When synthesising a sound of any type there are three main factors which must be correct if this final result is to be reasonably convincing and acceptable. The first is simply that the appropriate fundamental frequency must be present. As a few simple examples, this would entail generating low frequency tone of reasonable purity for drum sounds or a high frequency sinewave to give a sound similar to that produced by a triangle. The second factor is the harmonic content of the signal, that is the number of individual harmonics, or sinewaves, which with the fundamental make the waveform. Complex metallic sounds such as those produced by a bell or gong can be simulated using a ring modulator as explained in last month's 'Electro-Music Engineer' article. The third, and equally important factor, is the envelope shaping of the signal. This simply means controlling the amplitude of the signal, or volume, throughout its duration.

Envelope shapers

An envelope shaper is used to give the required changes in signal amplitude, and the most useful type for synthesized sounds is the ADSR (attack, decay, sustain, release) type. This gives an envelope shape of the type shown in Figure 1a. After the initial attack, or rise in amplitude, the signal decays to a sustain level, and then releases to an insignificant level.

Most ADSR envelope shapers have the durations of all four stages of the envelope independently adjustable from around 2ms to 10 seconds for A, D & R and 0 to 100% for sustain. This gives immense versatility, but at the cost of complexity, and even the most simple ADSR circuits are still quite complicated and consequently expensive to build. Fortunately it is possible to obtain good results using relatively simple envelope shaping circuits as it is hoped that this article (and part 2 next month) will show. Using more simple envelope shaping circuits the AD envelope shape obtained is usually something along the lines of Figure 1b, and this is quite satisfactory for the production of simple drum, cymbal and similar percussive sounds. This envelope has a preset fast attack of about 10mS and a variable decay. Very simple envelope shaping circuits using bipolar transistors and various types of field effect transistor (FET) tend not to give acceptable results, with the decay tending to be too abrupt. With careful adjustment of circuit values it is usually possible to obtain improved results, but adjustments are normally very critical with subsequent poor long term stability. Circuits based on integrated circuit VCAs tend to be more reliable and give better results, and only circuits of this type will be considered here.

Control voltage

Basically ar envelope shaper just consists of a circuit which generates a control voltage plus a Voltage Controlled Amplifier (VCA) which is fed with this control voltage. Assuming that the VCA is a type which provides high attenuation with a control voltage of zero, and decreasing attenuation as the control voltage is increased, the control voltage must initially rise rapidly to a fairly high value, then decay quite slowly and steadily. In other words the control voltage must follow much of the same path as the envelope shape of Figure 1b.

A simple control generator circuit which gives an output of this type is shown in Figure 2. This circuit can be triggered either by operating S1 or by applying a positive pulse of a few volts in amplitude to the input terminal. The duration of the trigger pulse is not critical, and anything over about one millisecond should be satisfactory. The circuit can also be triggered by tapping microphone Mic 1. The circuit operates in much the same manner whichever method of triggering is employed. IC1 is biased so that its output is normally at the negative supply potential, but operating S1 or providing an input trigger pulse results in a short positive pulse being fed to the non-inverting input of IC 1. This pulse is greatly amplified so that the output of IC1 goes fully positive, but for no more than a few milliseconds. Tapping the microphone results in a series of positive and negative pulses being fed to the non-inverting input of IC1, but the negative pulses have no effect and the positive ones cause the output of IC1 to go fully positive provided the microphone is tapped hard enough. Again the duration of this output signal will only be a few milliseconds at most since the microphone will provide only a very brief signal.

With the output of IC1 fully positive C4 rapidly charges by way of D1 to a potential almost equal to the positive supply rail voltage. When the output pulse ends C4 cannot discharge into the output of IC1 as D1 blocks any significant current flow in this direction. C4 therefore only discharges through R6 and VR1, and the rate of discharge depends on the effective value of VR1. At maximum value the decay time is about 4 seconds, but this reduces to less than a hundredth of this figure at minimum value and this control permits large changes in the nature of the sound obtained. The voltage across C4 falls quite rapidly at first since the high charge voltage produces a comparatively high discharge current, and as the charge voltage decreases so does discharge current and rate of discharge. Most natural percussive sounds decay in a manner which is analogous to this and good results are obtained using this method.

An important point to bear in mind is that the output of the circuit must feed into a load impedance of at least several meg ohms, and a buffer amplifier will normally be needed at the output. If necessary the control voltage can also be fed to a voltage controlled oscillator or filter to give falling pitch drum sounds etc.

The microphone can be a crystal type, a ceramic resonator, or a crystal earphone used in reverse as a crude microphone. If necessary the sensitivity of the circuit can be increased or decreased by raising or lowering the value of R5, and to an extent the circuit is touch sensitive.

VCA

Many Electro-Musicians probably have one or more MC3340P VCA integrated circuits in their possession, and the popularity of this device makes it an obvious choice as the basis of a matching VCA for the circuit of Figure 2. However, this device has a fairly low input impedance at the control input, and requires a control voltage which varies from about 3.5 volts (maximum gain) to 6 volts (minimum gain). This is obviously incompatible with the circuit of Figure 2, and to successfully interface the two it is necessary to have a buffer stage that inverts the control signal, provide level shifting, and also gives a certain amount of attenuation. This can actually be achieved using a simple circuit based on an operational amplifier, as shown in the circuit diagram of Figure 3.

In order to give R2 the correct setting first set the wiper of this component well towards (but not at) the negative supply end of its track, and then adjust the slider just far enough towards the opposite end of the track to give maximum attenuation.

The MC3340P gives around 13dB of voltage gain at maximum, and it can handle input levels of up to 500mV RMS without clipping.

Next month's article will describe alternative VCA's based on operational transconductance amplifiers, plus a simple VCF circuit.