Hopefully you’ve had the time to read my articles on “Mixing with the Professionals” and “Mastering with the Professionals” at this point, and have really dug in and tried out some of the concepts. This article will be a bit different than the previous two - I’d like to introduce several forms of synthesis and then go over the creation of a few common types of sounds you hear in modern electronic music.
The following is an expert guest post by Jim Emerson. Jim holds the title “Executive Audio Engineer” at Reasonzone Audio. With a deep-rooted passion for his work, he has had the opportunity to help musicians around the world realize their creative potential.
Table of Contents:
Unlike my previous articles, this topic isn’t as easily grouped into chapters, but I’ll do my best:
- Chapter 1: Basics of synthesis
- Chapter 2: Common types of synthesis
- Chapter 3: Sound creation
- Chapter 4: Sound enhancement
As you continue reading, I understand that the article may start off a bit basic, but I feel it important to establish a few simple concepts for anyone that isn’t familiar before proceeding. If you keep reading, I’m sure you’ll find something beneficial!
Chapter 1: Basics of synthesis
I will attempt to be brief here, because I’d like to focus more on the creation of a few popular types of sounds than bore you with the details behind how synthesizers work. Nonetheless, a few key points can’t hurt to go over.
Oscillators and basic sound waves
Before you can understand synthesis, I believe it’s important to start with synthesis at its simplest. Have you ever taken the end of a garden hose and shaken it up and down to make those waves travel along its length? You’re wiggling a garden hose, but an oscillator wiggles an audio signal. Sounds that have a definite pitch will have a repeating waveform, and an oscillator is the tool we use to make these waveforms. This signal will then commonly travel into a filter bank, various effects, envelopes, and then to an output. I’m sure an overwhelming majority of the readers are already aware of this signal flow and terminology, but it’s helpful to introduce the idea of basic waveforms before proceeding. A basic oscillator will produce a sine wave, triangle wave, square wave, or sawtooth wave.
Sine waves, as you can hear, are the simplest and most pure waveform shape. There isn’t much to say here, other than the fact that a sine wave is only one frequency, and is the basis for all of the other basic wave shapes. As a side note, because a sine wave is so simple and pure, there isn’t much of a point to compressing them. Some may compress a sine wave sub-bass, for example. This is totally fine and if it sounds good, you should go for it, but all you’re really doing is adjusting the envelope here. You can just as easily go into your synthesizer and edit the attack, decay, sustain, and release of the sine wave to (likely) get the same results.
Triangle waves sound like a cross between a square wave and a sine wave. If you played several sine waves together with frequencies at the odd harmonics of your fundamental frequency, you could build a triangle wave from scratch. For example, a sine wave played at a fundamental frequency of 100Hz, added to sine waves at the odd harmonics of 300Hz, 500Hz, 700Hz, 900Hz, and so on could create a triangle wave. Similarly with square waves - these are also composed of several sine waves at the frequencies of odd harmonics of the fundamental frequency. The difference is the volume of the odd harmonics in a square wave don’t drop off as quickly as they do in a triangle wave. The volume of each individual harmonic is extremely important, but honestly, you don’t need to know that math with any modern synthesizer. In other words, simply playing a bunch of sine waves at these frequencies probably won’t create something that sounds like the triangle or square wave coming out of your synth, but they’re in the right spot and just need the amplitude adjusted.
The most harmonic rich waveform is the sawtooth wave. These are composed of a set of sine waves starting with the fundamental frequency and continuing up the spectrum at each harmonic with a decreasing volume. For example, an ideal sawtooth wave with a fundamental frequency at 100Hz would have sine wave harmonics at 200Hz, 300Hz, 400Hz, and so on. As with the triangle and square waves, the volumes of the individual harmonics matter a great deal in determining the timbre or character of the wave, but that math isn’t necessary with modern synthesizers.
This is obviously an extremely rudimentary run-through of these basic concepts, but they may come in handy later when we’re actually creating and enhancing sounds.
Chapter 2: Common types of synthesis
There are many different types of synthesis: too many to list and explain. Instead, it makes a ton more sense to go over the ones you will likely encounter and leave it at that. With that being said, let’s jump right in! I’ll keep these brief.
Additive synthesis, while not the most common, is definitely prevalent in the production world. FL Studio comes with Harmor, a powerful additive synth, and Propellerhead has Parsec. These types of synths work in very much the same way as the earlier discussion on sine wave harmonics: additive synths add sine waves together to create their sounds. This method of synthesis can be a bit more difficult to master, but the sounds you can create are often very unique.
I like to think of additive synthesis as creating something from nothing, and employ a pottery-maker metaphor. Obviously it isn’t “from nothing”, but as a pottery-maker adds and shapes clay to form a pot, additive synths add partials, or individual sine waves together until the desired sound is achieved.
Subtractive synthesis is arguably the most common form of synthesis today, though wavetable synthesis (discussed later) is catching up in popularity. Synths such as Propellerhead Thor or u-he’s Diva utilize a powerful subtractive engine to create a wide variety of sounds. Subtractive synthesis works on the principle that a harmonically rich audio signal is attenuated by filters to change the timbre and character of the signal until you’ve reached the desired sound.
The easiest way to describe subtractive synthesis is to compare it to the process of sculpting. A sculptor starts with a block of marble and chips away at it until his work is finished. This process is very similar to a subtractive engine - instead of adding partials together, a more simple approach is employed, often making subtractive synthesis a quick and efficient way to approach sound design.
Granular synthesis is not very common, but is extremely powerful in the right hands. Granular synthesis can be explained with a rock quarry: rock is mined and broken down in a quarry, and the small broken down pieces are then used to create concrete and concrete structures. In very much the same way, granular engines break down sounds and redistribute the “grains” to form a desired sound.
Wavetable synthesis, as I mentioned earlier, is becoming increasingly common. Synths such as Xfer’s Serum and Propellerhead Thor (again) use this powerful synthesis engine to create a lush plethora of sounds. This type of synthesis engine uses a bank of existing sound waves from instruments and other sources to form new sounds. These sound banks are combined, stretched, and otherwise manipulated for an endless amount of creativity and control.
I may be reaching a little bit, but I like to think that the process of wavetable synthesis is very much like the process of repurposing old pallets as furniture or old wine bottles as vases. Get on Pinterest, you’ll see some examples of what I’m talking about. This repurposing to me is a very easy way to visualize what’s going on.
Chapter 3: Sound creation
In chapter two, I glossed over a few types of synthesis that are popular, but not within the scope of this article to explain. For example, FM synthesis is relatively popular and powerful, but anyone that uses it already knows what it is, and I won’t be using it in any of the following sounds. With that being said, I found it difficult to choose which types of sounds I wanted to go through the process of creating in this article. Eventually I settled on a solution: I’d like to use each type of synthesis mentioned above to create several patches that you might find in a modern electronic track. A couple different bass patches, some pads, a patch for supporting harmonic content, and a lead for melodies and accents.
Where else to start but the foundation of a good song (in my opinion)? The low end is arguably one of the most important aspects to nail down - this is what your listeners will physically feel, and as we move up the register from sub-bass to bass, set the tone of the track. Obviously, vocals, melodic structure, key signatures, etc., all “set the tone” so to speak, but I feel that a lot of the mood of a track comes from the bottom.
So let’s start at the bottom. A simple sub-bass to layer beneath your bass patch allows better control of the low end. This is very straightforward, but in tune with the beginning of the article and my philosophy of starting basic, this is the most logical way to begin. I’ll be using Propellerhead SubTractor for this, as it’s a very straightforward and powerful subtractive synth. Though I’m in SubTractor, any synth that can produce a sine wave will work just fine here. As you can see in the photo below, I have oscillator 1 on a sine wave at octave 4, and oscillator 2 on a sine wave at octave 3.
SubTractor Step 1:
Next, I turned the amp envelope sustain all the way up, and added a small amount to the amp envelope attack to get rid of the “pop” at the beginning of the note. As you can hear, the increase in sustain causes the patch to retain its volume over time and the increase in attack will prevent any clashing later on when this patch is mixed with a kick drum or other elements with sharp attacks.
SubTractor step 2:
I said I wanted to start simple, so there it is. Instead of compressing the sine sub-bass to make a “full” sounding patch, adjusting the amp envelope gets us the sound we want. Afterword, depending on taste, you may want to set the synth to legato and add some portamento for a more “gliding” sound between notes, but that’s personal taste.
Moving on from the sub-bass, now we can create a bass sound that adds some character to the track. For this, I’d like to use granular synthesis. No wait, wavetable. Actually, I’ll be using a mixture of the two - Propellerhead Malström. This synth is termed a “graintable” synth, which is a mashup of the two types of synthesis pioneered by Propellerhead. As you can see in the photo , I’ve set oscillator A to “Sawtooth*16”. Take a listen with the default settings.
Malstrom Step 1:
Alone it already sounds pretty cool, but I want to add some complexity to the patch. Turning on oscillator B, I‘ve changed the graintable to “TibetanMonks” and turned the motion knob all the way down. Then, I swept around the table, listening to individual grains of the sound until I found one that meshed well with the sawtooth. As you can hear, this adds an interesting layer to the bass as the vocal formants from the graintable blend in with the rest of the sound.
Malstrom step 2:
If you’d like to hear what just oscillator B sounds like, check out these clips:
TibetanMonks graintable with full motion:
TibetanMonks graintable with -64 motion and at an index of 77:
As you can hear, this type of synthesis is extremely powerful, allowing you to zone in on particular vowels or consonants, or just particular slices of a wavetable in general. This bass patch still isn’t where I’d like it, however. I’ve sent oscillators A and B into Filter B, and applied a filter envelope with sustain all the way down, and a moderate decay. This adds a bit more of a punch to the beginning of the patch, and makes it more of a plucky sound. In the following audio clip, I’ve layered this now plucky patch with our sub-bass and created a simple bass melody.
Malstrom step 3:
With this setup, the filter envelope sustain and filter position can be automated up for swells and more pad-like bass sounds. Whether or not these two sections will be next to one another in your song, I don’t know, but I felt it important to showcase a little flexibility with the patch.
Malstrom step 4:
For this patch, I’d like to use Propellerhead Thor. It’s a workhorse, with several different types of synthesis engines built in that can be used at your discretion. For this patch, I’ll be using subtractive synthesis and wavetable synthesis. As you can see and hear, I have two oscillators activated, passing through a standard lowpass filter. The grey oscillator uses a subtractive engine, and the blue is a wavetable oscillator. Together, they sound a little weird at this point, but we’ll fix that.
Thor step 1:
I’ve changed the filter from a lowpass to a notch filter, and sent the oscillators through a parallel filter, also set to notch. As you can see, the filter frequencies and notch settings are inversely set. In other words, both oscillator 1 and 2 are traveling through filter 1 to the output, and they’re both traveling through filter 2 and to the output. As far as the output signal is concerned, it is 4 separate oscillators stacked up on top of one another, when in reality, we only have two.
Thor step 2:
Next, I’ve automated LFO 1 to oscillate each bar and shift the frequency values of each filter, created a third oscillator that bypasses filters 1 and 2, and highpassed the entire patch with filter 3. Alone, it will sound thin and weak, but when layered with our sub-bass and character bass, it should add some nice movement to the entire bassline.
Solo accent bass patch:
Mixed in with our previous basses:
Pads are often overlooked by new producers. Often this leads to tracks that don’t sound as full or “professional” as they could be. Pads are an extremely important layer of a song: picture a bucket full of large rocks. These rocks are your basslines, melodies, vocals, drums, etc. All of the important elements. Adding pads to a song is like pouring sand into this bucket - it fills all the space between the rocks and fills the bucket the rest of the way. I apologize for using a dated metaphor, but it’s very apt in this case. Anyway, let’s get started!
I’ll first create a pad that adds supporting harmony to the audio clips. For this, I’m going to use subtractive synthesis again. Particularly, I will use Synapse Antidote. Antidote is one of the most powerful synths on the Rack Extension market, and can create huge, lush pads. First, I’ve separated oscillator 1 and 2 in the stereo field - one of them is panned all the way left and the other all the way right. Thereafter, I’ve set them both to “Analog Saw” and set the count to 12. At this point, there are 24 oscillators playing at once. Oscillator bank 1 is set to +7 cents and oscillator bank 2 is set to -7 cents. This further detunes the overall sound and widens the pad out.
Antidote step 1:
Now, I’ve turned the “DYAD” function for each oscillator back on to +24. This now means that the entirety of the oscillator banks playing have been duplicated and transposed up 2 octaves. As a sidenote, this means that in the audio clip, each individual note is activating 48 separate oscillators. Since the chords playing are dyads, this means that each chord is activating 96 oscillators. That’s a lot of power.
Antidote step 2:
With a slight “drift” added to simulate an analog synthesizer and some filter envelope adjustments, this pad is nearing completion for now. Later in the article, I’ll go over some post-processing for each patch that needs it, but for now, here’s our harmonic pad:
And mixed with our basses:
For this patch, I’d like to finally utilize additive synthesis. The purpose of an atmospheric pad is to add depth and air to a track. Additive synthesis is great for this because of the unique and evolving sounds you can easily create. I will be using Propellerhead Parsec. First, I will set the generators. These are the equivalent of oscillators in additive synthesis. For this patch, I’m setting both generator A and B to “Dual Saw”, and changing the detune to 50% on A and 100% on B.
Parsec step 1:
Next is the most important step - the sound engine modulators. As you can see, both sound engine A and B are set to unison and then highpass. This drastically changes the sound from a full saw wave to an airier and thinner pad-like sound.
Parsec step 2:
Finally, the amp envelope and reverb can be altered. For this sound, I liked how it sounded to turn the attack, decay, and sustain all the way up, and set the release around 4 seconds. Then, I set the built-in reverb to large size, long decay, minimal damping, and high level.
Parsec step 3:
Mixed in with the rest of the song, you can hear how it adds some supporting harmonics, but the primary focus is the wavy, airy, spaciousness near the top of the spectrum:
Without atmospheric pad:
With atmospheric pad:
Supporting harmonic content
All this means to me is something like a “keys” type of patch that plays your chord progression. In this instance, I’ll be creating some plucks with subtractive synthesis that will play the progression along with our bass and pads. This sound will sit on top of the pads and be the focus of the track at this point.
To start, I’m loading up an instance of Korg Mono/Poly. This synth is a true emulation of the vintage Korg Mono/Poly hardware synthesizer, and is perfect for pluck sounds. The first thing I did is turned the unison and voicing all the way up and lowered the volume respectively. Thereafter, I copied over the pad channel, turned it into an eighth note progression, and changed the notation from dyads to triads.
MonoPoly step 1:
Now, to spread the oscillators out a bit and turn this into a pluck. I left oscillator 1 alone, increased oscillator 2 one octave, changed oscillator 3 to a square wave, and changed oscillator 4 to a triangle wave (+1 octave). After this, I lowered the filter cutoff down to around 3, increased the filter resonance to about 2.5, and turned up the filter envelope intensity to around 3. Then it was just a case of changing the filter envelope decay to around 2.5 to get the plucky sound you hear below.
MonoPoly step 2:
Here, you can hear this pluck with a filter automation playing with the rest of the layers so far:
I’m not very good at melodies, so bear with me on this lead patch. I’ll be using a combination of wavetable synthesis and subtractive synthesis with Thor. This patch is a bit more complex, so again, bear with me on this. First, I’ve set oscillators 1 and 3 as wavetable oscillators, and sent them straight to filter 3, bypassing filters 1 and 2. They’re both set to “Basic Analog”, with differing wavetable positions, and each is only sent to either left or right filter 3 input. This means that oscillator 1 is only playing on the left speaker, and oscillator 2 is only playing on the right speaker. This, paired with the differing wavetable positions, creates a nice wide sound to start with.
Thor step 1:
Next, I’m activating oscillator 2, setting it to a subtractive engine (sine wave), and sending it to both filter 1 and filter 2. Filter 1 is a basic lowpass with a cutoff of about 8kHz, and filter 2 is a formant filter, imparting some vowel tones to the sound. Filter 1 is sent into the shaper on “Wrap” mode, and then sent into filter 2.
So, we have:
oscillator 2 > filter 1 > shaper > filter 2 > filter 3
oscillator 2 > filter 2 > filter 3
Filter 3 is receiving signal from oscillators 1 and 3 directly, as well as the summation of filter 1’s version of oscillator 2 and filter 2’s version of oscillator 2. Here you can see the patch and hear what it’s like so far.
Thor step 2:
Next, I’ve routed the filter envelope to the oscillator 1 and 3 input volume at inverse levels and modified the amp and filter envelopes. This sounds confusing, but here’s what is happening: as you can see in the following photo, I’ve set the “scale” for oscillator 1 and 2’s input to filter 3 as the filter envelope. They are at opposite values, meaning that oscillator 1 volume is inputting to filter 3 at the scale of the filter envelope, and oscillator 2 volume is going in at the inverse of the filter envelope. This makes oscillator 1 bump up as a pluck sound going in, and oscillator 2 have a somewhat “sucking” sound coming in. It simply adds some more width and creates a nice stereo effect.
Thor step 3:
Finally, I’ve turned on the built-in chorus and delay, set the patch to mono legato with portamento on, and created a tentative melody. The melody sounds weird by itself, but accents the pluck and pads nicely when blended:
This is obviously not supposed to be exhaustive by any means - I wanted to go over the creation of a couple patches in hopes that you can take away something other than how to make that exact patch. My goal isn’t to say “these are the patches you need to know how to make”, but rather illustrate some techniques that can be applied to any of your own patches.
Chapter 4: Sound enhancement
This section will be relatively straightforward. The patches we made in chapter 3 were basic in that they were straight from the synthesizers. In this section, I want to briefly touch on a couple techniques to enhance what’s coming out of the synths and polish them up. Let’s start with the basses.
Here I mainly want to focus on the “Character Bass”. The “Accent Bass” adds some depth to the low end, but I want to add a bit of reverb anyway. A common piece of advice thrown around is to avoid reverb in the low end at all costs. I’m not arguing - reverb in the low end is a surefire way to muddy up the mix. However, if we add it carefully and EQ properly, we can really add something to the track. First, I created a parallel “Character Bass” channel. Now there are two copies of this signal playing. On this parallel channel, I’ve inserted an instance of Softube’s TSAR-1 True Stereo Algorithmic Reverb. This unit is insanely lush and full sounding, and is always my go-to reverb. As you can see, I’ve set the density to “High” and turned the “time” all the way down. Then I shifted early reflection mix towards “early Refl.” and away from “Reverb”. Finally, I highpassed this mixer channel around 250Hz to get rid of the low end mud.
The first 4 bars of this audio clip is without the reverb, and the second 4 bars is with:
What a huge difference! In my opinion, the only other thing that could benefit the low end at this stage is tasteful compression with Cakewalk’s RE-2A Leveling Amplifier. This is a faithful emulation of the original LA-2A compressor, and sounds great on basses.
Cakewalk step 1:
Again, this should be pretty straightforward. All I want to do here is match the depth of the basses to tie these layers together and add a bit of airy reverb in the background. I start by creating a parallel channel of the pad bus. I’m not creating a parallel channel of either the harmonic pad or the atmospheric pad, but of a channel that has them merged. On this parallel channel, I’m simply copy and pasting the TSAR-1 reverb from the bass parallel channel earlier and adding back the highpass filter. This channel will help to unify the bass and pad channels, and create a more cohesive mix.
The first 4 bars of this audio clip is without the reverb, and the second 4 bars is with:
After this, I want to create a bit more detail in the pads. I’m employing the use of one more parallel channel containing an instance of “Echobode Frequency Shifter Delay”, by Sonic Charge. On this unit, I’m actually using the factory preset “Dream Delay”, which alone is enough reason to buy the Rack Extension. As you can hear, the summation of these changes makes a large difference in the audio quality and depth. The first ~23 seconds of this audio clip is the basses and pads together with no enhancement, and the rest of the clip is the basses and pads with enhancement:
I’ve normalized the perceived volume of each clip so you can accurately hear the differences and determine for yourself which you think sounds better.
Finally in this cursory run-through of sound design, we arrive at the enhancement of the lead and pluck patches we created earlier. Starting with the pluck patch, I’ve added u-he’s Uhbik A Ambience Processor on a parallel channel to the pluck. If I had to pick a second favorite reverb unit, it would be this one. First, I’ve turned the dry/wet all the way up and set the density to 100%. Then, I set the reverb to 100% late and changed the “Operation” value to “Open”, as well as turned the spread all the way up. I know these sound like extreme settings, and they are, but we will remedy this in a minute.
This is what the reverb channel sounds like alone, after a channel highpass:
After this, I sidechained the reverb to the original signal and adjusted the volumes. Sidechaining the reverb to the original signal means that when the pluck is playing, it pushed the reverb out of the way based on the compressor settings I dictate, but cames back into focus when the pluck isn’t playing. This allows the pluck “plucks” to push the reverb out of the way during the fade in at the beginning and ensures that the reverb doesn’t drown out the dry signal when the filter is open. Here is the before and after.
The first 8 bars are dry, and the last 8 are with the sidechained reverb:
That sounds pretty good to me, so I moved on to the lead and enhancement thereof. All I wanted to do here was create cohesion, so I created a parallel channel of the lead and copied everything over from the pluck.
Here you can hear everything together. The first 24 seconds are without the enhancements, and the rest of the clip is with:
The art of sound design is a huge topic. Way bigger than can be covered in a single article. My goal here was not to teach you individual patches, nor was it to adequately explain the different types of synthesis and when to use them. All I wanted to do here was show you some techniques that you can transfer over to your own sound design process.
Please remember that I am an audio engineer, not a composer. I understand synthesis and dabble in production, but I’m no composer. My chord progressions are simple and my melodies are plain. I know this, but I wanted to try and build somewhat of a listenable loop so you, the reader, can understand how these layers lock into one another and use this in your own work. If you use Propellerhead Reason, included is a download to the Reason file I used for this (see button below). You will need several Rack Extensions to make it work, but I highly recommend all of them.
If you enjoyed this article, stand by for one more entry: I will be going over a ton of different production tips and tricks to make your music stand out!