Better Late Than Never

Image placeholder title

Musicians are forever looking for ways to improve the sounds that they create. Many of the methods they use fall under the heading of “effects,” which can be free-standing hardware devices, specialized chips built into synthesizers, or software processes running on a computer. The purpose of effects, wherever they're found, is to alter a signal in some way.

Image placeholder title

FIG. 1: The stereo delay in Cubase SX3 includes a separate delay time and output panning knobs for the two delay lines. The wet/dry mix slider is at the right side of the screen.

One of the most important categories of effects is the delay line, also known as a DDL (digital delay line) or a delay. Delay effects are easy for manufacturers to create, and they have a number of musical uses. If you use any modern music-production software, it's almost guaranteed that you'll find a delay among the built-in suite of audio effects.

Simple Delay

In the simplest type of delay, a signal is sent through the delay line, which delays it for a short period of time before passing it on to the output. The delayed signal is an unchanged copy of the original. It is typically blended with the nondelayed signal, and you'll usually be able to adjust the relative loudness of these two signals using a wet/dry, a balance, or an output-level control.

The amount of time that passes before the delay line outputs the delayed signal is often measured in milliseconds (1,000 ms equals one second). If the delay time is only a few milliseconds, the delayed signal will not be perceived as separate from the original dry signal. Instead, the combined signal will have a slightly different tone color than the dry signal. A short delay causes tone-color changes due to the phase cancellation of some overtones. (For more about phase, see the Square One column “About Phase” in the May 2004 issue of EM.) Another term for this type of timbral change is comb filtering.

Slightly longer delays, between 50 and 100 ms, create an effect called doubling. Doubling is sometimes used to fatten up vocal tracks. It creates a sound similar to that of a singer recording the same part twice, onto two separate tracks whose outputs were then mixed. The main difference is that when the doubling is done with a delay line, the effect is more perfect and therefore sounds less natural.

A delay of more than 100 ms is perceived as a separate echo that follows the original sound. A delay-based echo can give a track a strong rhythmic character. Because that use of delay is so important in music arranging, your delay effect will synchronize with the tempo of your DAW (digital audio workstation) software. When the delay is synced, you'll set the delay-time parameter in musically sensible units such as quarter notes or eighth notes instead of in milliseconds. The abbreviations “T” for triplet and “D” for dotted values are sometimes seen in delay effects, as shown in Fig. 1.

Short delays are not often used on drum tracks because they smear the crisp attack transients of the percussion sounds. That type of doubled attack, which is called flamming, is usually an undesirable effect. Longer rhythmic delays, however, can be useful with drum tracks.

Delay with Feedback

Most delay effects have a feedback parameter. When that is turned up, a certain amount of the delay's output is fed back into its input. With long delay times, feedback produces repeating echoes. With short delay times, it intensifies the comb-filtering effect and produces a metallic ringing sound.

If the feedback is turned up too high, the delayed signal can get louder over time rather than softer. Such runaway feedback probably won't damage your audio software; it could, however, damage your speakers (and your ears), so be careful when applying large amounts of delay feedback.

A delay line with moderate feedback and a moderate-to-long delay time can be used to simulate a reverb effect. The sense of ambient space is not as convincing as it is with reverb, but a delay line uses fewer CPU cycles, which may be helpful if you're trying to get the most out of a slower computer.

Filtering and Delay

Some delay lines have a lowpass filter in their feedback circuit. As the amount of filtering is increased, successive repeats of the delayed signal have fewer and fewer high overtones (see Web Clip 1). Less often, the filter in a delay may have a highpass mode, in which successive repetitions lose their lows, leaving only the most brittle high overtones.

Image placeholder title

FIG. 2: Shown below is the frequency-based delay in iZotope Spectron. The black curve shows the frequency spectrum of the input signal, and the white peaks show areas where feedback is being applied.

High-end delay software, such as iZotope Spectron, lets you isolate various bands within the frequency spectrum using bandpass filters (see Fig. 2). Using this type of software enables you to give each band more or less delay and feedback, creating exotic shimmering or gargling tone colors.

Modulated Delay

Some delay lines allow the delay time to be modulated, (most often by an LFO), so that the delay time gradually increases and decreases. That has the effect of lowering and then raising the pitch of the delayed sound. If you turn the wet/dry knob to 100 percent wet to listen to only the delayed signal, a modulated delay will sound like vibrato. But when the wet and dry signals are mixed, a modulated delay introduces flanging or chorusing (see Web Clip 2).

In chorusing, two or more signals that are similar in most ways, but slightly different in pitch, are combined. The differences in pitch cause an acoustic phenomenon called beating. A dedicated chorus effect may have two or more modulated delay lines, each with its own LFO. Combining multiple delayed signals in that way creates a fat, acoustically active sound.

The main difference between flanging and chorusing is the width of the modulation. The modulation in flanging is very slight, so that the effect is heard mainly in the overtones as a sort of ringing or swooshing. In chorusing, it's deeper. Flangers often have only one delay line, not multiple delays, and they usually allow feedback. Chorus units generally don't produce feedback. But many delay units can produce either flanging or chorusing as needed, along with other delay effects.

Stereo Delay

Many delay effects offer separate processing for the left and right halves of a stereo signal. Stereo delays are most often used with long, synchronized delay times for creating dynamic and exciting stereo rhythmic effects. A stereo delay may offer three feedback controls: one for the right channel, one for the left channel, and another, called cross-feedback, that sends a portion of one channel's output into the other channel's input.

A ping-pong delay is a stereo delay effect in which the delayed signal with feedback appears alternately in the left and right channels, giving the impression that it's bouncing back and forth.

Multitap Delay

A delay effect that has two or more delay lines, each with its own parameters for delay time, output level, output panning, and feedback, is called a multitap delay. Many multitap delays have four or even more “taps” (outputs). By panning various outputs to different positions in the stereo field and giving each its own output level, it's possible to create striking rhythmic effects, such as syncopated rhythmic echoes that get louder and then softer.

Like other effects, delays can be overused. If you apply rhythmic delay to all of your tracks, the result will be a crowded, cluttered mix. But when judiciously applied, delay is a humble yet powerful effect that can do a lot to spice up your sonic palette.

Jim Aikin writes, teaches, and plays music in Northern California. For more on his activities, visit him online


Many older delay lines, especially hardware units, can't be set to musically sensible time values such as quarter notes. Here's an easy formula for converting rhythmic values into milliseconds:

60,000 / bpm = ms per quarter note

If you divide the number 60,000 by the tempo in beats per minute, you'll get the number of milliseconds in a single quarter note. To get the correct ms value for an eighth note or other subdivision of the beat, simply divide or multiply the quarter-note value as needed.