In previous installments we have explored the use of eq and dynamics processing to intentionally modify the character of a sound – i.e., as effects (TCRM 16 and 18 respectively). Other common effects types include delay, reverb, chorus, flange, phase, modulation, emulation, and pitch-shifting. This column will outline these effects types before discussing the specifics of their usage in TCRM 20.
Delay is a fundamental effect, as it is a building block for many others - reverb, flange, echo, pitch-shifting, exciting, doubling, comb-filtering and chorus all rely on delays in one form or another. A basic delay is very simple; it takes an incoming signal and waits a specified amount of time before sending it out. By itself, this can be useful for such things as equalizing latencies or aligning phase differences between microphones. When combined with tricks like recombination, feedback, modulation, and multiplication, delay can be used to create a wealth of diverse effects.
Here are some parameters in a common delay effect:
Delay time – the amount of time a single delay line will hold a signal before sending it out.
Feedback amount – Many delays include an option for taking some of the delayed signal and returning it back into the delays input. This is called feedback. The level of the signal being sent back to the input is known as the feedback amount (alternately feedback level or gain).
Feedback phase – Sometimes the phase of the feedback signal is also controllable; when any signal is mixed back with a delayed version of itself, interference occurs between the two waveforms causing severe peaks and dips in the frequency response. This is further accentuated by the numerous repetitions of a feedback loop. Phase control lets us modify the character of this sound.
Number of taps – A tap is a single delay line, named back in the days of tape-loop echo machines, where multiple playback heads would “tap” into the loop at different points. One single tap will produce a single delayed version of the input sound (sometimes referred to as slapback). When feedback is added, you get a traditional echo, with multiple repeats spaced evenly and fading away. Some delays, however, offer more than one tap, each with its own independent delay time, mix amount, and feedback. These multitap delays allow us to create interesting rhythmic patterns of echo and emphasis.
Delay equalization – To emulate the uneven decay of the audible frequency range in acoustic environments, as well as certain types of musical and audio gear, eq is often added to the output of a delay, or to each tap, to adjust the tone of each repetition.
Modulation – Many delays also offer the ability to change (modulate) the delay time automatically using a low-frequency control signal. This alternately increases and decreases the delay time repeatedly, causing variations in both pitch and filtering that can vary from subtle to extreme. When a delay offers modulation, there will be parameters for modulation rate or frequency and for modulation depth or amount. There may even be controls for the waveform of the modulation control signal, such as sine, triangle, or other shapes).
Combining one or more modulated, delayed signals back with an original, un-delayed source creates an effect known as a chorus (so called because, when used on solo vocals, it sounds like numerous people singing, as in a classical chorus). The modulation (continual variation) of the delay time(s) causes Doppler shifting of the frequencies of each delay line. This means that both time and frequency vary in comparison to the original and make it sound like two or more, slightly different, performances. Chorus effects can be achieved with either multiple discreet delays, and/or by feeding some of the delayed signal back into the input of the delay.
Common control parameters for a chorus include delay time, modulation frequency (often just labeled frequency or rate), depth (or amount). Delay time is the initial time differential between the original signal and the affected one. Modulation frequency describes how fast the delay time moves through its range. Depth determines the overall range of the delay, which determines how lightly or heavily one hears the chorus effect at work on the original signal.
A chorus often spreads the various delayed and modulated signals across the soundfield, creating a wider stereo image. Be aware that tricks like this can sometimes cause extreme cancellations when mixed to mono, especially if left and right signals are phase inverted from one another. It’s best to be aware of the mono-compatibility of such an effect you use in a mix.
A flanger uses a modulated delay, mixed back with the original signal (usually in mono), to create a sweeping comb-filter effect. First, the signal is split into two and one is delayed. The delay time is varied (modulated) automatically and continually (usually by a low-frequency sine or triangle wave) within a range of 0-20ms or so (though many digital flangers allow delays of as much as 500 ms). The delay range is controlled by two functions: the delay time (which generally represents the center of the delay range) and the depth (or amount) of the modulation, which determines the range of the delay time variation. There is also a control for modulation speed or frequency.
When the delayed signal is mixed with the original, the differences in phase cause interference between the two signals, creating both boosts and cancellations of frequencies that are whole number multiples of each other – our old friend/enemy comb filtering….
Sometimes, a portion of the output of the flanger is sent back through the effect. This is called feedback and it can make the flanger resonant by accentuating the filtering effects. When feedback is part of a flanger’s design, the phase of the reinserted signal is often also assignable. This allows extra control over exactly which frequencies are affected by the feedback.
Phasers and Flangers are often confused, as they can be used for similar effects and have the same basic physics at their heart (a moving series of filter notches). But they are actually very different in operation and character, easily differentiated by the trained ear.
A phaser (short for phase shifter) isn’t actually a delay device at all, but a series of filters. These filters are called stages and most phasers have from 2 to 24 of them. The frequencies of the stages are modulated by a control waveform, as on a flanger, and when the filtered signal is mixed back with the original signal there are phase cancellations that change in frequency over time. Phaser controls are similar to those on flangers: center frequency, modulation rate and depth, and sometimes feedback.
Rather than the regular spacing of a flanger’s comb-filtered notches, a phaser’s notches are spaced unevenly and are much fewer in number than a flanger’s. This is what creates a very different sonic character between the two. It’s worth experimenting with both phasers and flangers to learn their individual sonic qualities.
The reverberation (reverb) effect is an emulation of a diffuse acoustic environment through multiple delays and eq. Reverb lends a recording its sense of space. The effects created can vary from highly realistic space emulations (small or large halls, cathedrals, nightclubs, etc.) to virtual environments that could not possibly exist in “the real world.”
Why add reverb? It may not be needed if you used microphones in a great sounding space where the natural environment becomes part of the recording. But as we use more and more sources like virtual, synthesized, and entirely close-miked or sampled sounds, we lose the natural acoustic context of space and may need to add our own back in.
An acoustic space exhibits several particular behaviors over time. First, an observer hears the sound coming directly from the source. Next, strong early reflections are heard bouncing off of nearby walls and other surfaces. Most often, these reflections are first-order, as they bounce off of only one surface before reaching the listening position. Finally, the reflections reaching the observer become more frequent and randomized, coming from all directions as they travel all around the space. They lose energy with the passing of time and the number of surfaces with which they interact. This third period is called the diffuse field (or reverb tail). The greater the number and randomness of the reflections during this period, the more diffuse it is said to be.
A reverb generator, whether it comes in the form of a hardware unit or as software on your computer, will let you control several aspects of the effect. Just how many controls you’ll find and what their exact names will be depends on the manufacturer. Here are some of the most common parameters:
Reverb time is the length of time it takes a sound to die away. The quantified version of this is RT60, the time it takes an incident sound to lose 60 dB of energy in the space. One of the biggest influences on reverb time is the size of the space (its volume in cubic meters). For this reason, controls for reverb time are sometimes just lumped in with the room size control. In other instances these two elements may be kept separate, as reverb time is influenced by more than just size.
Most reverbs have controls for early reflections and/or pre-delay (or initial delay). Pre-delay is the time between the creation of the sound and the onset of the reverb diffuse field. This influences your perception of room size, complexity, and our listening position within a space. The strong, highly directional early reflections are simulated with two or more delay lines, often with user-definable delay times and levels.
The controls for the diffuse field may include diffusion (possibly amount or percentage), density, and level (or balance) as well as possible eq. A diffusion control usually affects how randomized the reflections are. The density of the diffuse elements is a measure of the number of randomized reflections per second. The more reflections, the higher the density. The average level of the diffuse field, specifically as compared to the direct sound or early reflections, is known as its balance. Again, this helps construct the particular character of the room as well as and the nature of the objects within it.
The reverb tail may also use eq to simulate how a room’s materials, as well as the air itself, damp certain frequencies faster than others.
A final master section may also include controls for wet/dry mix (mix of reverb versus incoming signal) and/or a gate. The gate can be used to truncate the reverb tail. This is a common way to reduce noise and improve mix clarity, as well as to achieve some interesting effects. (More on this in TCRM 20).
Pitch processors have become an absolute studio necessity. They can be used to create harmonies and chorus effects, as well as to correct questionable tuning. The subcategory of pitch-shifters intended for chorus and harmony effects are often simply called harmonizers (after the trademarked device of that name made by Eventide). Devices designed to calculate the fundamental frequency of incoming audio and pitch-shift it to match a pre-selected reference and scale are known collectively as intonators or automatic pitch correctors. They are used to automatically fix out-of-tune notes.
Subharmonic synthesizers are also a great way to add depth to an instrument. They can help increase the booty-shaking coefficient significantly by adding simulated lower harmonics of bass frequencies.
Finally, it is now important to include the ubiquitous Auto-Tune effect in any list of pitch-based effects. When Antares first released their automatic pitch corrector, called Auto-Tune, the intent was simply to help tune recorded performances, especially of vocals. Once people discovered the type of robotic vocal effect that could be made by using more extreme, or “incorrect” settings, a new trend took hold across various pop music, especially R&B and Hip-Hop. This effect is now also known as either the “Cher” or “T-Pain” effect, depending on who you talk to.
Every passing month brings more software offerings attempting to recreate the particular sounds of acoustic phenomena or analog devices - such things as analog tape, tube preamps, guitar and bass amplifiers, analog synthesizers, acoustic instruments, electric instruments, microphones, mic preamps, classic compressors, and classic eq. Some even attempt to emulate human performers and musical styles.
Compared to purchasing (or renting or hiring) all of the above, software emulators provide quite the bargain. Here, however, is an area where people are most tempted to buy the latest cool stuff, and yet use it the least (not to mention the time wasted arguing over how accurate these emulations are or are not). This is an area where it’s best to test first and buy with caution, so you reap the benefits of the technology rather than wondering why you spent the cash… especially if you already own some of the gear that’s being modeled!
Amplitude, frequency and ring modulation
As discussed earlier, to modulate means to change over time – a singer’s vibrato, a speakers change from loud to soft and from accented to non-accented syllables, a pianist’s controlled articulation of each note with just the right expression. All of these are musical types of modulation that appeal to listeners.
Synthesizers offer modulation techniques to allow the user to turn even a basic sine tone into something more interesting, complex, and hopefully musical.
When the amplitude of an input signal (called the carrier) is modulated by another waveform (called the modulator), we refer to it as amplitude modulation (AM). When the frequency of the modulator is low (between around 0.2 and 5 Hz) we hear it as tremolo. To produce this, many synths and effects modules use an LFO (low frequency oscillator).
If an LFO is used to modulate the frequency of a carrier (called FM or frequency modulation) instead of it’s amplitude, the result is a vibrato.
When the frequencies of the carrier are brought above this range, both AM and FM techniques begin to create additional spectra, called sidebands. These are entirely new partials, which do not necessarily follow the harmonic series, and are created symmetrically both above and below the frequencies of the carrier. At this point, whole new sounds emerge.
Even more extreme (dare I say alien?) are the results of ring modulation, which is similar to AM except that the original source is not heard at all, only the sidebands.
Basic digital audio theory gives us a bag of tricks for creating lower-fidelity sounds. As more bits and higher sample rates give us more accurate digital audio, fewer bits and lower sample rates, as well as many forms of data compression, create artifacts and byproducts that make for interesting effects. Using fewer bits creates quantization errors (a unique form of distortion), and lower sample rates produce aliasing (new tonal content related to the input signal in rhythm and dynamics but with completely different frequencies). Even the warbling sound of the low data rate mp3, though annoying when you want to hear a precise, clear recording, can be a cool effect when a track calls for it.
Well, that’s it for now. In TCRM 20 I’ll discuss the ways effects are created in a DAW and added to the mix, and offer examples of how they are used. For even more information on particular effects you may want to check out the book Sound FX by Alex Case (Focal Press).
John Shirley is a recording engineer, composer, programmer and producer. He’s also a Professor in the Sound Recording Technology program at the University of Massachusetts Lowell and chairman of their music department. You can check out some of his more extreme uses of effects processing on his Sonic Ninjutsu CD at http://cycling74.com/products/c74music/