, the two FX processors (FX1 and FX2) can be configured in SERIES or PARALLEL modes of operation:
This mode lets you have for example FX1 providing reverb, with FX2 providing chorus on top of that (in other words, the two FX processors are acting as one big combined effect).
A diagram helps visualise this (better ones are in the manual and displayed graphically on non-SR Waves):
WS Sound Generator | | |----> FX BUS A ---> FX1 (left) ----> FX2 (left) ------> Output 1 (left) |----> FX BUS B ---> FX1 (right) ---> FX2 (right) -----> Output 2 (right) | |----> FX BUS C ---------------------------------------> Output 3 -----> FX BUS D ---------------------------------------> Output 4(Remember that I'm temporarily ignoring the FX mix parameters for now, and thus have removed them from the diagrams).
In series mode, because busses A and B go through the effects, busses C and D go directly to outputs 3 and 4. This is a handy mode to have because you can have one sound going through the FX units and emerging at the main outs, with another seperate sound going to buss C and D and hence emerging at outputs 3 and 4.
In practical terms then, you can have a multi setup that sends one patch through the two effects units (as per POLY mode), and another that goes to the other two outputs to be processed by an external FX unit (I'll describe how to setup all these things later).
WS Sound Generator | | |------> FX BUS A -----> FX1 (left) -----------> Output 1 (left) |------> FX BUS B -----> FX1 (right) ----------> Output 2 (right) | |------> FX BUS C -----> FX2 (left) -----------> Output 3 (left) -------> FX BUS D -----> FX2 (right) ----------> Output 4 (right)This means you have two independent processors. One could be providing reverb for one sound, with the other providing echoes for a different sound.
Using this table in conjunction with the effects diagrams you should be able to decide which settings you need for a particular routing.
Let's have an easy practical example, then. Suppose you want to, within a performance, have one patch (say, ROM-04 WS Strings) going through a reverb and being output on the main stereo outs, and another patch (say, ROM-21 E.Piano) going through a chorus and being output from the sub outs. For this you need to have the FX in parallel mode. Setup FX1 with a reverb, and FX2 with a chorus. Now set part 1 (the string patch) to "50/50" and part 2 (the elec piano patch) to "C + D".
This arrangement of routing each patch to different FX bus combinations is quite flexible, however there is extra flexibility in the "FX Bus: PATCH" setting mentioned a short while ago. From the performance edit page, select the patch you want to edit with the cursor, hit the PATCH and then "FX BUS" softkeys. Now each oscillator within a patch (1, 2 or 4 oscillators per patch) can be independently sent to any or all of the four FX busses.
WS Sound Generator | | |----> FX BUS A ---> FX1 ------/--------> FX2 ----/-----> Output 1 (left) |----> FX BUS B ---> FX1 ------|------/--> FX2 --|---/--> Output 2 (right) | | | | | |----> FX BUS C ------> MIX3 -/------|----------/---|---> Output 3 -----> FX BUS D ------- MIX4 -------/---------------/---> Output 4The MIX3 section mixes some of bus C's signal into the bus A signal, and there are two points of entry. You can set the FX MIX3 parameter to "OFF" (no mixing at all, C does not get mixed into the A bus), "DRY" (bus C is mixed after FX2, and therefore does not go through either FX processors, but is still output from the main outputs), "WET" (the signal is mixed in before FX2, and therefore does go through the second FX processor), and various combinations between dry and wet. You can also modulate the FX mix amount as it has a modulation source, so the mod wheel could control the mix amount.
With the "DRY" setting, it means that patches routed to bus C will not be processed by FX1 (whatever effect type it's set to) but will be processed by FX2. MIX4 does the same thing but mixes bus D's signal into bus B, again before FX2.
This effects routing is useful if you want to have a specific effect on one patch (say a really drastic distortion), and then a blanket reverb processing on everything else. You'd send the patch to be distorted to busses A and B, and everything else to C and D, with FX1 doing the distortion and FX2 the reverb, with the final stereo mix output from outputs 1 and 2 (the main stereo outs).
Let's now have a look at the parallel routings, including the FX mix:
WS Sound Generator | | |------> FX BUS A -----> FX1 ------/--------/-------> Output 1 (left) |------> FX BUS B -----> FX1 -----|---/----|---/----> Output 2 (right) | | | | | | MIX3 MIX4 | | | |------> FX BUS C -----> FX2 ------/---------|------> Output 3 (left) -------> FX BUS D -----> FX2 ---------------/-------> Output 4 (right)Now you can see that the FX mix lets you mix in the output from FX2 into the A and B busses, chiefly to offer independant FX processing from the two FX units and have them all come out of the main stereo outs. In this mode, both MIX3 and MIX4 have settings of "OFF" (as before, no signal from C/D is mixed into A/B), and various pan settings from hard left to hard right (unlike the series mode routings, bus C can be mixed into both A and B, and panned between them (dynamically as well, using the mod source), whereas in series mode C is only mixed into bus A).
So in summary, the FX MIX is simply a little mixing section that mixes busses C and D into A and B (and only that way around - A and B don't get mixed into C and D).
Now the main problem here (apart from not having enough FX processors!) is that some performances use specific effects routings/configurations to achieve their sound, and consequently sometimes it is impossible to get the routings you want whilst still keeping the same processing. This is one of the reasons why routing is not always straightforward. The other is that the FX bus routings are stored within each performance, and to change the routing you have to edit each performance (and if it's one of the ROM ones, you have to then resave the edited version in a RAM slot). And to make matters more complex, if the FX bus setting is "PATCH" then we may have to delve into patch editing to change the bus assignments. Khew, are you still reading this far?
Well anyway, let's start with a practical example. Let's say we want our multi to have four preset performances (on midi channels 1 to 4), with each one coming from a separate output, with no effects processing (useful for processing on the desk).
The performances that we would like to use are (all from the ROM bank for ease of demonstration): 4 Mini Lead, 17 Octave Strings, 26 Tine Piano and 37 Round Wound.
So switch to multi mode and place the performances in the appropriate multi slots, and init the effects.
Now if we have a look at the effects routings of each of these performances - to do this we have to temporarily switch back to poly mode, otherwise we can't inspect each performance's effects settings. Select the desired performance and hit the EDIT and DETAIL softkeys, and step through each of the parts to examine their FX bus parameter:
Now if we go back to multi mode we have each of our four sounds being routed to A, B, C and D respectively, meaning they will emerge from outputs 1, 2, 3 and 4 respectively. Now you can play with the effects to add any processing you need. If you select two dual-mono effects you can process each of the four sounds independantly.
As you can see from all of this, the effects section and routing is quite complex, and I cannot describe every possible way of using it here. I hope I have given some hints and tips on using the effects, and if you understand how it works you should be able to set up the routings you need.