In previous chapters we have looked at different processors commonly used in recording studios (EQ, compressors, noise gates...) while discussing some interesting techniques. In this third and final chapter we look at time-based effects, reverbs and delay, and how to use them to create space and give character and grandeur to a mix. * Photo: Musigrama studios.
Delay
In essence, a delay is a processor that generates a copy of the original sound and shifts it later in time. Although this seems like a simple process, we shouldn't be fooled, since delay lines offer many possibilities for manipulating sound and creating different effects.
Much of this potential comes from the ability to modulate the delay time with an LFO (Low Frequency Oscillator). This oscillator causes variations in the delay time, making it fluctuate around the assigned value. The amplitude of the modulator determines how far we move from the initial delay, and its frequency sets how often this variation repeats.
Parameters
Although there are some truly complex processors in the world of delays, most of them are based on these controls:
- Delay Time: the time, usually in milliseconds, by which we delay the original sound. Depth: the possible variation around the delay time. It is usually given as an amount in milliseconds, or as a % of the original delay.
- Rate: the frequency of the oscillator, which determines how fast we move between the two values set by Depth.
- Feedback: the amount of signal at the processor's output that is fed back into its input.
- Mix: the proportion between the original and delayed signals at the processor's output.
Types of delay
Different effects are obtained depending on the delay time used. Stereo positioning, tonal effects and the popular flanger or flanging generally use delay times under 10 milliseconds. Delays between 10 and 50 ms give us doubling and chorus effects, and beyond that value we get tempo delay or echo.
Stereo positioning
One of the cues that allows the human brain to determine the position of a sound source is the difference in time with which the sound reaches each ear. If, for example, the sound arrives slightly later at the right ear, that means it has traveled a longer distance to get there, and so the brain interprets the source as being on the left.
This psychoacoustic phenomenon explains why, if we take a guitar track, duplicate it and pan one copy left and the other right, the net result is that the guitar sounds in the center of the stereo image at a higher level, rather than two guitars sounding at each end of the panorama: the same sound reaches both ears at once, so the brain interprets the guitar as being in the center.
Let's imagine an example with a guitar panned left and an exact copy panned right. For the first 4 seconds there is no delay, and the guitar sounds in the center of the panorama. From the fourth second on, we sweep the delay time from 0 to 5 seconds of delay; the guitar would gradually move to the left, because we have applied the delay to the right channel.
So when we want an instrument to sit on both sides of the stereo panorama, the usual approach is to record it twice. Even when the two performances are the same, there will always be small differences that tell our ear there are two sources instead of one, so each one sounds on one side of the stereo field.
Tonal effects
When two identical sounds are mixed together but slightly offset in time, comb-filtering frequency cancellations occur. The name comes from the shape the cancellations take in the frequency domain: deep, narrow "valleys" are created, as if a comb had been run through the spectrum. Sometimes that special touch we can't get with normal EQ appears when we apply a delay of, say, 0.86 milliseconds.
Flanger
Continuing with effects that use short delay times, the flanger or flanging also adds the LFO, that is, it uses variable delay times.
Originally, this effect was achieved by playing the same material on two analog tape machines simultaneously. By pressing and releasing one of the reels (the metal rim is called the flange, hence the name of the effect), different delays were introduced in one of the machines, causing variable comb filtering. In general, the most striking flanger effects are achieved with high feedback and modulation values.
Doubling and chorusing
Going back to recording the same performance twice, another use of this technique is to achieve the doubling effect. It gives the instrument more body and interest, as the different nuances of each take enrich the overall sound. Lead vocals and guitars are some of the instruments usually doubled on a recording.
This effect, doubling, can be emulated with a somewhat longer delay time than those used so far, usually in the 15-40 ms range. It also uses the LFO, with fairly wide modulation, but it is important that there is no feedback.
When we add feedback to the previous delay, we move into chorus territory. Because feedback creates more than one repetition, it produces an effect that makes it seem as if several instruments are playing the same part (hence the name "chorus").
Tempo delay or echo
50 milliseconds is a key point in the way we perceive sounds, as it is the threshold of the human ear's Integration Time. This means that when two similar sounds occur simultaneously or less than 50 milliseconds apart, the brain integrates both sounds and interprets them as a single source, whereas if the time gap is greater, the ear perceives two independent sounds.
That is why the effect achieved with delay times above 50 milliseconds is called echo, since with these delays one or more repetitions of the original sound can be clearly heard.
It is also called tempo delay because the delay time is often calculated to generate repetitions according to the song's tempo, creating echoes in quarter notes, eighth notes, dotted sixteenths, etc.
Most current processors let you sync the delay to the song's tempo and select the note value you want (eighths, sixteenths, quarters...). If you don't have this feature, here is one way to calculate the delay for a given note value:
Let's take an example with a tempo of 118 quarter notes per minute. This means there are 1.97 quarter notes in one second (118/60≈1.97). Since eighth notes last half as long as a quarter note, there will be 3.94 eighth notes in one second (1.97x2=3.94). To find out how long an eighth note lasts, we calculate the inverse of this amount: 1/3.94≈0.254 seconds; that is, the eighth notes repeat every 0.254 seconds or, in other words, 254 milliseconds. If we wanted the delay in sixteenth notes, it would be 254/2=127 milliseconds, and in quarter notes it would be twice the eighth-note time, 254x2= 508 ms.
Reverb
Reverberation is an acoustic phenomenon that occurs in enclosed or semi-enclosed spaces, caused by sound reflecting off the different surfaces. These time-delayed copies mix with the original sound, prolonging its duration and giving it new qualities.
Because the reflections depend on the size of the room, the position of its boundaries and the materials they are made of, each room has a characteristic sound, and some of them are especially well suited to music.
Reverb processors let us recreate different types of rooms and place our instruments in them, giving them a more "real" sound. They create a sense of space and grandeur, and are the basis for creating different sonic planes in a mix.
Parameters
As with delays, today's reverb processors let you adjust a large number of parameters to control their behavior. However, even with complex units such as convolution reverbs, the essential controls on this type of processor are the following:
Algorithm or reverb type
As we have already mentioned, the way reflections are generated in a space gives it a particular sound. So it is common to find algorithms that simulate concert halls, theaters, churches, rooms, recording studios, etc.
Two special types are spring and plate reverbs. Before hardware or software reverb processors existed, this effect was achieved with devices that used springs or metal plates to simulate it. Since these two types of reverb have been used so much (if you are a guitarist, you have probably heard the sound of the springs rattling when you move the amp), most current processors include algorithms that emulate their sound.
Another special type of reverb is gated reverb. Widely used in the '80s, especially on snare drums, this effect consists of altering the natural decay of the reverb with a noise gate. Usually a somewhat exaggerated reverb is used to get lots of early reflections, and the rest of the sustain is cut off with the gate. Because we are altering the natural decay of the reflections, this effect is also known as non-linear reverb.
An example of non-linear reverb on drums could be the following: the reverb has been applied in different amounts to the snare, hi-hat and overheads on the even passes (2 and 4), while the original drums are heard on the odd ones (1 and 3). The reverb gives the snare body and punch, while slightly taming the hi-hat.
Reverb time
This parameter controls how long the sound from the reflections remains audible. Long times give the feeling of being in large spaces, and shorter times simulate smaller spaces.
Predelay
This is the time between the original sound and the start of the reverb. This parameter is very important for conveying how close to or far from the source the listener is, since the higher it is, the more distance the sound is perceived to have traveled and therefore the farther away we seem to be. It also lets us control intelligibility and clarity by delaying the moment the reverb begins.
Other parameters
Diffusion is a control that regulates the density of the effect's reflections. Usually, a high value gives a greater number of reflections at the start of the effect, and a low value means a lower density of reflections. It is also common to find another parameter, Size, which lets us control the apparent dimensions of the room along with the predelay and reverb times.
Considerations when using reverbs and delays
Because these processors consume a lot of CPU resources, they are usually used in a send/return configuration; that is, instead of inserting a reverb on each channel, an aux channel is created for the effect and the other tracks are sent to it. This way the effect is shared among all of them, and the amount applied to each track is controlled by its send level, much as you would work in live situations.
The aux channel's return can be processed to get different results. For example, it is common to compress the signal coming from the reverb with a fairly fast attack and a long release to get a more consistent level. Likewise, if we want the effect to be less obvious, we can use an EQ to remove high frequencies from the return, or if we want it to be clearer, we can clean up the low frequencies.
Similarly, the send to the effect can also be processed. Placing the same compressor before the reverb/delay will keep the processor from having to respond to wide dynamic ranges, while, for example, adding a high-pass filter at 200 Hz will probably give us a clearer, cleaner reverb.
Sonic planes
The sense of depth in a mix is achieved through the use of reverbs and delays. Generally, the aim is to bring the main instruments to the front and place the rest "behind," in order of importance. Some people even advise imagining the instruments on a stage and then applying effects to achieve that layout in the mix.
In general, instruments with fewer effects applied tend to sound in front of the rest. The "drier," the further forward.
However, we will usually have to establish an order among several instruments that all use some reverb or delay. In this case, longer times (reverb or delay) will make instruments seem farther away, while short times will bring them closer. In a similar way, EQing the effect return can also help: the brighter we make the return, the more prominence we give the effect, and the darker it is, the more it will go unnoticed.
Bearing in mind that reverbs and delays are also used to make an instrument "bigger" (usually by using short times but a high effect level), a good idea is to use one type of processor for this purpose and the other to create depth. So we can use delays to make instruments bigger and reverbs to establish the sonic planes, or vice versa.
*Photos: Estudios Musigrama. Paco Ortega, Musigrama.
