This section covers the general principles of electronic sound generation and processing in more detail, including references to Summit’s facilities where relevant. It is recommended that this chapter is read carefully if analogue sound synthesis is an unfamiliar subject. Users familiar with this subject can skip this section and move on to the next.
To gain an understanding of how a synthesiser generates sound it is helpful to have an appreciation of the components that make up a sound, both musical and non-musical.
The only way a sound may be detected is by air vibrating the eardrum in a regular, periodic manner. The brain interprets these vibrations (very accurately) into one of an infinite number of different types of sound.
Remarkably, any sound may be described in terms of three properties, and all sounds always have them. They are:
What makes one sound different from another is the relative magnitudes of the three properties as initially present in the sound, and how the properties change over the duration of the sound.
With a musical synthesiser, we deliberately set out to have precise control over these three properties and, in particular, how they can be changed during the “lifetime” of the sound. The properties are often given different names, e.g., Volume may be referred to as Amplitude, Loudness or Level, Pitch as Frequency and sometimes Timbre as Tone.
As stated, sound is perceived by air vibrating the eardrum. The pitch of the sound is determined by how fast the vibrations are. For an adult human, the slowest vibration perceived as sound is about twenty times a second, which the brain interprets as a low bass sound; the fastest is many thousands of times a second, which the brain interprets as a high-pitched sound.
If the number of peaks in the two waveforms (vibrations) are counted, there are exactly twice as many peaks in Wave B as in Wave A. (Wave B is actually an octave higher in pitch than Wave A.) The number of vibrations in a given period determines the pitch of a sound. This is the reason pitch is sometimes referred to as frequency. It is the number of waveform peaks counted during a given period of time which defines the pitch, or frequency.
Musical sounds consist of several different, related pitches occurring simultaneously. The lowest is referred to as the ‘fundamental’ pitch and corresponds to the perceived note of the sound. Other pitches making up the sound which are related to the fundamental in simple mathematical ratios are called harmonics. The relative loudness of each harmonic as compared to the loudness of the fundamental determines the overall tone or ‘timbre’ of the sound.
Consider two instruments such as a harpsichord and a piano playing the same note on the keyboard and at equal volume. Despite having the same volume and pitch, the instruments still sound distinctly different. This is because the different note-making mechanisms of the two instruments generate different sets of harmonics; the harmonics present in a piano sound are different to those found in a harpsichord sound.
Volume, which is often referred to as the amplitude or loudness of the sound, is determined by how large the vibrations are. Very simply, listening to a piano from a metre away would sound louder than if it were fifty metres away.
Having shown just three elements may define any sound, these elements now have to be realised in a musical synthesiser. It is logical that different sections of the synthesiser ‘synthesize’ (or create) each of these different elements.
One section of the synthesiser, the Oscillators, generate raw waveform signals which define the pitch of the sound along with its raw harmonic content (tone). These signals are then mixed together in a section called the Mixer, and the resulting mixture is then fed into a section called the Filter. This makes further alterations to the tone of the sound, by removing (filtering) or enhancing certain harmonics. Lastly, the filtered signal is fed into the Amplifier, which determines the final volume of the sound.
Additional synthesiser sections - LFOs and Envelopes - provide further ways of altering the pitch, tone and volume of a sound by interacting with the Oscillators, Filter and Amplifier, providing changes in the character of the sound which can evolve over time. Because LFOs’ and Envelopes’ only purpose is to control (modulate) the other synthesiser sections, they are commonly known as ‘modulators’.
These various synthesiser sections will now be covered in more detail.
The Oscillator section is the heart of the synthesiser. It generates an electronic wave (which creates the vibrations when eventually fed to a loudspeaker). This waveform is produced at a controllable musical pitch, initially determined by the note played on the keyboard or contained in a received MIDI note message. The distinctive tone or timbre of the waveform is actually determined by the waveform’s shape.
Many years ago, pioneers of musical synthesis discovered just a few distinctive waveforms contained many of the most useful harmonics for making musical sounds. The names of these waves reflect their actual shape when viewed on an instrument called an oscilloscope, and they are: Sine waves, Square waves, Sawtooth waves, Triangle waves and Noise. Each of Summit’s Oscillator sections can generate all these waveforms, and can generate non-traditional synth waveforms as well. (Note that Noise is actually generated independently and mixed in with the other waveforms in the Mixer section.)
Each waveform (except Noise) has a specific set of musically-related harmonics which can be manipulated by further sections of the synthesiser.
The diagrams below show how these waveforms look on an oscilloscope, and illustrate the relative levels of their harmonics. Remember, it is the relative levels of the various harmonics present in a waveform which determine the tonal character of the final sound.
These possess just one harmonic. A sine waveform produces the “purest” sound because it only has this single pitch (frequency).
These contain only odd harmonics. The volume of each decreases as the square of its position in the harmonic series. For example, the 5th harmonic has a volume 1/25th of the volume of the fundamental.
These are rich in harmonics, and contain both even and odd harmonics of the fundamental frequency. The volume of each is inversely proportional to its position in the harmonic series.
Square/Pulse waves contain only odd harmonics, which are at the same volume as the odd harmonics in a sawtooth wave.
The square waveform spends an equal amount of time in its ‘high’ state as in its ‘low’ state. This ratio is known as the ‘duty cycle’. A square wave always has a duty cycle of 50% which means it is ‘high’ for half the cycle and ‘low’ for the other half. Summit lets you adjust the duty cycle of the basic square waveform (via the Shape controls) to produce a waveform which is more ‘rectangular’ in shape. These are often known as Pulse waveforms. As the waveform becomes more and more rectangular, more even harmonics are introduced and the waveform changes its character, becoming more ‘nasal’ sounding.
The width of the pulse waveform (the ‘Pulse Width’) can be altered dynamically by a modulator, which results in the harmonic content of the waveform constantly changing. This can give the waveform a ‘fat’ quality when the pulse width is altered at a moderate rate.
A pulse waveform sounds the same whether the duty cycle is – for example - 40% or 60%, since the waveform is just “inverted” and the harmonic content is exactly the same.
Noise is a random signal, and does not have a fundamental frequency (and therefore has no pitch property). Noise contains all frequencies, and all are at the same volume. Because it possesses no pitch, noise is often useful for creating sound effects and percussion type sounds.
A Ring Modulator is a sound generator that takes signals from two oscillators and effectively “multiplies” them together. Summit’s Ring Modulator uses Oscillator 1 and Oscillator 2 as inputs. The resulting output depends on the various frequencies and harmonic content present in each of the two oscillator signals, and will consist of a series of sum and difference frequencies as well as the frequencies present in the original signals.
Another method of combining the signals from two sources is Frequency Modulation, or FM. In this technique, the frequency of one oscillator – sometimes referred to as the “carrier” - is dynamically varied about its nominal “centre” value by an amount corresponding to the instantaneous amplitude of the signal from the second oscillator. Summit has a set of controls on the panel dedicated to adding FM effects.
The precise sonic result will depend on the wave shapes of each oscillator, their relative pitch, and the maximum amplitude of the modulating signal: on Summit this latter parameter may be controlled manually, and may be further varied by both LFO and Envelope.
The result of frequency modulation is the generation of a wide range of additional harmonics (in fact, theoretically infinite), both above and below the pitch of the oscillator being modulated. In FM language, these harmonics are often referred to as sidebands. The number of “significant” sidebands is proportional to the amplitude of the modulating signal and inversely proportional to the frequency difference between the carrier and the modulator. If the modulator is already rich in harmonics, e.g., something other than a simple sine wave, each harmonic creates its own set of sidebands, further enriching the spectral content of the result.
To extend the range of sounds you can produce, typical analogue synthesisers have more than one Oscillator (Summit has three for Part A and three for Part B). By using multiple Oscillators to create a sound, it is possible to achieve interesting harmonic mixes. It is also possible to slightly detune individual Oscillators against each other, which creates a warm, ‘fat’ sound.
Summit’s Mixer allows you to create a sound consisting of the waveforms of Oscillators 1, 2 and 3, a Noise source and the Ring Modulator output, all mixed together as needed.
Summit is a subtractive synthesiser. Subtractive implies part of the sound is subtracted somewhere in the synthesis process.
The Oscillators provide the raw waveforms with plenty of harmonic content and the Filter section subtracts some of the harmonics in a controlled manner.
There are three basic filter types, all of which are available in Summit: low-pass, band-pass and high-pass. The type of filter most commonly used on synthesisers is low-pass. In a low-pass filter, a “cut-off frequency” is chosen and any frequencies below this are passed, while frequencies above are filtered out, or removed.
The setting of the Filter Frequency parameter dictates the point above which frequencies are removed. This process of removing harmonics from the waveforms has the effect of changing the sound’s character or timbre. When the Frequency parameter is at maximum, the filter is completely “open” and no frequencies are removed from the raw Oscillator waveforms.
In practice, there is a gradual (rather than a sudden) reduction in the volume of the harmonics above the cut-off point of a low-pass filter. How rapidly these harmonics reduce in volume as frequency increases above the cut-off point is determined by the filter’s Slope parameter. The slope is measured in ‘volume units per octave’. Since volume is measured in decibels, this slope is usually quoted as so many decibels per octave (dB/oct). The higher the number, the greater the rejection of harmonics above the cut-off point, and the more pronounced the filtering effect. Each of Summit’s filter sections has a 12 dB/oct slope, but two of the same type can be cascaded (placed in series) to produce a slope of 24 dB/oct. Summit also allows two different types of filter to be cascaded, or even to be placed “in parallel”, so the mixer output is treated by both.
A further important parameter of the filter is Resonance. Frequencies at the cut-off point may be increased in volume by advancing the filter’s Resonance control. This is useful for emphasising certain harmonics of the sound.
As Resonance is increased, a whistling-like quality will be introduced to the sound passing through the filter. When set to very high levels, Resonance actually causes the filter to self-oscillate whenever a signal is being passed through it. The resulting whistling tone being produced is actually a pure sine wave, the pitch of which depends on the setting of the Frequency control (the filter’s cut-off point). This resonance-produced sine wave can actually be used for some sounds as an additional sound source if wished.
The diagram below shows the response of a typical low pass filter. Frequencies above the cut-off point are reduced in volume.
When resonance is added, the frequencies around the cut off point are boosted in volume.
In addition to the traditional low-pass filter type, there are also high-pass and band-pass types. On Summit, the filter type is selected with the Shape switch .
A high-pass filter is similar to a low-pass filter, but works in the “opposite sense”, so frequencies below the cut-off point which are removed. Frequencies above the cut-off point are passed. When the Filter Frequency parameter is set to minimum, the filter is completely open and no frequencies are removed from the raw Oscillator waveforms.
With a band-pass filter, just a narrow band of frequencies centred around the cut-off point is passed. Frequencies above and below the band are removed. It is not possible to fully open this type of filter and allow all frequencies to pass.
More complex relationships between volume and frequency can be obtained by using simple filters of the types described above in combination. Summit allows you to “cascade” two filters of different types, creating a “series” combination. Such a combination will generally result in more frequencies being removed than with a single filter section, as both filters are subtractive. However, interesting results can arise if the two filters have different cut-off frequencies.
For example, if a low-pass filter is followed by a high-pass filter, the low-pass filter
will pass only very high frequencies to the high-pass filter, which will remove some of them, leaving a narrow band of frequencies “between” the cut-off frequencies of both filters. The width of this band depends on the difference between, or “separation of” the two cut-off frequencies.
Combining the same filters in parallel produces quite a different result, as the responses of the two sections are effectively summed together. Low frequencies will be passed by the low-pass filter and high frequencies by the high-pass filter, resulting in a dip or a hump in the response in the area between the two cut-off frequencies.
In earlier paragraphs, the synthesis of the pitch and the timbre of a sound was described. The next part of the Synthesis Tutorial describes how the volume of the sound is controlled. The volume of a note created by a musical instrument often varies greatly over the duration of the note, according to the type of instrument.
For example, a note played on an organ quickly reaches full volume when a key is pressed. It stays at full volume until the key is released, then the volume level falls instantly to zero.
A piano note quickly attains full volume after a key is pressed, but gradually falls in volume to zero after several seconds, even if the key is held.
A string section emulation only attains full volume gradually when a key is pressed. It remains at full volume while the key is held down, but once the key is released, the volume falls to zero fairly slowly.
In an analogue synthesiser, changes to a sound’s character which occur over the duration of a note are controlled by sections called Envelope Generators. One of these (Amp Env) is always related to the Amplifier, which controls the note’s amplitude – i.e., the volume of the sound - when the note is played. In Summit, each envelope generator has five main parameters, which determine the shape of the envelope; these are referred to as the AHDSR parameters, or the envelope “phases”.
Adjusts the time it takes after a key is pressed for the volume to climb from zero to full volume. It can be used to create a sound with a slow fade-in.
This parameter is not found on many synthesisers, but is available on Summit. It determines for how long the note’s volume remains at its maximum level following the Attack Time, before commencing the volume drop set by the Decay Time.
Adjusts the time it takes for the volume to fall from its initial full volume to the level set by the Sustain control, while a key is held down.
This is unlike the other Envelope controls in that it sets a level rather than a period of time.
It sets the volume level that the envelope remains at while the key is held down, after the Decay Time has expired.
Adjusts the time it takes for the volume to fall from the Sustain level to zero once the key is released. It can be used to create sounds that have a “fade-out” quality.
You will notice the diagram also includes a further, initial phase, Delay. This is how long it takes for the Attack Time – and hence the entire AHDSR sequence - to commence after the key is struck. This is another envelope phase which is not generally found on other synthesisers, but is available in Summit. The addition of a Delay Time leads us to rename the envelope sequence DAHDSR for completeness (though many users will continue to refer to it by the more traditional term ADSR).
Most modern synthesisers can generate multiple envelopes. Summit has three Envelope Generators: Amp Env has a dedicated set of hardware ADSR slider controls (Delay and Hold are controlled separately via the menu), and is always applied to the amplifier to shape the volume of each note played, as detailed above. The two Modulation Envelopes (Mod Env 1 and Mod Env 2) share an identical set of controls, with an assignment switch selecting the envelope being controlled. Modulation envelopes can be used to dynamically alter other sections of the synthesiser during the lifetime of each note. Summit’s Mod Env Generators can be used to modify the filter cut-off frequency, or the pulse width of the Oscillators’ Square Wave outputs, for example.
Like the Envelope Generators, the LFO (Low Frequency Oscillator) section of a synthesiser is a Modulator. Instead of being a part of the sound synthesis itself, it is used to change (or modulate) other sections of the synthesiser. For example, the LFOs can be used to alter Oscillator pitch, or Filter cutoff frequency, as well as many other parameters.
Most musical instruments produce sounds that vary over time both in volume and in pitch and timbre. Sometimes these variations can be quite subtle, but still contribute greatly towards characterising the final sound.
Whereas an Envelope is used to control a one-off modulation over the lifetime of a single note, LFOs modulate by using a repeating cyclic waveform or pattern. As discussed earlier, Oscillators produce a constant waveform, which can take the shape of a repeating sine wave, triangle wave etc. LFOs produce waveforms in a similar way, but normally at a frequency which is too low to produce a sound the human ear could perceive directly. As with an Envelope, the waveforms generated by the LFOs may be fed to other parts of the synthesiser to create the desired changes over time – or ‘movements’ - to the sound.
Imagine this low frequency wave being applied to an Oscillator’s pitch. The result is the pitch of the Oscillator slowly rises and falls above and below its original pitch. This would simulate, for example, a violinist moving a finger up and down the string of the instrument whilst it is being bowed. This subtle up and down movement of pitch is referred to as the ‘Vibrato’ effect.
A waveshape often used for an LFO is a Triangle wave.
Alternatively, if the same LFO signal were to modulate the Filter cut-off frequency instead of the Oscillator pitch, a familiar wobbling effect known as ‘wah-wah’ would be the result.
A synthesiser can be broken down into five main sound generating or sound modifying (modulating) blocks:
-
Oscillators that generate waveforms at a various pitches.
-
A Mixer that mixes the outputs from the Oscillators together (and add Noise and other signals).
-
Filters that remove certain harmonics, changing the character or timbre of the sound.
-
An Amplifier controlled by an Envelope generator, which alters the volume of a sound over time when a note is played.
-
LFOs and Envelopes that can be used to modulate any of the above.
Much of the enjoyment to be had with a synthesiser is with experimenting with the factory preset sounds (Patches) and creating new ones.
There is no substitute for ‘hands on‘ experience. Experiments with adjusting Summit’s various controls will eventually lead to a fuller understanding of how the various synth sections alter and help shape new sounds.
Armed with the knowledge in this chapter, and an understanding of what is actually happening in the synth when tweaks to the knobs and switches are made, the process of creating new and exciting sounds will become easy.