Making sounds with analogue electronics – Part 3: Envelopes
[Part 1 briefly reviews the differences between analogue and digital synthesis, and discusses "one of the major innovations in the development of the synthesizer" – voltage control. Part 2 begins a look at subtractive synthesis with a discussion of VCOs, waveforms, harmonic content, and filters.]
3.3.6 Envelopes
An envelope is the overall ‘shape’ of the volume of a sound, plotted against time (Figure 3.3.17). In an analogue synthesizer, the volume of the sound output at any time is controlled by a voltage-controlled amplifier (see VCA) and the voltage that is used is called an envelope. Envelopes are produced by ‘EGs’ and have many variants. EGs are categorized by the number of controls which they provide over the shape of the envelope. The simplest provide control only over the start and end of a sound, whilst the most complex may have a very large number of parameters.
FIGURE 3.3.17 The ‘envelope’ of a sound is the overall shape – the change in volume with time. The shape of an envelope often forms a distinctive part of a sound.
Envelopes are split into segments or parts (Figure 3.3.18). The time from silence to the initial loudest point is called the attack time, whilst the time for the envelope to decrease or decay to a steady value is called the decay time. For instruments that can produce a continuous sound, such as an organ, the decay time is defined as the time for the sound to decay to the steady-state ‘sustain’ level, whilst the time that it takes for the sound to decay to silence when it ends is called the release time.
FIGURE 3.3.18 Envelopes are divided into segments depending on their position. The start of the sound is called the ‘attack segment’. After the loudest part of the sound, the fall to a steady ‘sustain’ segment is called the ‘decay’ segment. When the sound ends, the fall from the sustain segment is called the ‘release’ segment.
Bowed stringed instruments can have long attack, decay and release times, whilst plucked stringed instruments have shorter attack times and no sustain time. Pianos and percussion instruments can have very fast attack times and complex decay/sustain segments. There is an almost standardized set of names for the segments of envelopes in analogue synthesizers, which contrasts with the more diverse naming schemes used in digital synthesizers.
Envelopes are usually referred to in terms of the CV that they produce, and it is normally assumed that they are started by a key being pressed on a keyboard. Envelopes can be considered to be sophisticated time-based function generators with manual key triggering.
The following are some of the common types of EGs.
Attack release
Attack release (AR) envelopes only provide control over the start and end of a sound (Figure 3.3.19). The two-segment envelope CV, which is produced, rises up to the maximum level and then falls back to the quiescent level, which is usually 0 volts. AR envelopes are often found on 1970s vintage string machines: simple polyphonic keyboards that used organ ‘master oscillator and divider’ technology with simple filtering and chorus effects processing to give an emulation of an orchestral string sound (see Section 3.4 for more information).
FIGURE 3.3.19 In an AR envelope the pressing down of a key (or a similar gating device on a synthesizer that does not use keys) starts the attack segment. When the peak level has been reached, then the envelope stays at this level until the key is released (of the gating signal is removed) and the envelope falls in the release segment. If the key is released whilst the envelope is in the attack segment, then the envelope normally moves to the release segment, and need not reach the peak level (see also Figure 3.3.27). Some synthesizers provide a control which forces the whole of the attack segment to be completed.
Attack decay
If the envelope moves into the decay segment as soon as the attack segment has reached its maximum level, then the decay time sets how long it takes for the envelope to drop to zero. This means that only percussive (non-sustaining) envelopes can be produced (unless the decay time is set to be very long, as in some attack decay release (ADR) envelopes). These two-segment attack decay (AD) envelopes (Figure 3.3.20) are often found connected to the frequency control input of VCOs, where the envelope then produces a rapid change in pitch at the start of the note, known as a ‘chirp’. This can be effective for vocal and brass sounds. Inverting the envelope can produce changes downwards in pitch instead of upwards.
FIGURE 3.3.20 An AD envelope is similar to an AR envelope, except that there is no sustain segment. When the peak level is reached, the envelope decays, even if the key is held down.
Attack decay release
The ADR envelope uses long decay times to simulate a high sustain level, in which case the resulting envelope is very much like an AR envelope, or else a percussive AD envelope by using shorter decay times (Figure 3.3.21).
FIGURE 3.3.21 The ADR envelope provides control over separate decay and release segments. This allows more complex envelope shapes to be produced than is possible with AR or AD EGs. If the key or gate is released during the attack segment, then the envelope moves to the release segment and ignores the decay segment.
Attack decay sustain
If a sustain level is added to an AD envelope, then the attack decay sustain (ADS) EG is the result (Figure 3.3.22). The attack segment reaches a maximum value and the decay time then sets how long it takes for the envelope to reach the sustain level. Some ADS EGs have switches that make the release time the same as the decay time or else have a very short release time. The type of envelope that is produced depends on the sustain level. If the sustain level is set to the maximum level (the same as the attack reaches), then two-segment ARtype envelopes are produced. If the sustain level is set to zero, then only twosegment AD envelopes are produced. With the sustain level set mid-way, then four-segment ADSR-type envelopes can be produced. If these have an initial attack and decay portion, then the sustain portion whilst the key is held down and then a release portion when the key is released.
FIGURE 3.3.22 An ADS envelope adds a sustain segment at the end of the decay segment. The ‘release’ time is normally set to the same as the decay time, although some synthesizers provide a switch which forces a fast release time regardless of the setting of the decay time. An ADS EG can be used to produce a wide variety of envelopes, including the ones which have many of the characteristics of ADSR (see later), AR and AD envelopes.
Attack decay sustain release
The most widely adopted EG is probably the ADSR (Figure 3.3.23). With just four controls, it is capable of producing a wide variety of envelope shapes; with only the attack decay 1 break decay 2 release (ADBDR) dual-decay variant offering superior flexibility at the cost of one extra control. The ADSR EG’s main weakness is that the sustain segment is static, it is a fixed level. For this reason, ADSR-type envelopes are not particularly well suited in producing percussive piano-type envelopes, where the ‘sustain’ portion of the sound gradually decays to zero. See ADBDR envelope later for a better alternative.
FIGURE 3.3.23 The ADSR envelope adds a separate control for the release time. This provides enough flexibility to produce a large number of envelopes with a small number of controls and the ADSR envelope is widely used in synthesizers.
Attack hold decay sustain release
Some envelopes force the envelope to stay at the maximum or peak level for a fixed time when the attack segment has finished and before the decay segment can start (Figure 3.3.24). These are called attack hold decay sustain release (AHDSR) envelopes. This is useful when a percussive envelope is set with very rapid attack and decay times, and the minimum length of the envelope needs to be controlled. For some sounds, an AD envelope with fast times (less than 10 ms) can be too short to be audible.
FIGURE 3.3.24 An AHDSR envelope adds a ‘hold’ segment at the end of the attack segment, rather like the sustain segment, but the length is set by a time rather than when the key or gate is released. As with other envelope shapes, if the key is released before the sustain segment, then the envelope moves to the release segment.
A variation on the hold segment being after the ‘attack’ segment of the envelope is the attack decay hold release (ADHR) envelope, where the ‘sustain’ segment is only held up to a specific time, after which it begins to decay. This is arguably better suited to percussive and piano sounds than the ADSR.
Attack decay 1 break decay 2 release
By splitting the decay segment into two portions, with a ‘break-point’ level controlling when one decay portion finishes and the other starts, a wide range of envelope shapes can be produced (Figure 3.3.25). By setting the second decay to a very long time, it can be used in much the same way as a sustain segment, although it has the advantage that it can still decay away slowly. This is arguably a better emulation of real-world envelopes for instruments such as pianos, where the sustain segment is actually a long decay time. In some implementations of ADBDR envelopes, this second decay is called the ‘slope’ segment to distinguish it from the decay segment.
FIGURE 3.3.25 The ADBDR envelope has two decay segments and the transition from one decay is set by a variable level control, rather like a sustain level control. By setting the decay time to a long value, they can be used as pseudo-sustain segments, and so an ADBDR envelope can produce similar envelopes to an ADSR type.
Advanced EGs
There are many sophisticated enhancements of the basic analogue ADSR EG (Figure 3.3.26). Most of these are ADSRs with the addition of initial time delay, break-points in the attack or decay segments and times for the peak and sustain levels. Although the extra controls provide more possibilities for envelope shapes, they also greatly increase the complexity of the user interface. Delayed envelopes (denoted by an initial ‘D’ in the abbreviation: DADSR for delayed ADSR) are used when the start of the envelope needs to be delayed in time without the need for using a long attack time, or where the attack needs to be rapid after the delay time.
FIGURE 3.3.26 Multi-segment envelopes can have several attack, decay and release segments, as well as hold and sustain segments. Break-points can also be used to split a segment into smaller segments.
Some of these EGs provide a break-point in the attack segment, so that two different attack times can be controlled. This is especially useful for long attack times, where the start of the audio signal is too quiet to be heard, and the initial portion of the attack segment is heard as a delay. By having a rapid rise to a level where the audio signal is audible, followed by a slower second attack portion, this unwanted apparent delay can be avoided. This extra break-point is also useful for simulating more complicated attack curves.
Break-points are not always explicitly named as such. The interaction between the gate signal and the envelope often has implied break-points at the transitions between attack, decay, sustain and release. These are frequently not documented in the manufacturer’s product information. The usual method of operation is shown in Figure 3.3.27.
FIGURE 3.3.27 The transition from the attack segment to the release segment when the key or gate is released can be thought of as adding in a break-point to the attack segment.
If the key is only held down for a short time, and the envelope is still in the attack segment when the key is released, then the envelope will go into the release segment. In this case the envelope may not reach the maximum level, although some EGs always rise to the maximum level. If there is a hold time associated with the maximum level, then this is usually not affected by the key being released. If the envelope has reached the decay segment, then when the key is released, the envelope will go into the release segment.
If the initial, final, peak and sustain levels are all controllable, then the envelope flexibility can become approximately equivalent to the multi-segment envelopes often found in digital synthesizers, although the terminology is normally very different. See Chapter 5 for more details on digital envelopes.
Some analogue synthesizers only have one EG, which is then used to control both the VCF and VCA. If two envelopes are available, then patching one to the filter and the other to the amplifier provides independent control over the volume and timbre. A third envelope could be used to control the pitch of the VCOs or perhaps the stereo position of the sound using two VCAs arranged as a pan control.
Linear or exponential?
Many real-world quantities change in a non-linear way. This can be due to the process involved or the way that the change is perceived. For example, the theoretical population growth curve of many animal species shows an exponential or power-law growth because the initial two animals produce two new individuals, who then eventually join the breeding population, and then these four individuals produce four new offsprings. The doubling of the population in each successive generation produces a rapidly increasing population curve. Conversely, because human ears perceive sound in a non-linear way, each doubling of the apparent volume level requires about 10 times the energy in the sound. Again, the relationship connecting the two variables is a non-linear one.
Many natural sound envelopes have non-linear curves. Changes are usually rapid at first and gradually slow down (Figure 3.3.28). This is particularly apparent with the attack segment of envelopes, where a linear rise in volume sounds too slow at first, whereas an exponential rise in volume sounds ‘correct’ – in fact, it sounds ‘linear’ to the human ear! Some EGs enable a switched selection between linear and exponential curves. EGs with breakpoints in the attack, decay and release segments can produce similar effects to exponential curves, albeit with a crude approximation.
FIGURE 3.3.28 An exponential envelope does not use linear slopes and often provides more realistic sounding envelopes.
Triggering
The initiation of an EG is often assumed to be caused by a key being pressed on a music keyboard. Although this is the way that many synthesizers are set up, it is not the only way that envelopes can be started – an LFO or a VCO could provide a trigger which will start the EG. In this case, the envelope is not tied to the keyboard and can be used when a complex repeated CV is required (Figure 3.3.29).
FIGURE 3.3.29 The retriggering of an EG can sometimes be used to add in a break-point and start a new attack, normally from the level which had been reached by the envelope. The overall length of the envelope is controlled by the key being pressed down, or a similar gate control in synthesizer which are not controlled by a keyboard. The retriggering of the envelope is controlled by a trigger signal which is generated by the start of each new note. This is normally found on monophonic synthesizers, where the gate is produced globally from any keys which are being held down, whilst the triggers are produced individually by each key.
When the keyboard is used to start an envelope, two separate signals are produced. The ‘gate’ signal indicates when the key is up or down, whilst the start of the key depression is shown by a ‘trigger’ pulse (Figure 3.3.30). The response of an EG to these two signals depends on how the EG is configured.
FIGURE 3.3.30 The gate and trigger routing from a keyboard to the EG is normally fixed, whilst the keyboard CV can be routed to a number of destinations.
‘Single trigger’ EGs start when they receive a gate and a trigger and progress through the envelope, entering the release segment when the gate signal ends to indicate that the key is no longer being held down. ‘Multitrigger’ EGs start when they receive a gate signal and a trigger pulse, but additional trigger pulses will restart part of the attack segment and the decay segment. These extra trigger pulses are normally produced by monophonic synthesizers (one note at once) only when a key is held down and another key is pressed.
‘LFO trigger’ or ‘external trigger’ EGs normally ignore the trigger pulse and treat the input signal as a gate. The width of the LFO waveform or the length of the external signal sets the length of the gate signal.
Whereas sources of audio signals or CVs can be routed to almost any destination in a synthesizer, the routing of trigger and gate signals is often much more restricted – usually they are hard-wired from the keyboard in performance instruments.
Voltage-controlled parameters
Some EGs provide voltage control of the segment times and levels. This enables the shape of the envelope to be changed with one or more CVs. One use of this facility is for ‘scaling’, where the length of all the times in the envelope are changed to imitate variations in envelope shape with pitch, in which case the CV would be derived from the keyboard pitch CV. This type of facility is much more commonly found in digital synthesizers.
Table 3.3.1 Summary of Envelope Segments 
Coming up in Part 4: Voltage-controlled amplifiers.
Printed with permission from Focal Press, a division of Elsevier. Copyright 2009. "Sound Synthesis and Sampling" by Martin Russ. For more information about this title and other similar books, please visit www.elsevierdirect.com.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
