Synthesis Basics

by Lorraine Quirke

Back to main page

Sound is created by synthesisers in two ways:

• digitally
• analog

Analog synthesis is when voltages control one another to create sounds. Sounds are created by electrical circuits that are sound generators or oscillators.

Digital synthesis is when the voltage levels produced by an electrical circuit oscillator in an analog keyboard are instead mathematically generated by a CPU. (sine waves are the easiest to generate, versus a sample which needs lots of storage space)

Sampling is another digital method of modifying sound, that bears mentioning in this context. Sampling is digital recording of an external sound.

Synthesis, whether digital or analog is similar in that sound elements are put together to create complex sounds.

These elements, in the earlier digital synthesisers, just consisted of sine waves. Later synthesisers use combinations of sine waves and samples to create complex sounds.

Sounds are shaped in several ways:

Envelopes are ways to shape a sound's amplitude

Here are contrasting examples of envelopes:

• a drum sound: has a sharp attack, fast decay, no sustain, no release
• a violin sound: has a slow attack, slow decay, medium sustain and medium release
• a distorted electric guitar: has a sharp attack, no decay, long sustain and short release

Here are some examples of timbres:

• sine wave: thin and flute-like
• triangle wave:
• sawtooth wave: sharp: violin-like
• square wave: reedy and hollow: clarinet-like
• white noise: FM radio static between stations

Partials are the components of a sound. Harmonics are partials that are mathematically related to the fundamental. They are multiples of the fundamental, eg. twice the frequency, four times, eight times, and so on. If a sound is consonant, they contain harmonics in a series, either even or odd or all harmonics, and they are not all the same volume.

The fundamental harmonic is the lowest in the series, and it gives the note its musical name, for example its frequency (eg. 440 hz) determines the perceived musical note (concert A)

All other harmonics add character to the sound. They transform it from a simple sound, such as a sine wave, to being a complex sound, such as a distorted guitar sound.

If the harmonics are the same volume or louder than the fundamental, the sound no longer has a distinctive perceived musical pitch, such as is the case with a crash cymbal.

Harmonics are what are filtered out when you use an equalizer or bass/treble controls on a sound.

Harmonics are used in drawbar organs to create different sounds. You can customize sound by pulling out the drawbars to different lengths, the farther out you pull it, the louder that particular harmonic is. Pipe and electric organs both use this principle for you to create custom sounds.

On a DX7, the six operators can be set to different frequencies, to act as different harmonics, if they are playing unmodulated. If they are modifying one another in the formula sin(a+sin(b)) then additional harmonics are created in more complex patterns, if two operators are modifying one another, more than two harmonics are heard.

Components of a Patch

Various keyboards use different methods to create sounds. Some keyboards use control voltages in combination to make a sound -- generally analog or analog-type keyboards. The Yamaha DX series of keyboards use sine waves of different frequencies mathematically modifying one another, or played at the same time. Other keyboards have several samples of real instruments that you can play at the same time, add reverb and flange effects to.

How it applies to programming keyboards

Some keyboards, when you are programming patches, will let you have more control over the sound, some only let you change a few parameters. Generally, here is what to look for:

• Several different "voices" or elements to a sound. These make the partials. The loudest will be the fundamental, the partial that gives the sound a perceivable pitch (musical note)
• ADSR (attack decay sustain release, envelope generator) envelopes for each "voice" - how quickly it fades up, then down. Play with envelope generators for each harmonic, if you want the timbre to change over time.
• Frequency control is octave or pitch control for each voice - to control general high or low overall pitch of a sound, different pitches for the harmonic partials give the patch a distinctive sound (its timbre) The frequency of the fundamental determines what octave the note is in, or in other words, whether it's low-range like a bass guitar sound or high-range like a flute or mid-range like a piano. The harmonics are generally higher frequencies than the fundamental, in order not to overpower it, however, don't be afraid to make exceptions and experiment.
• LFO (low frequency oscillator) control to make vibrato (when controlling pitch) or tremolo (when controlling volume) it's an oscillator that doesn't actually make sound, it controls a voice that does make sound.
• Scaling is when one voice is louder for the higher notes or lower notes - this can give a patch a lot of character: eg. a piano patch may have some of a certain harmonic only in the upper midrange, and at the very top of the keyboard most of the harmonics will drop out giving it a thin sound at the top notes.
• Harmonics vs Fundamental: Fundamental should be loudest in relation to the other harmonics, most of the time, so that the listener can distinguish the note's percieved pitch. The other harmonics add "timbre" or color to the sound. "Timbre" refers to the recognizeable sound of a sound- what makes it sound like a piano, or whatever else. The fundamental should also be the lowest-pitched voice, except in certain cases like if you are making a clarinet or oboe sound, in which case the mid harmonics are louder than the fundamental, which is what gives it its nasally sound.
• Detuning can add a warmth or special effect to a sound. This is detuning of one voice in relation to another, for example, don't touch the fundamental, but detune the harmonics slightly. A slight detuning may add warmth to the sound. Some keyboards may especially need it, like the DX7, whereas the keyboards with built-in reverb and if the patches are built with samples, the warmth of the sound comes from that. More extreme detuning can give you a honky-tonk piano or toy xylophone effect. Most "real" instruments are not precisely in tune, so when you use detuning you are making it feel less like a machine-made sound. It works both ways, if you want a cold sound, then maybe stay away from mild detuning.
• Touch Sensitivity is the ability of the keyboard to sense when you are playing the note harder or softer. The patch may be able to be programmed to respond, like for a slap bass guitar sound, the higher frequencies would be louder (the brightness of the finger hitting the string against the frets) the harder you hit the note. Some keyboards have Velocity Sensitivity, which is the ability of the keyboard to determine the speed at which you press down the note. Sometimes this can be confusing but they are two different things.

The amplitude of an operator can be modified over time (with an envelope generator) to create sounds that change over time. The fundamental will be louder.

Knowing how sounds are made will enable you to have a lot of control over what sound you actually are able to create by synthesis, and will enable you to re-create favourite types of sounds on new keyboards, and to customize existing sounds on pre-programmed keyboards. You can also break all these rules to make some really weird sounds but I hardly think that I needed to say that.

Learn as much as you can about how sounds are made on acoustic instruments, then you will be able to re-create them on a machine, your friends will think you are really cool. Then you can add to a natural-sounding sound something that is really weird. People notice when you put a new twist on something that is familiar.

Back to main page
Contact: Lorraine Quirke