In the previous tutorials we have manipulated Oberon in ways very similar to classic subtractive synthesis. For this third tutorial we will start using some features more specific to additive synthesis: bend, harmonize and partial clipping. All these features modify the spectrum of the sound at a partial level. But then again, under the hood, in additive synthesis, everything comes down to modifying the partials.

First of all, let’s start with a clean sheet by creating an Oberon instance and reseting it. Set the wave start of Oscillator 1 to 50% so as to use the second shape of Wave 1. You should hear a classic unfiltered triangle wave sound.

Partial Clipping

The first partial modifier we will explore is located at the bottom left of the Oscillator 1 section and is labelled “P.Clip”. To better understand what is happening, you can open the spectrum EQ window on the Oberon rack or just rely on your ears. Let’s first play a note while cranking up the P.Clip knob:

As you can hear, the sound gets brighter and harsher with increased partial clipping. If we put back the wave start to 0% to play a saw wave (the default Wave 1 shape 1, remember), the effect is even more noticeable and the sound is almost nasal:

Looking back our spectrum EQ window, here are the two saw spectrums, one without partial clipping and the other with a full effect:

As you can maybe deduce, partial clipping has the effect of soft clipping the partial levels. In other words, it will boost low level sections of the spectrum while slightly lowering or at least maintaining high level sections. But then again, if we don’t want to delve into that much detail, we can also just see this as a “brighter” or “harsher” knob!

Let’s set back the wave start to 50% to have a triangle wave and put the partial clipping to around 75%. This will be our base timbre for the tutorial and it should sound like this on a C2 note:

Bend

For the moment we have just played around with spectrum levels but another nice aspect of additive synthesis is that we can also change the pitch of parts of the spectrum. Remember, additive synthesis lets us control individual portions of the spectrum through partials. This is done with Oberon with the Bend feature.

Again, prior to any explanation, lets just hear what this does:

It kind of sounds as if we were upwards pitch bending our sound but not quite as the timbre also changes to become metallic or bell like. Again, if we take a look at our spectrum window before and after bending, something pops up:

The peaks identifying the harmonics of the sound seem to have been displaced to the right of the spectrum, that is, to the higher frequencies. The effect seems also more pronounced as we go up in frequency. Let’s play a small melody, without and then with 50% bending:

As we can hear, the pitch or notes remain the same. It is just the timbre that changes to become really more metallic or glass like. That is pretty normal as the sound becomes less and less harmonic as the partial pitches are displaced from the harmonic frequencies, which is exactly what happens in bells, metal rods or very rigid strings: the frequencies produced are not harmonic. If you increase your guitar string tensions to an insane amount, apart from pitching the note up, you will also notice this effect.

If you don’t want to go into much detail, you can just stop there and consider the bend modifier to be analog to a “string tension”, “metallize” or “glassify” effect. If you are more curious and power hungry, let’s dive.

Bend curve

You may have noticed in the series of display buttons, next to the envelopes and waves, two buttons labeled “Bend”. It’s time we take a look so lets press the one on the left, associated with Oscillator 1. The top display changes to this:

Again, this should look familiar now that we’ve manipulated the envelopes and wave display but a big question arises: what does that curve manipulate? First of all, let’s play around with the different curve templates on the left.

To hear the full effect put back the oscillator 1 bend to 100%. Then let’s select “Buzz” and play:

A low frequency buzzing or beating can be heard, depending on the note played.

Now select template “Octave” and play around again:

This time the sound is harmonic but it has lost its characteristic “triangle / square” sound to become closer to saw timbre. Intriguing…

Finally, select template “Stiffness” and play:

That sounds like the original bend effect we started with and, as the name suggests, could correspond to an increasing string stiffness / tension.

Now, what exactly is going on here? Bending has the effect of increasing or diminishing the frequency spacing between partials in your oscillator’s spectrum. Said differently, it has the effect of contracting or stretching the frequencies of parts of your spectrum. With it you can mimic physical behaviors, like increased string tension, or completely crazy unpredictable stuff.

Curve points above the middle line will increase the frequency spacing and points underneath will contract it. The left start point of the curve is always the fundamental frequency of your note and that cannot be bent so as to keep your overall pitch intact. As we go towards the right side, the curve impacts higher harmonics.

With all this, we can understand the bend curves better. Lets start with the first “Buzz” curve. The lower harmonics are unchanged but higher harmonics are contracted, giving the illusion of a low frequency sound, hence the buzz or beating we hear.

With the “Octave” curve, we have the inverse effect. All harmonics except the fundamental have their pitch scaled up. We can’t really see it in the display but, believe me, they are scaled up by one octave. This is why the effect retains the harmonic character of the original sound. To position your bend curve points on octaves (either up or down), simply enable the “Snap Y” checkbox at the upper right of the curve display. This will have the effect of snapping your points vertically to octave positions.

Finally, with the “Stiffness” curve, we can see that, as we deduced from looking at the spectrum EQ window, as we go higher up the spectrum, the partial pitches are more and more stretched upwards. This, again, replicates the harmonic distortion of a very stiff string.

Ok, enough theory and curve manipulation.

Creating a bell sound

With this bending feature, it is extremely easy to create bell sounds. Just crank up the bend and add some filtering. Thankfully, there is a “Bell” envelope preset we can use both for the amplitude and modulation 1 envelopes. Here’s our try on a bell patch with bend set to 100% with the default “Stiffness” curve, the filter 1 frequency set to 700Hz and both amplitude and modulation 1 envelopes set to “Bell”. We’ve just moved the amplitude envelope’s sustain point to around 3 seconds for a longer decay. Listen to this:

I don’t know what you think of this but, although it may sound realistic, there is a slight off-tune character to the patch that I don’t really like. Let’s see how we can fix this.

Harmonize

That slightly off-tune characteristic of our patch is due to its inharmonic nature. We could counter balance this by reducing the bend but that might alter our timbre too much. Thankfully, we have a second option with the Harmonize knob.

Harmonize is a very simple but effective mean of forcing your oscillator back to musical territory. It gently pitches back all your partials to the nearest harmonic frequency. We can hear this by increasing the harmonize strength as we play:

There is a slight change of timbre but much less pronounced than with altering the bend. If you are still not satisfied, we have another trick up our sleeves. By default, the harmonize feature forces back the partials to harmonic frequencies of the current note. We can instead ask to use the harmonic frequencies of the same note but one or even two octaves below. This will keep our sound harmonic but with sub-octave harmonics. Since sub-octave harmonics are closer appart (the distance between two harmonics is exactly the note frequency, remember), our harmonized oscillator will have a higher chance of sounding like the original oscillator.

Let’s try this on our tune with harmonize set to the full value at the default “0” octave setting, then the “-1” value and finally the “-2” octave setting:

Still not satisfied?

Limit

We have another trick with the Limit setting of our harmonize feature. By default, the harmonize works on all parts of the spectrum. But with the limit knob we can decide to limit the effect to the low end thus keeping some inharmonic partials in the high frequencies while retaining the benefit of harmonize and keep a more distinct note pitch which is essentially determined by the lower harmonics.

To hear this, put the octave setting back to 0 and play with the limit by turning the knob to the left until you are satisfied. We want to keep a metallic aspect to our sound but we wish to keep it in tune. To our ears, putting the limit around 33 sounds just fine:

In the previous tutorial, we explored how to define the basic timbre of an oscillator. Although Oberon is an additive synthesizer at its core, it has a number of features that resemble the classic subtractive architecture that is the basis for most of the synthesizers from the very start. Most importantly, it also has filters.

Filters are great to quickly sculpt your sound by removing or boosting large portions of your spectrum. Classically, a synthesizer filter consists of a low pass profile which attenuates frequencies above a certain frequency value. Modern synthesizers can switch to alternate profiles like high pass or band pass.

Oberon offers two filters which are identical in nature. As we will see, each of them can alter the sound of both oscillators. But lets see how things are done in Oberon.

Frequency

This should be very familiar to you as this controls the cutoff frequency of the filter. By default, both filters have a low pass profile. You can verify this by pressing a key and playing around with Filter 1’s frequency parameter. In the following example, we’ve set the frequency value to a low value of around 200Hz. If you proceeded from tutorial 1, you should hear something like this, much less brighter than the sound we had at the end of tutorial 1.

Envelope modulation

If you listen carefully, you should notice that the attack of the sound is much brighter than the sustain part. Something is amiss with the filter frequency. There is no bug or magic there but just the effect of modulation envelope 1 modifying the filter frequency over time. Both filters 1 & 2 have their frequency parameter modulated by respectively modulation envelope 1 & 2. The amount of modulation is controlled through respectively the Mod Env 1 & Mod Env 2 knobs to the left of the filter frequency knobs. As you can see, by default there is a fair amount of modulation going on. Try setting it to 0 (middle position) and you should hear a sound that is fairly dull and static:

The modulation depth of the envelope can furthermore be modified by the velocity. This is great for dynamic bass sounds for example and is controlled by the Vel knob just underneath the modulation envelope knobs. If you set the velocity knob to full value and the modulation envelope depth back to the original value of 65, a low velocity note will have almost no envelope modulation, while high velocity notes will have the full modulation depth of 65. This is an example of what it should sound like playing notes with gradually more velocity:

Editing the envelope

To edit the modulation envelope 1, which is the one modifying the filter 1 frequency, click on the Mod 1 button of the Envelopes display section just underneath the top central display. You should see this:

Just like the wave curve display, to the left are a list of envelope templates, crazy or classic, to let you quickly set your envelopes. But of course, you can also modify them by moving the points around and adding / removing them. You should notice something new, though.

First of all, above the curve frame is a small “S” label. This identifies the sustain point. If you wish to modify this, just click on one of the rounded boxes above each point. In Oberon, all points before the sustain point are considered part of the attack while all points after are considered part of the release.

Lets try to create a more “picked bass” sound by shortening the envelope attack. You can do this by dragging the sustain point to the left or by using the quick edit button to the right of the curve display labelled “Shorter Att.”. This will scale down in time all points before the sustain point. Here’s our modulation envelope after shortening the attack:

To have even more dynamic velocity in our sound we’ve increased the filter 1 modulation envelope 1 depth to the full value and brightened it up a little by increasing the filter frequency to around 560Hz. With our new envelope, here’s our sound with increasing velocity:

Filter Curves

Now that we’ve played around with the cutoff frequency, the most acute readers among you should wonder if Oberon has any resonance setting. The answer is: yes, and even more. To the top right of the frequency knob you should see a small knob labelled “X-Y”. Try increasing it to the full value and listen to the sound:

Sure enough, we recognize the typical funky squeal sound characteristic of a high resonance value. So why call this knob “X-Y”?

In Oberon, there is no such thing as “resonance”. The synthesizer uses a more broader scheme based on two filter profiles for each filter. Each filter has an X profile and a Y profile. By default, filter 1, for instance, has a low pass profile assigned to the X profile and a resonant low pass profile assigned to the Y profile. Changing the X-Y knob towards the right (towards Y) gradually morphs the final filter between this flat low pass profile to a squeaky resonant low pass profile.

Now the interesting part is that each X and Y profile can be edited for both filters the same way as wave forms or envelopes can. This implies that you can very well decide that the X profile is a low pass profile and the Y profile a high pass profile. Lets see this in action by clicking on the X button of the Filter 1 display. Here’s how your display should look like:

All this should now be familiar as you recognize a list of curve templates to the left, with the classic low pass, band pass, high pass profiles and some other more esoteric ones. Sure enough, the X filter curve which defines the X profile looks definitely like a low pass filter. The low end is at 0dB gain. Every point above it creates a gain (like a resonance), while every point underneath creates a cut (like a notch). In fact, if you are accustomed to graphical equalizers, you should feel right at home.

At the center of the display, labelled “0 semi” is the cutoff frequency. Everything to the right is above the cutoff frequency and everything to the left are lower frequencies. The labels “-4 oct.” and “4 oct.” indicate that at this horizontal position, the frequency is 4 octaves lower or higher, respectively than the cutoff frequency.

Lets switch to the Y curve by clicking on the neighboring Y button:

The peak is our resonance and as we can see, it is perfectly centered on our cutoff frequency. This is not an obligation and we can easily change this with the usual point dragging and “Add / Remove point” features. Lets try something complicated with two resonance-like peaks like this:

With the filter 1 X-Y knob fully turned to the right and with a slight global volume attenuation to prevent clipping, this is what it sounds like:

Now with the X-Y knob on filter 1 we can go from a normal low pass filter (approximately sounding like a 24 dB/octave filter) to a low pass filter with character. If you wish you can link the modulation wheel to the filter 1 X-Y parameter with the modulation matrix.

Repeat

There is another little thing we can do with filters and that is to repeat the filter response. You may have noticed this weird “Repeat” button under the filter curve to the left. Pressing this will change how the curve is applied as a filter response. When activated, the response is repeated across the frequency spectrum with a period defined by the filter frequency. For instance, if we activate repeat on our current filter setting with a filter frequency of 550 Hz, the filter curve controls the gain and attenuation of the spectrum for a 550 Hz wide band that is repeated every 550 Hz.

Again, using the spectrum EQ window will makes things clearer but here is a diagram showing the differences with non repeated filter curves first then repeated filter curves.

Repetitive filters like this are not uncommon in classic subtractive synthesis. Comb filters behave exactly the same way with repetitive peaks or repetitive notches.

Lets listen to our patch with repeat activated:

Filter Modes

To finish off this tutorial on filters, we will take a look on how both filters are applied on oscillators.

By default filter 1 & 2 are applied in sequence / serial. Pressing the button over the “Filtering” label between the oscillator and filter sections reveals a popup menu with all the available filtering modes for each oscillator.

For performance reasons, it is sometimes interesting to switch to no-filtering or to a single filter only if the other has no effect. Just to be tidy, we will switch to Filter 1 Only. Hopefully the sound should remain the same!