Synthesizers: an overview
Created | Updated Nov 10, 2008
Are you having trouble with your new techno track? Yes? Maybe you ought to stop faffing with that ludicrous oboe and start getting to grips with a synthesizer instead. Synthesizers are electronic sound creation devices often found in the form of a keyboard type thing with hundreds of intriguing knobs and flashing lights on them. Synthesizers are also found in the form of weird little boxes, rack-mount units and pieces of software.
How do I make it go?
Steady there. You might want to plug it in to a few things first. You will find a hole or two on the back marked 'Output'. Liberate the correct audio cable from your big mess of cables and use it to connect the synthesizer to a mixer or amplifier. If you are using a piece of software, or soft-synth, the sound will come from of the soundcard's output. All you need to do now is trigger a sound. To do this press a key on the synthesizer's keyboard or send the synthesizer a note-on message using MIDI. MIDI stands for Musical Instrument Digital Interface and is a protocol that has been the industry standard for transmitting parameter values between synthesizers and sound modules since 1983. If you are using MIDI to trigger the synthesizer externally you will need a sequencer of some sort and the MIDI cable to connect it to the synthesizer. Once the note message been received a sound signal will be produced, this is the equivalent of where a guitar or piano create sound with a selection of vibrating pieces of wire displacing air at different frequencies, synthesizers use oscillators, essentially a pieces of circuitry that create cyclical signals at various frequencies, to be reproduced by your giant atmosphere-shunting funk cones (Speakers)
That sounds ghastly!
It sounds a bit raw, that's all. It's time we made a few changes to the sound. To do this we need to know what parameters we can control. Let's start with the most common one. The envelope. The envelope controls one aspect of the shape of the sound triggered. This aspect is the sound's volume over time. To do this it uses the (almost) standardised system of four parameters called Attack/Decay/Sustain/Release or ADSR. Each of these parameters affects the volume of the note from the time triggered in a different way.
- Attack changes how long the sound takes to get to full volume. A high attack parameter will create a sound that gently rises in volume. Lets turn that parameter down to zero. Now the sound is at full volume as soon as it is triggered.
- Decay changes how long the sound takes to drop to the level of the Sustain parameter. Let's turn this parameter down quite a lot but not all the way. Now the sound drops quickly to the sustain level. What is the sustain level?
- Sustain is the volume the note will play at as long as the key is pressed down or the MIDI message is on. Let's leave this at half way. Now when we press the key we get a sharp peak in volume and a medium volume note for as long as we hold the note down. When the key is released the Release parameter comes into play.
- Release is the parameter that effects how long the volume takes to fade right down to zero after the note has been released. Makes sense doesn't it? Let's turn this right the way up. Now the sound gently dwindles away to nothing when the note is turned off.
It still sounds ghastly, but now its shorter
Excellent! We're making progress. The reason the sound has remained 'ghastly' is because the character or timbre of the sound hasn't actually been changed. It is the process of changing the timbre1 that gives synthesizers their name. There are ways to alter timbre that are not strictly synthesis techniques and collectively these are called effects or signal processors. Many synthesizers have these effects, such as compression, flange or reverb, built in but essentially effects are a peripheral way to alter the sound. To make changes directly to the sound's timbre we need to know which type of synthesis we are using.
I care not, just make it stop!
Subtractive synthesis is the most common form of synthesis and is often used in combination with other forms. The process works by using a number of oscillators that create a variety of waveforms. Each different waveform has different harmonic properties that produce different sounds. Most synthesizers have at least two oscilators that produce these waveforms at various volumes or phase positions. This is done because a combination of waves in turn effect each other to further alter the harmonic structure of the sound. This process is referred to as additive synthesis and is used to produce a wide range of different timbres. In addition to the different timbral effects of combining different wave forms each oscillator may be tuned (or detuned) independently and often features a variety of other means to modify the sound such as wave sync2, stereo phase3, pulse width modulation4 or internal ring modulation5. Most subtractive synthesizers are first additive unless they have only a single oscillator. The basic waveforms are:
- Saw-tooth. This waveform is especially well suited to subtractive synthesis as it has equal quantities of odd and even harmonics to be creatively removed. These harmonics give saw-tooth waves a strong, bright and almost rasping sound.
- Sine wave is the waveform with no harmonics at all which gives it a very pure sound much like the sound obtained from blowing over the top of a beer bottle.
- Square waves are, well, square waves. They have only odd harmonics close to the fundamental frequency. Square waves sound a little like a sine but with a much crisper, ringing quality. They are good for bass notes.
- Triangle waves have only higher odd harmonics and also sound a little like a sine wave but with a reedy quality.
- Pulse waveforms look a bit like a square wave but the peak to trough ratio is not equal. You could call them rectangle waves. They sound like a square wave but with a thinner less natural edge.
- White/Pink noise is available on some synthesizers and can be used to make percussion sounds or stupid wind noises.
It is after this process of designing timbre by harmonic interference that the sound will be routed through one, or number of filters. This is where the subtractive element of the synthesis begins.
Filters are inherent to the process of synthesis in general and are where subtractive synthesis derives its name. This is because filters, you may be scandalised to find out, filter frequencies out of a sound and in a number of ways can highlight the remaining frequencies further modifying the sound's character. Which frequencies are removed depends on the type of filter and how its settings are configured. The basic settings of a filter are:
- Cutoff. This is the point in the spectrum at which the filter begins removing frequencies
- Pole number or dB. This is the breadth of frequencies encompassed between the cutoff point and the point in the spectrum where frequencies are reduced to zero in volume. Standards are 6dB, 12dB or 24dB slope per octave. Higher pole filters have less breadth between these points than low or single pole filters.
- Resonance or 'Q'. This is the setting that controls, for resonating filters, that produce feedback at and around the cutoff frequency, the level of amplitude surrounding the cutoff frequency. This has the effect that frequencies will be at normal amplitude before rising, to a greater degree as the resonance is increased, just before the cutoff and then falling to zero.
The way in which these settings will effect a sound also depends on the type of filter being used. These are the most common:
- Low pass filter. These filters remove the frequencies above the cutoff frequency. This has the effect of making the sound seem less harsh and more rounded and bassy
- High pass filter. The opposite of a low pass this removes frequencies below the cutoff to create a bright and hollow sound
- Band pass filter. This filters frequencies at either side of the cut off, often creating a muffled sound but can create dynamic 'warping' noises when the cutoff is swept with a high Q
- Notch filter. The opposite of a band pass, this filter removes frequencies directly at the cutoff leaving those above and below. This filter can be used to remove unwanted resonance and noise from a signal and also to make awesome buzzing sounds
- Comb filter. A comb filter is a series of high resonance notch filters fixed a small distance apart in the spectrum and swept simultaneously. These filters can often be inverted so they resonate only and do not remove frequencies or remove all frequencies except those within the "teeth" of the comb. They also often have controllable feedback. Comb filters sound odd. They can be used to make freaky metallic gurgling noises and on models with a high degree of control-ability can be used to do a whole host of weird things such as resonating chords out of noise or creating robotic effects on vocal parts
Praise Bob, it's stopped
Bob Moog is considered by many to be the father of the modern synthesizer. This is due to the popular, user friendly and accessible Mini Moog and the terrifying and baroquely interconnect-able Moog Modular analogue synthesizers he produced during the 1960's. These synthesizers, as used by a distinguished list of artists (including Stevie Wonder) had a significant role in popularising the synthesizer and combined many trademark and pioneering features,(such as a keyboard) which have become indispensable in making modern synthesizers more like actual musical instruments and less like monsters out of Dr Who. Despite early analogue synthesizers being monophonic, tonally unstable and without a patch memory or a generic control system they remain popular for their "organic" or "fat" sound which some feel is lacking from the more reliable modern digital synthesizers. Bob Moog's modular synthesizer was the first modular to be available commercially. Modular synthesizers feature the same electronic units or modules as normal synthesizers with the distinction that in modular synthesizers it is decided by the user how these are connected to each other. In original analogue synthesizers connections would be made with wires and in the newer digital-modular and soft-synth models virtual wires or connection lines. The modular synthesizer has progressed to such an extent that huge virtual-modular programs exist, such as Native Instruments Reaktor, that allow people to build the synthesizer of their dreams (or nightmares,more frequently) using only several hundred modules, macros and frightening boxes containing "algorithms". Some of the simpler6 modules are often marked with acronyms. Knowledge of these arcane looking symbols and devices can be invaluable to someone wondering why their sound has gone "all bitty". A list of common module acronyms follows.
- VCO : This stands for Voltage Controlled Oscillator the voltage is the means used to control the pitch or note of the sound.
- VCF : Stands for Voltage Controlled Filter. This is the filter as described earlier
- VCA : This controls the volume of the sound and can be used as a general control in addition to the Envelope or for amplitude modulation and stands for Voltage Controlled Amplifier.
- LFO : Thought you were getting the hang of it? LFO stands for Low Frequency Oscillator. It is an, erm, oscillator that oscillates at a low frequency. These are used to create slow fluctuations in other parameters. When applied to pitch an LFO will create a Vibrato effect at the one frequency and amplitude or a siren at another frequency and amplitude
- SL : This term is sometimes found but more often the full term Slew Limiter is used. This creates smooth transitions between voltages and can create sweeps and slides of other parameters including pitch to create a portamento effect.
- CV : This stands for Control Voltage and was how old school synthesizers were controlled remotely. Wiretastic it was too.
Lucifer's ding dong!
Frequency Modulation synthesis, sometimes refered to as phase distortion synthesis, was invented by John Chowning in 1967 at Stanford University and is a form of digital synthesis. FM synthesis works by using one, or more, waveform as a carrier signal and one or more other frequency to modulate the frequency of this carrier. This modulation, depending on the frequency and amplitude of the modulator will produce different harmonic or inharmonic partials7 at either side of the carrier frequency altering the timbre of the sound. FM synthesis can be quite hard to use as like many other non-linear synthesis methods it is a relatively unintuitive method and only modulation signals that are harmonically related to the carrier wave will create harmonic tones, this potential inharmonicity often causes a built-in propensity for FM synthesizers to create weirdly dissonant digital sounds often in the form of frightening doorbell noises or tones similar to what a looped sample of old school modem would sound like on dial-up if you poured milk in it [citation needed]. Earlier user programmable FM synthesizers were often slightly pointless compared to their cheaper non-programmable cousins as the difficulties in creating pleasant sounds often made them reliant on expertly programmed presets. FM synthesis however was responsible, in part due to the preposterously successful Yamaha DX7, for the sound of a fair proportion of the electronic music of the 1980s.
Can I go home now please?
Wavetable synthesis is slightly confusing. This is because there are several different types of wavetable synthesis. The simplest form works by cycling between wave forms in either a preset or user designed sequence. For example: instead of a sine wave that repeats over and over a sequence may go sine_square_saw_pulse before repeating. This creates a different sound than any single wave from the sequence would on its own and in a way ever so slightly, but not really, similar to how a rapid sequence of still frames creates an illusion of movement. This type of wavetable synthesis often includes a cycle speed setting (which changes the wave at a distinct frequency from the frequency of the note to create extra harmonics) and a phase setting, which changes the start point each wave re-triggers at (also to create extra harmonics). The most widely heard instance of this form of synthesis is probably the ground breaking MOS-6581 SID (Sound Interface Device) chip. This is the chip that creates all the sound and music for the mind meltingly successful Commodore 64 home computer and uses wavetable synthesis to create a wide range of crunchy digital textures. The other less awesome, less C64 related form of wavetable synthesis isn't really a form of wavetable synthesis, merely mistitled sample-based synthesis, and, mysteriously, uses samples as a starting point. Sample based synthesis works by digitally chopping up recorded sounds into tiny chunks the order, number of repetitions, length (pitch) or phase positions of which can be changed to fulfil the role of synthesizer parameters. For example if someone were to say the word "quack" 8 and record it into a sample based synthesizer we would, for ease of explanation, have five components Q-U-A-C and K. If the length of each were shortened the word would be reproduced the same but at a higher pitch. If the number of cycles for the U and A were to be increased, for example, the sound would be reproduced as Q-U-U-A-A-A-A-A-C-K. This, however, is a stupid analogy. The technique of splitting sounds into tiny chunks and repeating them is the also the basis for granular synthesis. The distinction between sample based synthesis and granular synthesis is that in the former the waves are translated mathematically into their harmonic components as opposed to being played back directly as portions of a sample.
All work and no play makes Jack a dull boy
Great synths of all time
The Access Virus
The Sequential Circuits Prophet 5
The Roland TB303
The Fairlight CMI (computer musical instrument)
The Moog MiniMoog
The Korg MS20
The Arp Odessy
Casio CZ series
NED synclavier system
Yamaha DX7
The Thingythingy Placeholder
1Not to be confused with timber.2This restarts the waveform at a given frequency altering the harmonics.3 this will alter the start point of the waveform independently for each stereo channel which will cause physical wave cancellation when reproduced.4PWM for pulse waveforms. Essentially changes the mark to space ratio of the waveform.5Don't ask.6not to be confused with the Ableton Sampler "Simpler".7If you don't know what these are it's best to think of them as extra frequencies8Or "quank" for those of a Gallic disposition
