Nick Magnus puts on his Indiana Jones hat and treks ever deeper into the heart of your digital synth. This month's Holy Grail is analogue brass.
Welcome to the third instalment of the Jurassic Park of synthesis — a computer‑generated voyage into sonic archaeology, if you will. Recreating analogue‑type brass sounds on one of today's digital 'Sample + Synthesis' instruments is the subject matter for this month.
Last issue, I discussed how the basic waveform of string sounds is arguably the most important aspect of synthesising them. Brass sounds, conversely, rely more on amplitude, filter and pitch envelopes to provide their all‑important aural signature. As before, I'll assume that our digital synth can use two or more oscillators, or tones, to make up a sound. A very useful first step is to program from scratch a 'vanilla' brass sound using one tone; this can be a basis from which to work. To put our modern synth firmly in the early '80s, start by removing all internal effects, stereo panning and velocity sensitivity. These could, conceivably, be reinstated later depending on the synth you're emulating (for example, a Yamaha CS80), but for now an entirely neutral starting point is all that is necessary; note that you should begin with the filter and amplifier envelopes at a standard gate (or organ type) setting.
Since a sawtooth waveform contains most of the characteristic harmonics of brass instruments, it is most often employed to create these sounds. And, being rich in upper harmonics, this waveform gives plenty of information for the filter to grab hold of for maximum timbral variation. Other waves, too, can be blended at various levels to fatten up the proceedings or subtly colour the tone. So, having made our one‑tone vanilla patch utilising a sawtooth wave, we are ready to proceed.
Brass instruments are very much identified by their timbral movement; typically, the initial portion of the sound is brightest, followed by a relatively quick drop to a mellower tone, which usually stabilises until the note is stopped. Of course a real trumpet, for example, takes a finite time to reach its brightest tone, due to the time it takes to build up air pressure when the instrument has been blown into. Thus we need to create an envelope to control a low‑pass filter over an appropriate time scale. Figure 1 is an example of what such an envelope might look like. In order to achieve this, the cutoff frequency of the filter needs to be set to the same as, or lower than, the final brightness of tone desired. The lower the cutoff frequency, the greater the timbral variation is a good rule of thumb. Figure 2 shows some suggested filter and envelope values to start you off. This envelope is not active, however, until we raise the amount of filter envelope modulation. This parameter is often called 'EG Gain' or 'Env Depth', or simply 'Modulation Amount'. As you set the value of this higher, so the effect the envelope has upon the filter becomes more apparent.
Having made a basic filter envelope, we turn our attention to the amplifier envelope. In a typical case, this could remain by and large as an ordinary gate envelope, but with a little softening of the attack portion. This can be set to the same value as the filter attack time, but is often just a 'gnat's' faster. This is so we can hear the entire filter movement at the front of the sound, but it prevents an unnaturally quick start and avoids the possibility of a key‑on click. A little time set on the release envelope of both filter and amplifier will also avoid an abrupt sounding key‑off. Figure 3 shows the relationship between the two envelope shapes.
Here's a programming tip: most digital synths have a 'global' editing capability — i.e. all selected tones can be edited to the same values, simultaneously. Even if only one tone is to be used, it saves a lot of time if all the tones are edited together. Most analogue synths put all their oscillators through a common set of envelopes; there are exceptions, but for now we'll stick with the classic Prophet/ Jupiter/ Oberheim OB series architecture.
We're now armed with a basic analogue brass simulation: here are some suggestions for ways to manipulate the sound, creating many variations of the basic.
- Try setting one tone an octave above the other. This will give more of a 'section' or 'big band' quality to the sound.
- The kind of fat analogue brass sound most sought after will, of course, use two oscillators. Make sure the second tone is identical to the first, or use the tone copy utility on your synth (if present) to copy the first tone to the second. If more than one sawtooth wave is offered in your synth, select the alternative one for your second tone and then detune the tones against each other to taste. Using two different saw waves can eliminate the phasing often encountered when detuning identical sampled waves.
- Needless to say, you don't have to stop at two tones; the MemoryMoog, for example, had three oscillators and a consequently huge sound. Try detuning two of the tones +6 and ‑6 cents with respect to the third.
- Make use of the envelopes' Time/Key follow parameter; if it's set so that lower key positions slow down the envelope movement and higher positions speed it up, it will more accurately represent how the real instruments respond.
- Adding some resonance to the filter will accentuate the 'synthy' nature of the sound. Extreme amounts of resonance will often require the filter envelope amount to be reduced to make things more manageable.
- French horn‑like sounds can be created easily by slowing down the attack and decay speeds considerably. To capture the mellowness of the sound, the filter envelope amount will need to be reduced accordingly, and a longer release time on both filter and amplifier is also very effective. You may also want to raise the filter sustain level to achieve a more consistent tone.
- Assigning a sine wave LFO to modulate the filter cutoff frequency very fast can produce a classic growl effect. Careful setting of the depth control is required here.
- Use both envelopes for effects such as sforzando and swells. Judicious setting of envelope times can put swells in context with the tempo of the relevant music. Figure 4 shows some examples.
- Especially important are the envelope times, especially the attack and decay times, the sustain level, and their relationship with the filter envelope depth. All these controls are highly interactive, so experiment, investigate, twiddle...
- For a more expressive effect, use the Pitch Envelope, if your synth has one. When real brass is blown hard, a certain amount of detuning occurs at the start of the note. Try making one of the tones slightly sharp or flat at key‑on, with a rapid return to normal tuning. If this pitch envelope is too slow, you will end up with a drunken brass section on your hands (then again, some people might say this is highly realistic). This detuning is very effective when the tones are an octave apart. Set the upper tone to start sharp, and the lower tone to start flat, and remember to make them return to pitch as quickly as is tasteful. See Figure 5 for an illustration of this.
- As mentioned last month, having an analogue synth to hand to use as a sound model is a real boon. A/B comparisons between the two for envelope speeds, LFO rates and so on, results in some wonderfully accurate simulations.
Next month, we'll ride the Turbo Elevator way on down to Engineering and take a look at recreating some favourite analogue‑style bass sounds.
A set of 64 classic retro sounds for the Roland JV‑series synths, including those sounds mentioned in these articles, is available free of charge from Roland dealers, or contact Roland UK on 0252 816181.
Analogue Feel
Some synths offer an 'Analogue Feel' parameter; this applies a very slow, shallow, random modulation to the tones, in an attempt to replicate the instability of Voltage Controlled Oscillators. Subtlety is the watchword once again, or else this parameter automatically renames itself 'Special Brew'. If this feature is not found on your synth, try the following:
1. Assign an LFO to all tones, selecting a random waveform. So much the better if there is a choice of stepped or smooth waves; the smooth random wave works best
2. Turn up the modulation depth until it is plainly audible then adjust the speed until a very slow, lazy pitch undulation is obtained.
3. Turn the depth right down again until this undulation is barely detectable.
4. Finally, if available, turn the LFO key sync mode to Off. This ensures that the tones modulate independently, providing the subtlest of chorus effects. Keep the depth low; as with many things in music, less is more.