Operator is Ableton’s FM synth and while it allows the user to design their own waveforms you might think it’s not possible to do PWM in Operator, in fact there appears to be no Pulse wave available in Operator so using Pulse Width Modulation to morph from a square wave to Pulse wave seems doubly impossible.
In fact there is a way to create a pulse wave and also pulse width modulation in Ableton’s Operator synth. This is possible by using two oscillators out of phase with each other and then harnessing the power of FM.
Routing of the Oscillators
Firstly you’ll need to put Operator in any of the the routing modes where oscillators A and B interact through frequency modulation. In simple terms any where the Green boox feeds into the yellow box. I’ve indicated any that you can use in this picture.
The Basic Waveform
Choose the basic waveform of a squarewave, as indicated below.
The square wave I used is the one with the D next to it which indicates “Double” or 128 partials, in simple terms this means it’ s the most detailed square wave that Operator can manage. ensure that the little “R” for re-triggering is activated, this means we’ll have a reliably consistent wave output that starts at 0 every time.
You can see that in my example I’m using a Max for Live oscilloscope by Jonas Obermeuller in order to show you the wave output, I’d suggest you also use an oscilloscope plugin of some kind to help you craft the sound. Any will do. If you have M4L you can grab the same one that I used.
The Second Oscillator makes a Pulse
Now, we add in the modulator waveform in Oscillator slot B. here again I use the square wave with the D next to it and again ensure that the retrigger is active, but here we use the phase offset to set a 25% offset . It’s equally important to set the volume of oscillator B to -22db
If you have an oscilloscope active you should now see a pulse wave, a narrowed square. If this is not the case check the previous steps and images carefully.
Now, try turning the volume of oscillator B down from its -22db position, listen and watch the oscilloscope. You will hear and see the disctinctive pulse width modulation as the pulsewave morphs smoothly to a square wave and back (as you return the Oscillator B volume to -22db)
Create Modulation
So, all we have left to do is to automate this volume control by use of the LFO. Copy the settings in this image carefully
Notice that all of the LFO “Destination A” settings are deactivated and we are instead using the “Destination B” section to route 100% of our modulation signal to the Oscillator B volume.
Now you should hear the sound of LFO modulated pulse width modulation in Ableton’s Operator.
This is just the start of creating bold new unexpected sounds in this powerful little synth. Try dropping the Coarse pitch oscillator B to create a rich complex waveform.
In this Ableton tutorial we’ll take a look at the Ableton Looper and how a hidden feature can enable it to become more of a creative tool for producing unique pads and effects.
This is the second in a series exploring the creative power of feedback in Ableton Live. In the first in this series Feedback in Ableton Live – making a “dub delay”, we learned how to activate the feedback mode of Ableton Live’s effects return channels and to use Ableton devices to shape audio in a feedback loop, it might help you to read that tutorial before this one, but it’s not essential.
In this tutorial we’ll look at carving the frequencies within a long loop to produce unique sounds.
I will show how we can take a simple Sawtooth wave like this
and use loop frequency carving using only an EQ to boost and reduce harmonics, and a saturator to add related harmonics to turn that simple sound into something very different.
About the Ableton Looper effect insert
The Ableton Looper device has an effect insert which enables us to place effects within the Looper feedback path. In the Live manual there is a small and mysterious section regarding the looper entitled “22.21.1 Feedback Routing”. and it is described like this :
“…[it] allows you to, for example, create Looper overdubs that continually feed back
through another track’s devices.
If you have followed the previous tutorial you’ll remember how we used a delay effect and inserted our own effects within the feedback path in order to shape the sound.
In this feedback effecting loop any changes the changes are iterative, this means if we sweep a filter inside the audio feedback loop that filter sweep sound will come back around as feedback and we can layer different filter sweeps onto the looping sound. Iterative changes to the sound in the loop mean we can carve the sound harmonically by listening and changing to our requirements. We can put effects an EQ inside the feedback path, and use them to shape the sound. We can reshape the harmonics which make up waveforms so we can create a noise which moves from a sawtooth to a square wave (for example)
carving the frequencies inside a loop
In the previous tutorial we worked in the Ableton Effect return channels and these require us to send our sound source off to the effects section – but the looper can exist anywhere, and its feedback insert feature relies on the track routing boxes – so lets begin making our turbocharged looper feedback loop.
Step One – create an instrument track with a Looper
Create a MIDI track and place an Operator device on it, followed by a Looper. You should match the Instrument and Looper settings to those pictured here.
If you do not have access to Operator any synthesiser with a basic sawtooth waveform will do. We use a sawtooth waveform because it contains all of the odd and even numbered harmonics and this makes it a perfect starting point to carve out our resultant sound. For example we might choose to slowly remove all of the even numbered harmonics from our looping sawtooth leaving only the odd numbered ones – and that would give us a morphing sawtooth-square wave!
The instrument needs to have a slow attack and release on the amplitude and no filter activated.
In my example the MIDI track is titled “Operator” and that name will help us identify the track in the next step.
Step Two – create the Loop insert channel
Create a new Audio track and title it “Loop insert”, open up the In-Out section so you can see the track routings, if it is not already visible you will find the show/hide In-Out button over on the right side of the interface, it’s a tiny little circle marked “I.O”.
make your Loop Insert track’s IO look like this, notice that the Loop Insert track’s volume is set to -6db, this is important.
Notice that in the the Audio From section of Loop Insert I have chosen “2-Operator”, and in the pulldown immediately beneath I have chosen “Insert-Looper”. In the Audio To section I have chosen “2-Operator”.
Now we ought to check if the routing is working before we proceed so … Flip over to your instrument channel and play a note or two with Operator, if you can hear something from Operator that’s a good start. If you can’t hear anything take a look at the Looper’s Input-> Output pulldown and set it to “Always”
Now press play to start the Ableton transport running and take a look at the Looper, hopefully that has started and appears to be following the main Ableton BPM. If it seems to slow or too fast, ensure the Tempo control is set to “follow song tempo”
Play a note from Operator and if all is going well you should hear the note repeat after a couple of bars, and the note should slowly fade away. If the note does not repeat – check that your Loop Insert channel Monitor section is set to “Monitor:In.”
Once you get the basic looping fade, proceed to the next step.
Step three – Loop Insert effects
On the Loop Insert channel is where we can interfere with our loop. Here we can drop effects in and twiddle with them and whatever we to the sound will come back around the next time the feedback loop returns.
I suggest inserting this basic toolkit. An EQ8 and a Saturator with these settings :
The drive of the saturator is set to +10 db and the output of the Saturator is set to -4 db, the reason for this is that these two will combine to create +6db of signal boost (saturated). If you have correctly set the Loop Insert channel fader to -6db that will add up to 0db of boost. That means the signal will be saturated, but not increase or decrease in volume.
Why choose an EQ and a Saturator?
An EQ8 can remove existing harmonics, or it can boost existing harmonics so the usefulness of this device should be obvious in this context. A saturator will add related harmonics where there are none – so even if we started with a simple sine wave at 100hz – by using a Saturator we could create harmonics for our EQ 8 to work with.
By choosing an EQ8 and a Saturator we have a good simple toolkit to amplify and attenuate existing and to create harmonics where there currently are none.
Step four – Sculpting sound
Now the fun begins! What we are going to do is iteratively carve and sculpt the audio feedback each time it passes around, this means that every filtering or tweaking of the audio will be cumulative. We can add sound in frequency ranges where there is no sound, and take away frequencies too. Carving and sculpting the audio.
For a nice rich source of harmonics to try this out on I recommend setting your synth to output a sawtooth wave, and for the amplitude envelope of the synth to have a slow attack and decay. Around 4 seconds for both will do it.
Perhaps the simplest way to examine how to carve sound with the setup you just created is to watch this video of me “performing” the system.
Here I play a sawtooth wave into the loop and switch to the EQ of the feedback loop channel, where I now have realtime access to the harmonics of the sawtooth. I can increase or decrease their volume, I can make harmonics from (near) silence, and I can wobble the harmonic to create variation.
The outcome shown in this video below is a basic demonstration of controlling the harmonics in a loop, for resynthesis. It should help you understand the potential of this method to create new and unusual sounds.
As soon as you have a loop you like you can drag it out of the looper and into a new Audio track, this way you will collect a bunch of looper takes which are harmonically related and the resultant soundbank can be assembled or selected from to form the basis of anything from ambient masterwork, a movie soundbed, or a rhythmic pumping technological atmosphere.
This tutorial is the first in a series covering feedback in synthesis and production, and how it can be a fun way to create new and interesting sounds with a seeming life of their own. Exploring feedback techniques can bring a complexity and uniqueness of sound to your tone palette, it is a sound with internal coherence so it might appear more “natural” than other types of synthesis or effect. If you are the type who learns by dismantling a working example, then there is a downloadable example file at the end of this tutorial.
What can I make with feedback?
The types of outcomes you can create with feedback range from Dub style delays, to no-mixer synthesis, to looping evolving ambient soundscapes.
here are three audio examples of the sound of feedback
Dub delay
Regenerative looping
A no-input mixer
In this first tutorial I will focus creating the Dub Delay and a little bit of regenerative Looping. The tutorial also shows you the basics which can be used in all of these cases, this involves activating feedback in Ableton Live’s effect return channels and showing you how to progressively reshape the sound produced until you have a unique effect. I’ll explain some pitfalls to avoid which we should bear in mind as our experiments become more complex. In later tutorials I will cover feedback matrices and automation, intermodulation and other more advanced topics.
A “Dub Delay”
A short history and explanation
In the recordings of Lee “scratch” Perry, King Tubby, and others, there is a very recognisable echo sound which has come to be recognised as the “Dub Delay”. Often the effect device used was a very simple type of analog echo device known as a bucket brigade delay, however the notable sonic features which made the distinctive dub sound was NOT inherent in the delay device itself but in the way it was routed. This is what we will recreate.
Traditionally the way to incorporate effects returns on an analogue mixing board would be to have a send dial on each track (such as Drums), the effect send dial would send a percentage of the track signal out to the delay device, the effected signal would return back into the mixing board onto a dedicated return channel. This return channel was a very simple channel with no EQ, or any other features common to the rest of the mixing board tracks. The return channel could not send to other effects.
What Lee Perry (and others) did was to route the effect output of the Delay unit into a normal track channel, thus allowing the effected Delay returns to be EQed and even sent back to themselves using the effect sends, delay-eq-delay-eq-delay-eq-…. . Additionally it meant that the mixing engineers could now send delay into reverb, or reverb into a flanger and then into the delay. This allowed the creation of very flexible feedback loops which could quickly get out of control and start making chaotic noises, but a skilled operator could manage this chaos into art.
In Ableton Live we have a (possibly) unique ability to allow return channels to send back to themselves and each other, but by default you will notice that the return channel send dials are deactivated, this is to prevent users encountering unavoidable issues. We need to activate the return channel send dials for our experiment to work. Create a new return channel and title it “dub delay”, in that new channel right click on the send dial which would send back to itself and choose “activate send” from the context menu.
One thing you should understand is that by allowing Ableton Live to send effects back to themselves we have chosen to make plugin latency calculation for this special track almost impossible, so Live will no longer do PDC on tracks with feedback activated. I will explain further in the section titled “warnings”
The basic effects
Now find the “simple delay” device and drop it into this track, turn the wet/dry on the effect to 100% wet and VERY IMPORTANTLY turn the feedback down to zero. The reason for that is we are going to take care of the feedback signal path, we are taking that away the effect. Set the left and right sides to the same speed.
In a normal delay effect device path the audio comes in and this input gets sent to a “buffer” where it is delayed for a specified amount of time and then this wet signal is mixed into the dry signal and output. The feedback control sends an amount of the delayed signal back into the input so that is delayed once again, and that mix appears at the output and again is sent back to the input.
That closed loop is no use to us if we want to take control of feedback tonality. We need to inject our own toys into the loop so we have much more shaping control over the outcome.
You will need a sound source as we proceed, and I recommend any kind of snare-drum. Load a snare into any channel and use the effect send dial on that channel to send to your “dub delay”
Here is a brief tutorial on making generative music with Live. This tutorial first appeared on the Ableton Live Forum, in 2007 – so the examples were made using Live 6.07 and will work with any version of live later than that.
What is generative music?
Generative music is where you provide logical or programmatic ‘seeds’ and the computer grows you some music based on the parameters you set out, Brian Eno is probably the most famous practitioner.
Feeling lazy?
Here is a LivePack to download with four examples of varying complexity.
Why make generative music, or get Live to make music for you?
Generative music is a different beast from making a track of your own, it is more like planting a garden. In fact a generative piece is like a glorified wind chime, so we could equally ask ourselves “why do people have wind chimes rather than stand in the garden hitting aluminium themselves?” The answer would be the same – the sounds which result may not be “music” but they can be good background noise and in that way quite beautiful and surprisingly interesting as ambience, furthermore the underlying generation can be tinkered with to deliver a wide range of what appears to be expression. A generative piece will sustain a low level of interest for hours!
Live is quite a good environment for creating generative music and I have two methods to do so, an audio based method and a midi method.
I will focus on the more midi-oriented method here.
There are limitations to how far you can go with Live and generative music, but what you can achieve is entertaining.
How it is achieved in Live
To make generative music we need to make Live play or do something whenever a condition is met, we get flexibility by giving the program some freedom. Instead of saying “EVERY time a bar starts play a C minor chord”, we want variation. An example might be “Sometimes play a chord (from a selection of chords) on either the first or third beat, and if you do then perhaps play one of these ralated chords after it, or perhaps think about playing this tune instead”
So now we have a random event which is constrained by a limited set of outcomes, it sounds passably like music.
Tools of the trade
Midi Devices
Random – to generate random pitches
Scale – to make those pitches be less random
Velocity – to make random note start triggers
Pitch – to transpose phrases chords and notes
Note Length – to give variation to the note length
Chord – a variety of uses
Audio Devices
AutoPanner – can be used for ‘cross-fading’ between one component and another
AutoFilter – similar but different
BeatRepeat – repeating random stuff makes it sound like you meant it!
obviously the other Live devices such as Reverb get used, but the above are the ones that are strictly ‘generative’
Getting started – ‘plinky plonk world’
Lets make a bad start! Drop a Midi instrument on a track so you can hear what is happening during the exploration process, use a simple patch like an electric piano (or a Simpler with a triangle wave in) and give it a long release. At the start the output will sound pretty dull, trust me it gets better.
we need a timing pulse to begin with, I use a single midi C3 note in a session clip, I loop this on a quarter bar.
This gives me a nice regular trigger at a 120bpm for the random events to take their cue from. Like this …
That’s all the midi programming we will need done! Next we need to make that ‘trigger’ random, to do this we use the Velocity device. Actually two velocity devices, like this
The first takes the incoming repetitive note and randomizes the velocity.
The second one only lets notes through which fall in the selected velocity range.
This is the core concept of how we will select and filter random values in Live. How we can decide “do I do something, or not?”
In the image that range is from 0 to 59. You will see that I assign the range to a macro called ‘chance’. Try this, it will soon become apparent what ‘chance’ does.
Remember to set the Velocity device to ‘gate’ mode.
You might notice I also set the ‘random’ high on the second Velocity – this is so that further down the chain I have a ready made scattering of random values to make use of for further filtering (you’ll see!)
Now we need to make the pitches change, the most brutal way is to use the Random Device, simply dropping it in the rack after the Velocity section will give you an annoying random ‘tune’.
Give it some pattern with the use of the Scale plugin, choose a nice preset or set your own if you know how … and there you have your first very basic generative thing. A bit shit though isn’t it. It plays random notes, they have random pitches which conform in a brutal way to a scale. It’s a start. It’s hardly music though.
Make it sound more interesting
There are a few things about our example above which make it a bit dull, unpleasant and it just .. well, un-interesting. It sounds very machine like in its random nature rather than musical, there is no tonal variance such as there might be in a human musical peice, there is no sense of drama, no sense of space or *cough* ambience. If we can simulate any of that, it would help.
Conditions of Interest
We used the velocity devices to set a ‘condition’, if the value is within a certain range the note sounds. We can expand this to become what programmers know as an ‘if – else’ conditional.
Here we set two ranges, one produces one outcome, the other will produce another outcome.
This can help us get some drama and tonal variance in there, for example: by allowing any (random) value between 60 and 70 to trigger one chord, while anything between 70 and 80 might trigger a different chord.
Those are ‘conditions’ , when a condition is met an action is taken. To a programmer this might read as if (var>60 && var<70) { do this } else if(var>70 && var<80) {do that}
We can make our piece more interesting by cascading the variations, or to put it in English ” if the value is between 60 and seventy then do either this chord or that chord, if you do that chord then consider these further three variations”
A second choice of conditions follows the first, cunning use of these cascades provides potentially musical output.
Cascading The Variations .
You saw how I set a ‘condition’ with two Velocity plugins, and how we can set two different outcomes using ‘if – else’. Now imagine dividing the random values up into many zones, this way you can create little themic areas.
You can start to go further down the fractal tree, each conditional zone can have a new random value generated to make new notes for itself. Each ‘conditional zone’ can be a different part of your song. The ‘riff’ , the ‘bassline’ , the ‘chords’ . Each of them can watch a zone and do some more complicated ‘riff’ or ‘chord’ related actions anytime that rack is triggered by the main condition.
Here we have the zones of a random seed divided into themes
and in this example the watched zones are further divided to play notes of a riff, each riff note is also conditional. The cascade effect first decides if a note should play on a particular beat and then it decides which note from its small pallet of notes. Out of view here, a further conditional decides whether to transpose the eventual riff.
This is making my brain hurt!
Although this sounds very technical, in fact (if you are anything like me) you will simply set up a few conditions and then add stuff where it seems right, until it makes a nice noise. Once you have an idea of the basic principles you can simply sculpt or garden yourself a tune in an ad-hoc manner. You will notice that my example files aren’t exactly exemplary of order and fore-thought.
But it certainly helps to have some kind of plan!
I think that’s enough to absorb for now.
So, here are some example files
Examples
Zawinul – a complex peice which should sustain a lengthy listen
Here’s something I’ve been working on recently, it’s called “Phases” and it’s centered around a lapsteel and pulsating bass with a laid back beat underneath it. It’s one of my favourite things I’ve done for a while.
I’ve been told that it’s evocative of cruising in slow motion through a hot rocky desert, well see if you agree on that – and hopefully you like it.
Here’s a preview, it’s of course available in the store right now. Or via the ‘add to cart’ button below
This is the home of Steve Angstrom, creator of all sorts of music & artwork. Listen to musical experiments, buy some tunes and read various thoughts.
Social or Instagram
The square wave I used is the one with the D next to it which indicates “Double” or 128 partials, in simple terms this means it’ s the most detailed square wave that Operator can manage. ensure that the little “R” for re-triggering is activated, this means we’ll have a reliably consistent wave output that starts at 0 every time.
Notice that all of the LFO “Destination A” settings are deactivated and we are instead using the “Destination B” section to route 100% of our modulation signal to the Oscillator B volume.
Download the Live Pack
