Having looked at the various types of effects available, Paul White explains the importance of the order in which these effects are applied.
Over the past two months, I've covered the effects blocks most likely to be found in a typical multi‑effects unit, and looked at the difference between the series and parallel connection of effects blocks, so now it's time to see how the various blocks can be used together to create different composite effects. One of the most important factors is the order in which blocks are connected, and this is particularly true in patches where reverb or delay is being used. For example, consider a dramatic effect such as a flanger, connected in series with a reverb block. Do you put the flanger before the reverb or after it? In fact you can connect them either way round, but the result will be very different.
Reverb creates thousands of delays which effectively scatter a signal in time. If you put the flanger before the reverb, the reverb reflections will effectively break up the cyclic nature of the effect, resulting in a more diffuse sound. Instead of the deep modulation effect associated with flanging, you'll get something more subtle, sometimes described as a reverb with shimmer. On the other hand, put the flanger after the reverb and the whole reverb tail will be flanged to produce the familiar 'whooshy' effect. Chorus is often placed before reverb to add a little movement and shimmer, and that can work well on more ethereal sounds, or pianos.
Substituting a delay block for the reverb produces similar, but perhaps less marked, differences. If the delay has feedback to create multiple repeats, and you put the flanger first, then each repeat will be, in effect, an image of the same section of the flanger sweep as the original sound, but delayed in time. As the repeats will tend to overlap when sustained sounds are being treated, the effect will be to overlay one flanger sweep with several more sweeps (decaying in level), but starting at different times depending on the delay time setting. This results in a weakening of the flanger sweep effect, but the complexity of the sound is increased, producing more movement and interest. However, some impression of flange sweep will probably remain, simply because multiple delays aren't as complex as reverb.
Putting the flanger last will simply impart the familiar sweeping sound onto everything that's fed into it. If the flanged effect is mixed in with the original dry signal, the result is a clean original sound followed by the flanged echo, which can still be more subtle than effects settings that treat the original sound as well. In terms of applications, chorus, flange or detuning (via a pitch‑shifter) can be placed before a delay or echo block to create an impression of space and distance.
So far we've been talking about series connection, but many effects units also include the facility for putting blocks in parallel. Sticking with delay and flanging for the moment, if we were to put these two blocks in parallel, the output would be a mix of two distinctly separate effects — a clean delay and a flanged original sound.
Let's stick with flanger before delay a while longer and see what else you might do with it. You could add a little reverb at the end of the line, because then each discrete delay (which is also flanged) would be diffused into a cloud of reverb rather than being a hard repeat. The effect is to further diffuse the flange effects and to increase the sense of space. Stereo reverb will also add stereo width to the sound, even if mono flanger and delay blocks come before it, though a stereo multitapped delay probably gives the best effect.
Most multi‑effects units also include EQ of some sort, but while the applications are obvious when you're working with guitar sounds, for example, it's not quite so clear how to use them in conjunction with other blocks to create specific effects. For example, say that you want to turn a digital delay into an approximation of a tape echo sound. You could put a very shallow chorus in front of a multi‑tapped delay, to simulate the wow and flutter that comes from a capstan that's not been cleaned for years and a pinch roller that's flattened out of shape, but you also need to recreate the very limited bandwidth of the old tape systems. Putting in a high cut filter to roll off the top at 3 or 4kHz provides a rough approximation, but to get it more accurate, you have to think in terms of what the original machine did. Remember that tape delays create multiple echoes in the same way as digital delays — they feed some of the output back to the input. In the case of the tape echo unit, though, the signal is being fed back into a system with a re‑cued bandwidth, so every time a delay circulates, it becomes duller. To emulate this with a digital system, you have to take the delay block's feedback path from after the filter — this way the delays pass through the filter every time they re‑circulate, progressively removing more top end. In fact, being really pedantic, you'd need to return the feedback path to the chorus input to add more 'capstan error' each time the sound goes around the loop — this gives a far better representation of the real thing. In some effects units, the delay block itself may include the option of filtering the feedback path. If not, you'll have to look through the manual to see if your particular unit will let you place an EQ block after the delay block, then take the feedback from the delay output. If you can't, you'll just have to settle for second best and filter the whole delay output, which means that all the delays will be equally equalised rather than becoming progressively less toppy.
Distortion is important to guitar players, but it can also be used more subtly, to add character to synth or sampled organ patches, or to make digital synths sound more analogue. As a general rule, distortion should come right at the front of the signal chain so that any following effects blocks can work on the new harmonics that the distortion introduces. For example, a flanger sounds far more dramatic on a harmonically rich sound.
Because distortion involves high gain, noise can be a problem, so noise gates are often included as part of the effects repertoire. These are usually best placed directly after the distortion block or, if there's an EQ block following the distortion, they could come after that instead. As a rule, gates come before compressors — because the output from a compressor has a smaller dynamic range than the input, it's harder to get the gate threshold right if you gate after compression. It also makes sense to put the gate before any effect that involves delay or reverb, because putting the gate at the end of the line may well chop off the ends of slowly decaying sounds. If you gate before delay or reverb, then even if a sound is cut off slightly early, the sustain of the reverb or delay (which now comes after the gate) will help disguise the fact. The simplest way to remember the basic connection rules is: distortion or other dynamic processors at the front of the chain and delay or reverb at the end.
Because these series connection rules apply in the majority of cases, some multi‑effects designs simplify the whole system by presenting the user with a predetermined chain of blocks which may be switched on or off. Additionally, there are often several alternative effects of a similar type that can be assigned to each block — for example, one block may provide modulation effects such as flanging, chorus, phasing or vibrato.
Most multi‑effects units contain modulation effects of one kind or another, the more obvious ones being chorus and flanging, but recently their designers have started to get more adventurous. For example, you may have a synth‑style resonant filter that can be modulated via an LFO, or even by the envelope of the sound being processed. This latter option needs an effects unit that includes an envelope follower. You can create synth‑like filter sweeps with these that follow the input signal level, say, or you can take the input envelope, invert it, and use that signal to control a gain block placed in front of your reverb block. This will pull the reverb down in level when the input is loud, but during pauses or quiet sections the reverb will increase. It's known as reverb ducking, and is useful for keeping busy mixes clean. If you're lucky, the whole ducking option will be built into the reverb block; if you have to build it yourself, though, at least you'll know what blocks you need and how to connect them.
Another useful facility is the sample and hold LFO, a modulation source derived from the old analogue synths. Like any other LFO, the source has a variable frequency, but instead of a repeating waveform, you get a series of random voltage steps, which produce interesting sounds if they're used to control filter frequencies, and instead of a continuous filter sweep, you get regular steps but at random frequencies — at high filter resonances, this can sound very effective. Other less obvious tricks include using the sample and hold waveform to trigger an autopan (sound jumps to random pan positions), or you could use the input envelope follower to control the pan, so that as the signals get louder, they move further across the soundstage.
The real trick to combining effects is to think through what you want to achieve, and then draw out a block diagram on paper that will let you achieve it. Then you'll have to work within the limitations of your effects processor to get as near to it as you can. Most of the time you'll have a fair idea of what the result is going to be, but when you start creating complicated feedback loops, the outcome is less predictable. For example, most pitch‑shifters include a delay and feedback facility, so if you set up a pitch shift, add a 200ms delay, then feed some of the output back to the input, each repeat will be shifted further than the preceding one by the same amount, until it finally spirals into inaudibility. Imagine, too, the further complexity you can introduce by putting one effect in the feedback loop of another, or by cross‑feeding signal feedback paths between the two arms of a parallel chain. Very soon you enter the world of chaos, where even the slightest parameter change brings about a huge change in the overall effect. To be fair, very few people experiment to this extent, but if you explore some of the more off‑the‑wall presets that inhabit your factory patches section, you can learn a lot about the less obvious ways to combine effects. Even if you don't feel up to creating your own from scratch, you should be able to load up a factory sound you like, then identify the most important parameters you need to tweak to customise it to your own needs.
Next month I'll be looking at ways to use MIDI control to make effects even more interesting.