Time-Related Effects

Time-Related Effects

Delay/Echo

A basic delay or echo effect repeats an input signal at a specified time interval. The most basic parameters are:

What delay does: Repeats the signal at specified time(s)

How to use it: Set a time interval, repeat and filtering options (if available), and mix (dry/wet)

When to use it: Use shorter time intervals to create doubling effects, and larger intervals for rhythmic effects. Use tempo-synced delays to create rhythmic patterns that synchronize with the tempo of your music.


  • Delay time: The length of the delay, usually in ms but sometimes syncable to the tempo of your project.

  • Dry/wet mix: Natural reflections are softer than the original sound, so it's common to reduce the amplitude of the delayed signal.

  • Feedback: A simpler means of presenting multiple delays is with a feedback loop. Instead of just feeding the signal into the delay once, feedback routes the output of the delay (the delayed and possibly filtered sound) back into the input of the delay, in a loop. A feedback level parameter controls the amount of signal routed from the delay output back to the input. Set to its lowest setting, you should hear only one delay. At its highest setting, delays will build on top of one another, and will often ultimately distort.

Delay and echo should not be confused with reverb : delays simply repeat the signal, which can simulate basic reflections in an acoustic space but without some of the timbral and spatial controls found in a reverb. (See p. 42.)

Variants on this theme give you different sets of options, all of which may be incorporated in software in different combinations:

  • Stereo delay: Allows special effects, either by providing separate left and right delay controls, or by automatically feeding delays to left and right channels at different times for a "ping-pong" effect.

  • Multitap delay: Repeats the delayed signal multiple times (or "taps"), often with individual level, pan, and feedback controls for each repetition.

  • Filtered delay: Inserts a multiband filter prior to the delay, so that different frequency components are delayed separately.

  • Tape echo: Produces a wide range of sounds from "spacey" vintage sounds to the distinctive sound of modern dub electronica, which in turn borrows from a Jamaican reggae tradition that used tape echoes. Vintage echoes ( Figure 7.30 ) relied on looping the signal through one or more reel-to-reel tape machines for warm, analog delays and unusual special effects. With feedback, the delay can continue to repeat over itself infinitely, even increasing in level until hair-raising distortion is achieved. Software emulations incorporate simulations of analog circuitry to create warm distortion.

    Figure 7.30. Early tape delays took advantage of the distance between the record head and the playback head of the tape recorder. A signal was fed to the record head (1), recording it on the tape. The tape then had to travel (2) the physical distance from the record head to the playback head. By the time the signal reached the playback head (3), the playback output was delayed. If you took that playback output and routed it back to the record head (4), you could create an endless loop of delays that built one atop another. The resulting warm echoes (from a single delay) or far-out effects (from feedback) are still sought after today in digital simulations. Digital delays simply create the delay through numberbased algorithms, but if you have a two-head tape recorder, you can re-create the scenario shown. (Just keep in mind, different recorders create different delay lengths if the distance between heads or the speed of the tape varies!)

Reverb

Real spaces sound the way they do because of the combination of closely spaced reflections of sound off of different surfaces, such as walls and other objects. When we hear sound, we hear a combination of:

What reverb does: Simulates the many combined reflections created by acoustical spaces such as concert halls and tile-lined bathrooms

How to use it: Use predetermined settings that mimic real-world spaces, and/or manually adjust reflectivity, room size , and reverb length, balancing the wet reverberation with the dry source

When to use it: To make sound seem as though it is heard in a different space than the one in which it was recorded, for subtle, natural effects, or for more dramatic special effects


  • Direct ( dry ) sound.

  • Individual reflections ( early reflections ) from nearby hard surfaces.

  • Other reflections so closely spaced and overlapping that our brain can't perceive them as separate. Instead, they combine to form an overall reverberation , which is filtered by the natural properties of the space.

Imagine the sound you hear in a large cathedral or stadium: there's the direct sound coming from the performer, distinct echoes (early reflections), and a general wash of reverberation, which gradually dies away.

The most direct, though not the easiest , way to create a reverb effect is to play back your source in a space. Some pro engineers still drag speakers and a microphone out to a concrete stairwell to get a reverb "just right."

Since running to a stairwell or a parking garage wasn't always practical, pre-digital reverbs resorted to other means. Plate reverbs mechanically vibrated a piece of steel , employing pickups to record the vibrations in the metal. The plate sound was typical of recordings in the '70s and early '80s, but it required big plates: around 6 x3 ! Spring reverbs worked on the same basic principle but used coiled springs, allowing them to be incorporated into instruments and guitar amps; if you have a vintage amp with a reverb, it's most likely a spring reverb. Since both methods create a distinctive sound, they're both commonly emulated in software.

Specialized digital reverbs produce more realistic results with far greater control over timbre.

When is reverb "in good taste"? Reverb is like ketchup on a burger. Most people like adding a little, but not everyone agrees on how much to use. Both analog and digital recordings can sound a little lifeless if they're completely dry (unless your recording already includes some natural reverb). A very short, low-level reverb can make a track sound warmer and more natural. The key is not to overdo it: add reverb in small amounts to taste. You'll also find you don't have to add reverb to every track in the mix: sometimes just adding a little reverb to one or two tracks provides the illusion of all the tracks being in a real space, especially if it's a subtle effect you want. Add too much reverb, and your sound will be overly muddy. Keep adding, and it'll sound like you used a large ice cave for a studioa bad idea unless you're specifically going for that effect!


Conventional digital reverb

Most digital reverbs (like Audio Damage's reverb in Figure 7.31 ) use a combination of digital processes to produce the illusion of natural reverberationmultiple delays, equalization, envelope shaping, special filters, and other processes. These blend into the sound of the reverb tail, the simulated reflected sound generated by the reverb. Although the underlying algorithm is mostly invisible to the user , a number of useful parameters are commonly found on a typical reverb:

  • Predelay: The interval before a reverb tail is added to the original signal; in real-world situations the sound of reverberation reaches your ears slightly later than the dry sound. In a large concert hall or gymnasium, where the walls are far from the sound source, the delay between the dry sound and the beginning of the reverberation may be more than 50 ms.

  • Early reflections: This parameter controls the level of discrete initial reflectionsthe clearly heard delays at the beginning of the reverb tail, before individual delays combine into a wash of reverberation.

  • Size: Overall size of the simulated space (sometimes displayed in feet or meters as the length of one side of a cube); typically determines the density of the reflected sound.

  • Time: The length of the reverb tail.

  • Diffusion/reflectivity: Simulates the extent to which surfaces diffuse or reflect sound (a stone cathedral will be highly reflective; a crowded club will be highly diffusive).

  • Width: Describes the stereo quality of the output signal.

  • Dry/wet: Particularly important in reverb: make your sound too wet, and you'll lose the clarity of your source.

  • High damping : Determines whether the high-frequency portion of the reverb tail dies out (is damped) more quickly than the low-frequency portion, or perhaps more slowly. Different rates of high damping are characteristic of different reflective surfaces: a hard surface such as tile will produce less high damping than a soft surface such as carpet.

Figure 7.31. Audio Damage's Deverb uses many of the concepts found in a typical digital reverb. The two most essential controls on any reverb are the (1) size controls for how big your virtual room will be and (2) wet/dry mix to balance the reverberation sound with your direct sound. Audio Damage also offers some unusual goodies : (3) dual filters with a "digital rot" knob for a grungier sound, plus (4) MIDI control for rhythmic reverb effects (www.audiodamage.com).

Of these parameters, the most critical are the size, time, and dry/wet settings; nearly all reverbs will provide at least these three. Reverbs also often include filter and sometimes even envelope parameters.

Convolution reverb

Conventional digital reverbs can sound very convincing, and we're now accustomed to hearing the effect they generate. Thanks to the proliferation of faster CPUs, a newer technique called convolution has also become practical in real time, allowing musicians to use hyperrealistic convolution reverbs and create special effects. Convolution reverbs are not necessarily better than synthetic reverbson the contrary, sometimes the "classic" sound of a traditional digital reverb is idealbut they can create effects synthetic reverbs can't come close to reproducing.

Get on the bus! Reverb processors are notoriously the most processorintensive of all effects. Some even provide controls for directly reducing CPU load by simplifying calculations at the cost of some reduction in sound quality. You'll usually want to use a reverb on an effect send bus rather than apply several of them independently to each channel. On the other hand, if you're using contrasting reverbs for different channels, then it may be worth the increased processor load to achieve the individualized sound you want.


Of course, if you wanted the ultimate realistic recording of the way your mix sounds in Notre Dame Cathedral, for example, you could simply drag some speakers into the center of the church, play back your mix, and record the result. Since the folks who run the church probably won't let you do that, the next best thing is to use a pre-recorded, single sound that represents all frequencies at all amplitudes in a single moment of time, a sound that models the way the space resonates ( Figure 7.32 ).

Figure 7.32. You may not be able to record your mix in the Main Hall of the Castle de Haar in Kasteellaan, Netherlands, but with a convolution reverb, it can sound as if you did. AudioEase's AltiVerb (www.audioease.com) was the first commercial computer convolution reverb and includes extras like 3D spectral profiles and (just for fun) photos of the sampled spaces.

This perfect sound, called an impulse , exists only in theory, but in the real world a sound designer can come close to producing it with a starter pistol. The sound of a starter pistol is loud (it saturates the range of possible amplitudes), noisy (it covers nearly all of the audible frequency spectrum), and brief (faster than a speeding bullet). It contains energy at the entire range of frequencies you would likely want to hear in the space. Record the reverberating sound of the starter pistol firing, and you've effectively recorded the entire frequency spectrum as it resonates in your chosen acoustic environment. This recording of the reverb tail of the impulse is called the recording environment's impulse response (IR).

Using a digital audio technique that combines the frequency spectra of two sounds over time (a process called convolution), the reverb can combine your input with the impulse response so that the reverb can model what your sound would sound like if it were in Notre Dame Cathedral, a cave, the bathroom in the headquarters of the software developer, or in any location for which you have an impulse file. Here's how it works:

  1. A sound designer creates or records the impulse , typically by recording the sound of a starter pistol or some other loud, brief transient.

  2. The impulse is played live in an acoustic environment, and the result is recorded; the new recording is that environment's distinctive impulse response (IR), modeling how it responds to all possible frequencies.

  3. The convolution reverb combines your audio source with the impulse response, convolving them into an output signal.

  4. The output you hear is an extremely accurate depiction of what it would be like if your source were actually heard in the modeled environment, provided the source and listener were positioned in the same locations as the impulse and microphone.

Most convolution reverbs include a set of IR sound files (literally, the recording of the impulse in physical spaces) as well as synthetic IRs. Convolution effects also usually include the ability to add your own recorded IRs in case you feel motivated to try recording them. (It's a tricky process, so if you're not up to the challenge, you can take advantage of a variety of impulse files that are available for purchase or free download on the Web.) The convolution process does much of the work, so the typical parameters for the reverb are often simply time (reverb tail length) and dry/wet mix. But many convolution reverbs also add extensive envelope shapers and filtering for finer control over output.

Free Mac convolution: Convolution of two audio files, complete with a unique time-smearing effect, was first popularized on the Mac with a free program called SoundHack (www.soundhack.com). It's still a favorite among sound designers for the bizarre effects it can create.


Convolution can be used to create other effects aside from reverbfor example for "morphing" sounds into one another. As seen in real-time digital reverbs, convolution is one of a class of spectral effects that use a digital-processing calculation called the FFT (see the sidebar "Fast Fourier Transform").

Fast Fourier Transform: Behind the Scenes of Spectral Processing

From essential to off-the-wall plug-ins, many real-time effects rely on FFTs or Fast Fourier Transforms . The FFT is a mathematical process that allows real-time transformations of audio signals. It's essentially a number- crunching shortcut that converts sound between the frequency and time domains. The algorithm can separate sound into component sine waves at multiple frequencies, allowing powerful spectral processing from convolution to other spectral effects and resynthesis. Most likely, if the name or documentation of an effect uses the word "spectral," it's an FFT-based effect.

If you've looked at a software real-time spectrograph of a sound, you've actually seen an FFT at work. Whereas other time- and frequency-based plug-ins, such as analog-style shelving and peaking filters, work with rounded areas of a signal's spectrum, FFT-based spectral processors can manipulate individual frequencies, convolve different sources, and allow wild sonic effects, all in real time. You can work with FFTs directly using modular sound systems like Cycling '74 Max/MSP and Native Instruments Reaktor, and find them in plug-ins like those from Delay Dots (www.delaydots.com), GRM Tools' ST Bundle (www.grmtools.com), and Apple Logic Pro ( Figure 7.33 ). FFTs are also at work in spectral analysis graphs, noise reduction tools, pitch correctors, and all sorts of other essential tools.

Figure 7.33. Logic's Spectral Gate uses FFTs to create, in Apple's own parlance,"some pretty wacky filtering effects." Powerful individual frequency control, thanks to the FFT, enables that wackiness.

Incredibly, although FFTs are behind much of the most cutting-edge sound technology today, genius mathematician and scientist Carl Friedrich Gauss had worked out the basic idea as early as 1805. Gauss was too busy discovering properties of magnetism , polygons, astronomy, and differential mathematics to publish his findings, so it took until 1965 for others to reinvent what he had discovered . Not until the last few years did computer processors make calculating FFTs practical in real time. (As recently as the mid-1990s, composers and musicians would leave computers running overnight to get the results they needed!) You can expect to see spectral effects multiply in the coming years as computers continue to get faster.




Real World Digital Audio
Real World Digital Audio
ISBN: 0321304608
EAN: 2147483647
Year: 2006
Pages: 96
Authors: Peter Kirn

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net