Effects: All You Need To Know... And A Little Bit More

Whether you're a beginner or a seasoned pro, there's always something to learn about adding colour to your mix.

The world of audio effects is one that can be confusing even for experienced engineers. Especially in modern computer-based recording systems, there's a bewildering array of options, and to add to the confusion, some effects are widely referred to by more than one name. In this article, I'll take you through the most common effects, explaining how they work and where you might want to use them in your music. Meanwhile, SOS 's team of writers has contributed a wealth of expert tips and tricks, which you can find in the boxes scattered through this article.

Effect Or Processor?

Let's begin by getting clear on what we mean by 'effect': an effect is a device that treats the audio in some way, then adds it back to a dry or untreated version of the sound. Echo and reverb are obvious cases, and you can use pitch-shift and pitch modulation in a similar way. 'Processors', by contrast, generally are those devices that change the entire signal and don't add in any of the dry signal. Things like compressors and equalisers fall into this category: as you'll see from the tips and tricks, processors can often be used as effects in their own right, or as part of an effect chain, but until you know exactly what you're doing and what the consequences are likely to be, it is a good idea to stick to these guideline definitions, as they dictate how you can connect the effects into your system.

If a device has a Mix control on it that goes from 100 percent wet (effect only) to 100 percent dry (clean only), then you can be pretty sure it is an effect. If it doesn't have a Mix control, and doesn't rely on electronic delays to create its results, then it is probably safe to assume it is a processor.

Effects can be connected via insert points, or the effect send and return loop that is included in most consoles and DAWs (Digital Audio Workstations). When effects are used in the send/return loop, their Mix control should be set to 100 percent wet, so you add back only effected sound to the dry sound, which comes directly through the mixer channel.

Processors, on the other hand, comprise an entirely different water heating appliance filled with piscean vertebrates, as they tend not to need any of the dry sound, other than in a few specialist applications. As a rule, processors such as EQ and compression are connected only via track, bus or master insert points — at least until you have the necessary experience to understand why you might want to break the rules once in a while. Having got that off my chest, let's look at some specific effects (we'll look more closely at processors another time).

Echo & Delay

Echo and delay are created by copying the original signal in some way, then replaying it a short time later. There's no exact natural counterpart, though the strong reflections sometimes heard in valleys or tunnels appear as reasonably distinct echoes. Early echo units were based on tape loops, before analogue charge-coupled devices eliminated the need for moving parts. Today, most delay units are digital, but they often include controls to help them emulate the characteristics of the early tape units, including distortion and low-pass filtering in the delay path and pitch modulation to emulate the wow and flutter of a well-used tape transport.

While pure digital delay produces perfect echoes, an analogue emulation can be more musically useful, as each successive echo becomes less distinct, creating a sense of distance and perspective. Hi-fi echoes tend to confuse the original sound, while the human hearing system seems better able to separate lo-fi echoes from the original clean sound.

The feedback control regulates the number of echoes by feeding some of the output back to the input. If you apply too much feedback the delay unit will self-oscillate — an effect often used in dub music. Delay normally relates to a setting with no feedback whereas echo uses feedback to produce a series of diminishing repeats.

You don't have to use long, distinct delays: short delays up to 120ms can be used to create vocal doubling effects, normally set with little or no feedback. Nor do you have to dedicate a delay to a single sound: you can configure it via an aux send so that several tracks can be treated with different amounts of the same delay or echo treatment, which not only saves on processing power (or buying separate units!), but can help to make elements of your mix work better together.

You can often use a tap-tempo or tempo sync facility to get your echoes exactly in time with the song if that's the effect you need, but many echo/delay plug-ins can be locked to your sequencer's master tempo, enabling you to create precise, rhythmic delay effects.

Reverb

I think it's fair to say that we all have a pretty good idea of what reverb is, though there are several ways of emulating it in the studio. Early reverb chambers, plates and springs have now given way to digital solutions, which fall into two main camps: synthetic and convolution. Synthetic reverbs take an algorithmic approach, setting up multiple delays, filters and feedback paths to create a dense reverberation effect similar to what you might hear in a large room. Though these often sound a bit 'larger than life', they've been used on so many hit records that we now tend to accept their sound as being the 'correct' one for pop music production. Most can approximate the sound of rooms, halls, plates and chambers, but in comparison with a real reverberant environment, the early reflections often seem to be too pronounced. The advantage of a synthetic reverb is that the designer can give the user plenty of controls for altering the apparent room size, brightness, decay time and so on.

In recent years, convolution reverbs have become both affordable and commonplace. These differ from synthetic reverbs insomuch as they work from impulse responses (or IRs), recorded in real spaces to faithfully recreate the ambience at the microphone's position when the IR was made. Sometimes these are referred to as sampling reverbs but there's no sampling involved as such, even though the process seems akin to sampling the sonic signature of a room, hall or other space.

Because IRs can be recorded in virtually any space, convolution reverbs generally come with a library of IRs ranging from small live rooms to famous venues, top studio rooms, forests, canyons, railway stations and just about anything else you can think of. They sound very convincing, and there's plenty of variety to be had, but once the IR is loaded, there's only a limited amount of editing you can do without spoiling the natural sound. Usually you can apply EQ and also change the envelope of the reverb decay to make it shorter, and adding pre-delay is not a problem, but after that you pretty much have to take what you get. Some companies, such as Waves, have managed to create additional controls but, as a rule, the further you move from the original IR, the less natural the end result.

Ironically, the sound of certain synthetic reverbs is now such an established part of music history that most convolution reverbs come with some IRs taken from existing hardware reverb units or from old mechanical reverb plates. Also, if you have a convolution reverb, it is worth checking the manufacturer's site, as additional IRs are frequently available for download.

All serious reverb units have a stereo output to emulate the way sound behaves in a real space and, in the case of convolution models, the IRs are often recorded in stereo, using two microphones. Some surround reverbs are also available.

Reverb creates a sense of space, but it also increases the perception of distance. If you need something to appear at the front of a mix, a short, bright reverb may be more appropriate than a long, warm reverb, which will have the effect of pushing the sound into the background. If you need to make the reverb sound 'bigger', a pre-delay (a gap between the dry and wet signals) of up to 120ms can help to do this without pushing the sound too far back, or obscuring it.

Though reverb increases the sense of stereo width, it dilutes the sense of stereo position. If you want to pinpoint the placement of something in a mix, you should consider using a mono rather than a stereo reverb, and panning this to the same place as the dry sound.

Most synthetic reverbs allow you to balance the level of the early reflections and the later, more dense reverb tail. If you want to keep the sense of space but without the reverb tail taking up too much space in your mix, you can increase the early reflection level and reduce the tail level.

As a rule, you don't add much, if any, reverb to low-frequency sounds, such as bass guitar or kick drums. Where you need to add reverb to these sources, short ambient space emulations usually work better than big washy reverbs, which tend to make things sound muddy. Taking this a step further, you can also make a mix sound less congested by EQ'ing some low end out of your reverbs.

Modulated Delays

Effects such as chorus, phasing, flanging and pitch vibrato are created using pitch modulation and, except in the case of vibrato, the modulated sound is added back to the original to create the effect. The pitch modulation is generated by delaying the signal by just a few milliseconds, and then modulating the delay time, using a low frequency oscillator (or LFO). For vibrato, this is all that needs to be done and, because the delay time is actually very short, the effect is perceived as happening in real time. The other effects, however, generally rely on an equal balance of the dry and modulated signals to achieve the strongest effect, so it is easier when working with plug-ins to adjust the wet/dry balance using the plug-in controls, rather than adding the wet only signal via a send/return loop. As a rule, these effects aren't very processor-intensive so, if you're working with plug-ins, you can probably afford to insert as many as you need into track or bus insert points as required. Stereo versions of these plug-ins may generate different modulated delays for the left and right channels to create a more dramatic spatial effect.

Chorus and flanging are created in fairly similar ways, the main difference being that chorus doesn't use feedback from the input to the output and generally employs slightly longer delay times. Phasing is similar to both chorus and flanging, but uses much shorter delay times. Feedback may be added to strengthen the swept filter effect it creates. Phasing is far more subtle than flanging and is often used on guitar parts. With chorus, phasing and flanging, the delay time, modulation speed and modulation depth affect the character of the effect very significantly. A generic modulated delay plug-in allows you to create all these effects by simply altering the delay time, feedback, modulation rate and modulation depth parameters. Most of the time, low modulation depths tend to work well for faster LFO speeds (often also referred to as the rate), while deeper modulation works better at slower modulation rates.

Chorus is useful for 'softening' rhythm guitar or synth pad sounds, but it does tend to push sounds further back into the mix, so it should be used with care. Adding more brightness to the sound can help compensate for this effect. Chorus also works well on fretless bass, but tends to sound quite unnatural on vocals. Phasing can be used in a similar way to chorus but, whereas chorus creates the impression of two slightly detuned instruments playing the same part, phasing sounds more like a single sound source being filtered, where the frequencies being 'notched out' vary as the LFO sweeps through its cycle.

Flanging is the strongest of the standard modulation effects. The feedback control increases the depth of the 'comb filtering' produced when a delayed signal is added back to itself. Because it is such a distinctive effect, it is best used sparingly, though it can also be used to process a reverb send to add a more subtle complexity to the reverbed sound.

Though modulated delays are essentially effects, the need to balance the dry and delayed sounds as a means of regulating the effect strength means that using these devices via insert points makes them much more controllable than trying to use them in an effects send/return loop. If you do use them as a send effect, you can achieve this balance by automating the send level.

Pitch-shift

Pitch-shifters work by slicing the incoming audio into extremely short sections (typically a few tens of milliseconds long) and then lengthening each section where the pitch is to be decreased, or shortening each section where the pitch is to be increased. Though cross-fading algorithms and other techniques are used to hide the splice points, most pitch-shifters tend to sound grainy or warbly when used to create large amounts of shift (a couple of semitones or more), though they can sound very natural when used to create subtle detuning effects, using shifts of a few cents. A refinement of the system, designed for use with monophonic sources, attempts to synchronise the splicing process with whole numbers of cycles of the input signal, which makes the whole thing sound a lot smoother but, as soon as you present these devices with chords or other complex sounds, the splices again become audible.

Though some sophisticated processors combine pitch detection with pitch-shifting, to generate musically correct harmonies in user-defined keys, simple pitch-shifters always change the pitch by the same number of cents or semitones. In musical terms, that means that only the octaves, parallel fourths and fifths are very useful. Other intervals tend to sound discordant, as they don't follow the intervals dictated by typical musical scales.

When using subtle detuning to thicken a sound, I suggest trying values of between five and 10 cents and, where possible, adding both positive and negative shifts, to keep the pitch centre correct. Then combine with the dry sound and adjust the level to control the subjective depth of the effect. This is very effective for fattening up guitar solos or backing vocals.

By putting a pitch-shifter before a delay, then feeding some of the output of the delay back to the input of the pitch shifter, you can create delays that keep climbing or falling in pitch as they recirculate. Though not always very useful in a musical context, this effect is often used in TV and film dream sequences.

Because large pitch-shifts can sound grainy, it is common to combine the effect with the dry signal, rather than using only the 100 percent effected signal, though ultimately this is an artistic rather than technical decision.

Though pitch-shifting is an effect, it is easier to control when used via an insert point. However, if you need to use the effect on several tracks in varying amounts, you can use it via a send/return loop, providing the shifter is set to 100 percent wet. That way, you can adjust the effects depth for individual mix channels by using the send control feeding the pitch-shifter.

Tuning Correction

By contrast, tuning (or pitch) correction processors and plug-ins are normally considered processors rather than effects, but they do have creative uses. The idea behind these devices is to monitor the pitch of the incoming signal, then compare it to a user-defined scale, which can be a simple chromatic scale or any combination of notes. Pitch-shifting techniques are then used to nudge the audio to the nearest semitone in the user's scale but, because the amount of pitch-shift required is usually quite small, the result doesn't sound grainy or lumpy, as often happens when large amounts of pitch-shift are generated. Because pitch tracking is used to identify the original pitch, only monophonic signals can be treated.

When used with the human voice, it is important that the pitch correction doesn't happen too quickly, otherwise all the natural slurs and vibrato will be stripped out leaving you with a very unnatural and robotic vocal sound. If only a few notes need fixing, consider automating the pitch-corrector's correction speed parameter so that it is normally too slow to have any significant effect, then increase the speed just for the problem sections. This prevents perfectly good audio from being processed unnecessarily.

If you stick to a simple chromatic scale (all the semitones), you also run the risk of the pitch correction moving the audio to the wrong note if the singer is more than half a semitone off pitch. A user scale, containing only the desired notes, generally works much better. Some systems also allow you to dictate the correct notes via MIDI. If the song contains sections in different keys or that use different scales, it is often simplest to split the vocal part across several tracks and then use a different pitch-corrector on every track, each one set to the appropriate scale for the section being processed.

If your audio track suffers from a lot of spill, or includes chords, the pitch correction may not work correctly. Where spill is loud enough to be audible, you'll hear this being modulated in pitch alongside the wanted part of the audio as it is corrected. As a rule, chords are ignored, so guitar solos, bowed stringed instruments and bass parts (including fretless) can be processed, and only single notes will be corrected.

The main creative application for pitch correction is the so-called 'Cher effect', which is achieved by setting the tracking speed as fast as possible to deliberately generate a robotic-sounding result. It's a matter of taste, but for me, this is one effect that has already been done to death!

Distortion

Technically, distortion is defined as being any change to the original signal other than in level. However, we tend not to think of processes such as EQ and compression as distortion, and the term is more commonly used to describe processes that change the waveform in some radical and often level-dependent way. These include guitar overdrive, fuzz, and simply overdriving analogue circuitry or tape to achieve 'warmth'. In the analogue domain, heavy overdrive distortion is usually created by adding a lot of gain to the signal to provoke deliberate overloading in a specific part of the circuit. Such high levels of gain invariably bring up the level of hum and background noise, so it may be helpful to gate the source. Though overdriving analogue circuitry is the traditional way of creating intentional distortion, we now have many digital simulations, as well as some new and entirely digital sound-mangling algorithms.

The most musically satisfying types of distortion tend to be progressive, where the audio waveform becomes more 'squashed' as the level increases. Hard clipping, by contrast, tends to sound harsh. All these types of distortion introduce additional harmonics into the signal, but it is the level and proportion of the added harmonics that creates the character of the sound. Harmonically related distortion can be added at much higher levels than non-harmonically related distortion before the human hearing system recognises it as such, so there is no way to define a percentage of distortion below which audio is acceptable or above which it is unacceptable. The reason that digital distortion has its own character, which most people find less musically pleasant, is because it is not usually harmonically related to the input signal. For example, quantisation distortion, which results from sampling at too low a bit depth, sounds quite ugly, though many dance and industrial music producers have found a use for it, and some plug-ins deliberately introduce it.

The use of overdrive distortion as a musical effect probably originated with electric guitar amplifiers, where the less pleasant upper harmonics created by overdriving the amp are filtered out by the limited frequency response of the speaker. If you use a distortion plug-in without following it up with low-pass filtering (or a speaker simulator) in this way, you may hear a lot of raspy high-end that isn't musically useful. This is why electric guitar DI'd via a fuzz box or distortion pedal sounds thin and buzzy unless further processed to remove these high frequencies.

The warmth associated with tube equipment and analogue tape is quite subtle when compared with deliberate overdrive effects. As a rule, if you're trying simply to warm up a sound and you can hear the distortion, you should back off it a little, as there's probably too much of it.

Adding a little distortion to sounds such as drums, electronic organs, and even vocals, can help them stand out in a mix, and give substance to a sound that's too thin or uneven. Software guitar amp models often sound more convincing if you use external guitar pedals to create overdrive prior to the audio interface.

A Word Of Warning

Effects are fun, and can make mixing a more creative process, but it's worth bearing in mind that they won't help in situations where the basic principles of recording have been ignored! Used with care, effects can help turn a good mix into a great one, but they are seldom successful in covering up other problems. It is also very easy to over-use them — sometimes their most valuable control is the bypass button, and it is certainly worth learning to use the basic effects well before throwing lots of complicated tricks at your sound. As long as you let your ears decide what is right, you should be OK, and a little critical listening to your favourite records will give you a feel for what works and what doesn't.