By Trent Rentsch
To say that a synthesizer creates a sound “from scratch” is about as true as a magician explaining his mysteries with the chestnut, “The hand is quicker than the eye.” In both cases, there are several elements in play that actually accomplish the finished effect. For the magician, misdirection is really the key element. For the synthesizer programmer, it is the oscillator.
What is an oscillator? Think of it as the first ingredient in a sound concocted in a synthesizer. It is a base sound, which can be altered and/or mixed with other oscillators to create the desired sound. There are many unique forms of oscillators available, but the standard ones are analog waveforms (and the digital versions you’ll find in a software synthesizer), and samples. Yes, samples, just like the ones you record in your audio software. One might go so far as to say you’ve already done synthesis (of a sort) with a sample based oscillator if you’ve ever altered a voice-over or sound effect to give it a different sound with the processing in your editor.
All synthesizers have one oscillator, most have 2 or more, some may have more than one type. Unless you’re trying to recreate a simple, punchy sound, the more, the better, as combining the output of several oscillators with different settings is a great way to create brand new noise. And of course, different types will yield different results.
Early analog synthesizers relied on a signal with a simple repeating pattern, which was to be known as a waveform. Frequency was (and is) changed by the speed of the pattern. Even in modern software synthesizers, you’ll find a switch to change waveform type, and another to change the frequency, which is really a speed control. Most also include a way to go from waveform to waveform, to create hybrid waveforms, or simply take advantage of the differences in character between different types within one sound.
The main analog waveforms are the sine (think of it visually as rounded hills and valleys, triangle (pointed hills and valleys), saw (still pointed, but the hills and valleys are angled), the square (ahhh… square hills and valleys), and pulse (still square, but the distance between the hills and valleys is noticeably different). Alone, they’re pretty boring sounds. The sine is just a dull tone, the triangle adds a few overtones, the saw is simply bright and buzzy with lots of harmonics (especially noticeable are the higher ones), the square only contains odd-numbered harmonics for a hollow tone. The pulse is a lot like the square, except the width between the top and bottom ends of the cycle are different, which changes the nature of the sound.
It might sound confusing on paper, but when you get in front of a synthesizer and listen to the differences between waveforms, you’ll begin to hear how each can be applied to add different character to a sound (an example of each appears in my current podcast at slysounds.com). As you listen, listen to the waveform at different frequencies, as this opens up the potential for new types of sound as well.
Digital synthesizers have opened more worlds of waveform oscillators. You’ll find many different waveforms… combinations of the original analog waves, completely new waves, even the ability to create your own. Many also let you include noise as a waveform, which opens up even more sonic avenues (think waterfalls and snare drums, for example).
The other oscillator world digital synthesizers created was that of samplers. Any sound could now be sampled (recorded), modified (processed), and played (played back). As I mentioned earlier, it’s a process that is familiar to anyone who works with digital audio. The advantage for musicians has been that, instead of a synthesized reproduction of an instrument, they can play the “real thing,” via a sample. Still, it has its drawbacks. Let’s say you’re using one piano sample, one note, middle C should do it. Now, as the musician plays up and down the keyboard, the notes should go up and down, right? Yes, but… and it’s a big but. Because sampling is changing the frequency of the sample for it to sound higher or lower, it’s basically changing the speed of the note. Imagine that C a few octaves down, plodding out a slow growl, with a few octaves higher it plays so fast it’s barely noticeable. The simple way around this issue was to come up with multi sample units, which play different samples depending on what key you play. Some even change the sample depending on how hard you hit the keys. Great solution, but all those samples become memory hogs. Now, none of this may be important to you if all you want to do is make out of this world and/or garbled noise, but it’s something to keep in the back of your mind.
Again, waveforms and samplers are the two most basic oscillators you’ll find on synthesizers, and the frequency (or speed control) is the most basic way to alter the sound of the oscillators. You may find a synth with a sampler oscillator which lets you redraw the sample, you may find one which lets one oscillator change the tone and characteristics of another oscillator, you may even run across one which lets you use the input of a microphone or any other electronic instrument as the oscillator for the synth. No matter how bizarre the configuration, just remember that the oscillator is always the base of the sound you are creating, and the more oscillators you have available, the more complex the sound you can create… like the difference between one Lego and a box of them.
How do we bend that basic noise into what we want? Like the wily magician, programmers have some tricks up their sleeve and I’ll share some of them next month.