Creature Sound Design
October 16, 2006

In Return of the King, the tremendous thumps of Oliphant footsteps were designed by dropping large boulders from a crane. In order to create a cleaner, more dramatic effect, sounds were applied to the front feet only.

Usually when watching fantasy, science fiction, or horror films, or playing video games, we never stop to wonder how the voices of the supernatural beings were created. In fact, if we do, it might be because the voice is so unnatural, that it's distracting. Instead we should be wanting to know more about the creature, because in some way, when we become engrossed, we suspend our disbelief that the creature could not exist outside of the screen.

In his book, The Practical Art of Motion Picture Sound, David Lewis Yewdall, supervising sound editor for The Thing, says:

"I have harped on this throughout this book and will continue to do so, because it is the keystone of the sound design philosophy -- if you take the audience OUT of the film because of bad sound design -- then you have failed as a sound designer. It’s just that simple."

Personally, I am distracted from film when I hear sound effects that are clichés of other films. To me, these are lost opportunities, and the worst examples of audio design. Although we should look to others' work for inspiration, we should be inspired by things outside of the entertainment industry so that we can bring in new ideas. There is a great deal of imitation going on in film and video games right now, but still, the only major achievements are done by innovators. Ben Burtt, sound designer for the Star Wars series, explained it like this:

"Your success as an artist, to say something new, ultimately depends on the breadth of your education. My recommendation would be to major in an area other than film, develop a point of view, and then apply that knowledge to film. Because if film is all you know, you cannot help but make derivative work. I found that what I had learned about sound, history, biology, English, physics all goes into the mix."

In this article, I would like to discuss some past and recent examples of creature sound design techniques. Also, I will introduce some alternative digital signal processing (DSP) methods that can add further FX potential than can be achieved by mixing. I have some thoughts to share on current effect mixing trends, and what to avoid in these trends. Additionally, I will provide insight into other fields of research whenever possible, and not give step-by-step methods, in order to inspire others to explore original ideas. The information here can be synthesized in individualized and unique ways, to produce new ideas for creature sound design.

Alchemy

"One becomes two, two becomes three, and by means of the third and fourth achieves unity; thus two are but one...." - Maria the Jewess, 300 A.D. (Sharp)

To be a sound designer is to gambol with the forces of alchemy. We venture into the world to find powerful new elements, contemplate their properties, speculate about their inter-conversions, then hide away in dark chambers to divide, filtrate, mix and ignite them. This is a dark art, pursued by sonic sorcerers. I am fascinated by the chemistry of SFX creation, an enigmatic, largely unexplored and undocumented art form that has only just begun.

When we listen to a recorded sound, it is not always obvious what we are hearing; without a visual cue, we might not know what the sound belongs to. We might hear the sound of crinkling cellophane and recognize that it produces a familiar sound yet not know what it belongs, or is related, to, but if this cellophane sound is placed on top of some scenery with a cloudy sky, we will believe that it is rain that we are hearing.

Picture a whale coming up for air. It's exhale is just a release of pressurized air, so we could use any similar sound in its place. With a little processing, we could swap it for a flamethrower burst and the scene would gain some energy. We could also layer in the sound of an elephant call, our motivation being that we want the sound to be larger than life. By replacing the softer bellow of air with something more thundering and textural, we’ve just made the scene more dramatic, without anyone knowing that they’ve just heard a special effect. Illusion is the basis of sound design, a magic trick that has not changed since rolling cannon balls provided the sound of thunder in Shakespearian theatre.

The most straightforward way to create original creature sounds is to remix and process recordings of animals, a technique pioneered by Murray Spivak, head of the sound department of RKO studios who mixed sound for the 1933 release of King Kong. Kong's roar consisted of a lion's roar played back at half speed, layered with a tiger played in reverse. This technique was mind-blowing for a time when silent films were still being released. Right up until the 60's, special effects sounds and even animal sounds were being done with foley effects gadgets and machinery, a method that was borrowed from theatre and radio. It is impressive what the sound editors were able to achieve with normal objects and magnetic tape recordings. David Lewis Yewdall spoke up for this ancient art, saying:

"When these fine craftspersons labored to make such memorable sound events for your entertainment pleasure, they made them without the help of today's high-tech digital tools -- without Harmonizers or Vocorders, without a Synclavier or a Fairlight. They designed these sounds with their own brains, understanding what sounds to put together to create new sound events -- how to play them backward, slow them down, cut, clip, and scrape them with a razor blade (when magnetic soundtrack became available in 1953), or paint them with blooping ink (when they still cut sound effects on optical track)."

Unfortunately though, these ingenious sounds were always pretty much the same, and eventually ended up in archives. Studios came to rely more on stock recordings from sound libraries, instead of reproducing the sounds that were needed every time. This indeed cut down on costs, but one could hear the same foley sounds across hundreds of films. It wasn't until such innovators as George Lucas joined the industry that we escaped the old school techniques and began to utilize new audio technologies to create sound effects. Lucas had the financial backing to initiate major developments in sound and fortunately, he had two audio geniuses at his disposal, Walter Murch, and Ben Burrt.

	Chewbacca's voice consists of several animal sounds, primarily a hungry bear.

Ben Burtt, inspired by the sound work in King Kong, took the technique of remixing animal sounds to a whole new level. Burtt was one of the first people in film to be responsible for all aspects of the audio track, as up until that point the task was collectively handled by sound recordists, editors and mixers, with sparse communication between. He started by building a personal sound effects library during the preproduction of Star Wars, traveling around to zoos, factories, airports, and museums with a microphone and tape recorder. The initial idea for Chewbacca's voice was realized when he heard a walrus stranded at the bottom of an empty pool at Marineland. After some experimentation, he decided on several ingredients including the moans of a hungry bear, walrus, dog growls, and lion roars. Numerous variations of these samples were meticulously clipped and layered in sequence across 4 or 5 tracks, in time with the actor's mouth movements, to develop a singular vocalization. The design of this voice is so smooth and consistent that we recognize it immediately as Wookie, without ever knowing that we are hearing 4 different animal sounds mixed together.

If you are an independent sound designer, without the help of a production crews, it would be quite difficult to gather all of your own recordings of animal sounds. You would have to do a lot of research and traveling, the locations would not always provide good recording conditions and you would need authorization for many sites. Furthermore, zoos, for example, are extremely noisy with lots of surface reflections and the expectation cannot be made that the animals would simply walk up to the microphones and make that sound that is sought after. As an alternative, we can do well by relying on licensed nature sound effects libraries for some of the staple animal sounds, then go out and do special recordings to supplement them.

Of course, I am just using animal sounds as one example. When we are making the sound of living organisms we can also mix together recordings of normal objects, music, our own voice...Anything. For example, a large flat stone, like a tombstone, sounds very much like a low growl when sliding over a concrete surface. If we take that rumbling sound, the sound of chanting Tibetan monks, and a great big sigh from our own mouth, we might have three very good ingredients for the exhale of a menacing demon.

The greatest sounds are the result of innovation, both deliberate and coincidental. Not only should we look to films for inspiration, but also approach our designs with speculative thought, personal experience, and with knowledge obtained from outside of the entertainment industry. We can certainly make a career out of doing imitations and spin-offs, but if we want to be creators then we had better put our goggles on and flare up the torch of invention.

Animal Communication

Much of the time, the director or designer will simply count on human voices for creatures. If this is the case, you will clearly have to make the voice intelligible. But consider for a moment the impact that creative sound design can have on the overall perception of the organism. The voice of Treebeard in The Lord of the Rings, The Two Towers, is a good example of human voice customization. His voice is deep and resonant, like it is really coming from within a tree, and mixed with his creaking limbs and thundering footsteps, a personality emerges that is thousands of years old and far from human.

Other times the voice will be semi-human, or non intelligible, animal-like. When we are to design sounds for more unusual organisms, a good place to look for inspiration is nature. In fact, recordings of animal sounds may be all the ingredients that you need to start creating your own unique beings. In order to get a better idea of our options, lets look at a few communication methods that animals have devised.

In general, when we think of sounds that animals make, we usually just consider what comes from it's mouth, such as a coyote howl. It lifts it's head upward, and with muscles compressing in it's diaphragm, air is pushed from the lungs and through the voice box. As the coyote tightens it's vocal cords, which resemble elastic bands, air squeezes through in a steady stream of short bursts, creating an audible vibration. This waveform is then shaped by muscles in the throat, the oral and nasal cavities, movement of the tongue, the lips, and numerous other factors that contribute to the sound that echoes out and disappears into the environment.

	In order to create the sounds of creatures in The Chronicles of Riddick (such as the Hellhound), audio designers used field recordings of a black leopard, Bengal Tiger, and several types of wolves. These were recorded directly to DAT using a Neumann 191 and Sennheiser 461 shotgun.

For various biological and behavioral reasons, animals such as big cats and howler monkeys, need to produce sounds at a much higher volume than vocal cords alone could handle. The roar of a lion, for instance, can reach 114 decibels, which is about the volume of a chain saw at full throttle. In order for this to be possible, the hyoid bones that makes up the shape of the lions larynx (voice box) is made of soft cartilage connected by a 6 to 10 inch long ligament. When the cat pushes air through its larynx and tightens this ligament, the entire larynx vibrates. This vibration is further amplified by a horn shaped waveguide, that helps amplify the roar so that it can be heard for miles around. This unusual voice box structure allows them a sound palette quite unique in comparison to animals that use vocal cords.

Lets consider methods of vocalization that do not require a larynx. Spiny lobsters, for example, do not have vocal cords, but they can produce shrieking sounds in the same way that a squeaky door hinge does. This sound is created by a continuous slip-stick-slip-stick friction between flesh and exoskeleton inside the base of the lobster's antennae. In order to do this, the lobster needs to push at least one of its antennae back and forth while contracting muscles on either side of it's eyes where the antennae protrude from. It is thought that lobsters produce these sounds to ward off enemies, but still, it is unknown.

Many living things create sounds by an action called stridulation, which is achieved by rubbing two bony organs together. This is most commonly known in insects and arthropods such as crickets, millipedes, and stag beetles. For insects, this is achieved by rubbing legs or talons together which have microscopic ridges on them. Many underwater inhabitants such as catfish and seahorses create clicks and squeaks in the same way, but by rubbing skeletal parts together, sometimes within their fins. These sounds are produced as alarms, to attract mates, as a fearful response, or to antagonize competitors.

Porcupines have special quills on their tails that are shaped like tubes, and which they can shake to produce a steady rattling sound, in order to warn others of their presence, similarly to a rattlesnake. They also stomp their hind feet, click their teeth, growl and hiss. They are actually quite timid in personality, but they are also talented performers, and can put on a convincing act of aggression.

	By pushing air through their blowholes, Blue Whales, the largest mammals on earth, produce subsonic booms that can be heard for thousands of miles around.

Whales produce a wide range of sounds through their blowhole, reaching far beyond human range of hearing. Male humpback whales use their blowholes to sing rich complex songs during mating season. These vocalizations can go on for twenty minutes, and they can repeat these performances verbatim, which may evolve for years on end. Like bats; sperm whales, killer whales and dolphins use echolocation clicks to find food and navigate through dark underwater environments. Blue whales can produce low frequency booms anywhere from 2-20 Hz, with enough force that they can be recorded half a globe away. Above the water, we can hear whales shooting huge plumes of vapor into the air, and quickly sucking in huge breaths before descending again.

Birds do not possess a larynx, instead they use what is called a syrinx which is positioned deep down within their chest. It is uniquely different than the larynx in that it is split into two halves, creating two vocal cords, which allows them duophonic note capability. Some birds, like thrushes can use their two vocal cords independently, such as singing falling and rising notes simultaneously. Some song birds alternate between the two vocal cords, use them in detuned unison, designate each side for a different frequency range, or even modulate the frequency of one side with the other. These methods are all remarkably similar to the way that dual oscillators are used in monophonic synthesizers. Most birds do not produce songs at all but instead call out in short unmusical vocalizations in order to communicate special calls for warnings, offense, territorial defense, feeding, nesting, flocking and even enjoyment.

There is some evidence that animals are capable of telepathic communication. In 1952, an island called Koshima in Japan, scientists were feeding sweet potatoes to Macaca fuscata monkeys and watching their behavior. The potatoes were dropped onto the beach, making them sandy, and one monkey, a young female, came up with the idea of washing the potatoes in the ocean water. Not only did it wash the sand off but it gave the potatoes a nice salty taste, which caused a potato washing craze that spread throughout the colony by evening. A few days later, something happened that caused monkeys on other islands and the mainland to start washing their food in the ocean, a trait that had never been observed before the occurrence at Koshima. The thought is that when a certain critical number of monkeys had reached awareness, the information spontaneously spread through the collective unconscious, from mind to mind. This famous and well-documented event is known as the "hundredth monkey" incident. Although telepathy isn't thought of as an audible phenomenon, sound would be a great way to portray its occurance in film and games.

Five DSP FX for Creature Creation:

Here is an assortment of advanced processing tricks for creature sounds. By manipulating the sounds of animals, the human voice, ordinary objects, or electronic sounds we can create the voices of fantastic beings. The main focus being on technologies that are more unusual than mixing and equalization. Also, these processes are very natural and can fully warp the input signal without sounding like effects. Some of these techniques would have been impossible just a few years ago, without the help of advanced processors and recent software developments. By combining these methods in different ways, we can come up with some truly original effects.

1. Morph

In 1988, for a film called Willow, Industrial Light and Magic created the groundbreaking digital effect of a human shapeshifting into an animal, earning them a Technical Achievement Award. This effect was coined "morphing" and soon appeared in television ads, music videos, and several other motion pictures. Another good example of this effect is the "mimetic polyalloy" robot in Terminator II, morphing from "liquid metal" into the various things that it samples through touch, such as other human beings. It's quite an impressive effect, even when used judicially, such as in the Lord of the Rings, when Gandalf releases King Theoden from Saruman's spell and Theoden's face morphs back to normal.

	Prosoniq Morph. Moving the grey dot around adjusts the degree that input signals are morphed. MIDI control signals can be assigned by the X/Y coordinates.

In the 90's, a medical software company called Prosoniq became interested in the application of artificial neural networks to sound manipulation. This algorithm models human perception by creating artificial nerve cells that can adapt and learn, in order to imitate such complex processes as the way hearing works. Prosoniq used this intelligent artificial awareness, to create a program called "Morph", which performs the sonic equivalent of shapeshifting. With Morph (Mac, PC), one can re-synthesize two audio signals in real time, and create hybrid versions of them, sharing the acoustic properties of both.

An older package for doing audio morphing, although it's algorithm is completely different, is Symbolic Sound's Kyma (Mac, PC), a hardware based visual programming language that has been used a lot of major motion pictures, including War of the Worlds, Lord of the Rings, and Star Wars: Attack of the Clones. For morphing, Kyma uses a Fast Fourier Transform model, which adds some amount of warbling and ringing in comparison to the "adaptive transform" of Morph.

Using morphing technology, one can create very natural sounding transforms between disparate sources such as human language, animal sounds and environmental effects. The effect can be a static mixture, or a morph from one state to another. This is a great tool for making unique creature sounds, especially since the effect is rather clean and unidentifiable. This effect is particularly alluring for people to experiment with because of its connection with metamorphosis, folklore, and magic.

2. Granulate
.

As the Zen Master Taisen Deshimaru once said, "Time is not a line, but a series of now points." What I think he means is that we should not be worried about the past and future, and even though we experience life as a sequence of events, but by being present and focusing on what lies right before us, we focus on the most purposeful things in our lives.

When editing, or even thinking about how we might do our edits, we are re-sequencing the time framework. Sitting in front of Final Cut with some bad film footage, for example, our thoughts jump around to various points in the shoot and we begin to imagine a new series of edits. As we cut, drag and re-order the shots we are giving the scenes new substance, new physical dimensions, a new passage of time. After we have edited out the bad parts and reconstructed the scene, we can sit back like Zen masters and experience the meaning that encompasses the film.

The smaller our edits, the more capability we have to manipulate the scene. With the help of an advanced sequencer, we can break sounds or images into molecular pieces, which gives us the ability to warp and coalesce our media like controlled vapor.

In 1947, a British physicist named Dennis Gobor proposed a theory in which any sound could be broken down into particles of time and frequency. To demonstrate his ideas, he created mechanical devices that allowed him to perform time compression and expansion of sound and moving pictures. In the 1970's, a composer named Curtis Roads, having a great interest in Gobor's ideas, started programming the first granular synthesis software, a technology allowing recorded instrument sounds to be granulated for the creation of music. It was a fantastic idea, perhaps a bit too complex for early computers and it was not until the 1990's that CPU's could handle the amount of data necessary to make this effect available to musicians and sound designers.

Granular synthesis is a technique that involves breaking sounds up into tiny grains and playing them back in a controlled manner to create longer events called clouds. The grains are basically just small portions of the waveform that loop and the cloud is a large number of these grains playing back at the same time. A good analogy is the way particles of water floating in the air look like mist when you see them from a distance.

The specific grain that is triggered at any given point is generally defined by the location of a short 10-50ms section within the larger waveform consisting of a start point and a stop point. Both of these points can either move around within the waveform independently, move around together, or be stationary. If the loop points were stationary we would probably just hear a steady buzzing noise, but as we sweep it around the waveform we hear flowing sonic movement. Each time a grain starts it overlaps the previous grain by a small amount so that the grains blend together. In this way all the grains sound like one continuous waveform when played back; a smooth flowing cloud.

	Native Instruments Travelizer offers the basic goods for granular synthesis. The loaded sample is visible on top and the X/Y pad on the right adjusts grain size and position.

We have control over the cloud densities, that is, how many grains are heard at one time which can be hundreds or thousands per second. The key to getting the most pleasing sounds out of granular synthesis is to keep the grains moving about and use smooth attack and decay envelopes so that when the grains are played in high density they all blend together nicely.

The end result is that we can completely alter the texture and shape of a waveform, independent of its timing. This means, for example, that you could play back all portions of a sound at the same time, sweep it both forward or in reverse, move around the waveform in complex patterns, or even compress and expand the length of a waveform without altering the pitch (the effect is commonly known as pitch shifting.) Depending on what granular synthesizer you use, this technology offers a tremendous number of sound shaping options.

I recommend Native Instruments Reaktor (Mac, PC) for making creature sounds. There are two ensembles included with the package: Grainstates and Travelizer. Grainstates is more of a step sequencer with multiple modules; well suited for musical sounds. Travelizer is a single module with very advanced capabilities. On both ensembles, the interface is an X-Y pad, which can either be controlled with a mouse, trackball, or graphics tablet. The X dimension is assigned to grain start, and the Y dimension controls the width of the grains. One important parameter is "inertia" which can smooth out mouse movement, resulting in sweeter, more flowing clouds.

3. Shift

	Serato Pitch 'n Time

Pitch shifters are effects processors that alter the pitch of a waveform without changing it's timing. The first commercially available pitch shifter was the Eventide H910 Harmonizer, an 8 bit audio processor released in 1975 that was first made popular by Bowie's producer, Tony Visconti, for the doubled snare drums in "Low". These days, pitch shifting effects are usually handled via software and include the option of formant correction, a process that preserves the natural characteristics of the human voice when forced into different pitches.

It is important to distinguish this effect from simply adjusting the playback rate, such as changing the playback speed of a tape or record, which also adjusts the length of the recording. When we pitch shift, in order to preserve the timing in the original sound, the waveform is broken up into tiny sections, a process similar to granular synthesis. Therefore, when the pitch of the sample is changed each of these tiny sections plays at a different speed, but their position in time, relative to one another, does not change.

In order to create the voice of the 2005 King Kong, sound designer Brent Burge used the live performance of Andy Serkis, the actor performing the physical motions of Kong, captured for the CGI animations. These recordings were made at 192 KHz directly into Pro Tools so that the pitch could later be dropped substantially without losing higher frequencies. His pitch shifted voice was fed to loudspeakers so that Serkis and the cast could hear an approximation of the effect in real time. Burge later added lion, tiger, elephant, bear, and gorilla sounds to further build and refine the character of the voice.

Pitch shifting plugins are pretty numerous and as Craig Anderton stated in the review of DSP/FX, "The only real question about a pitch-shifter is how badly it messes up the sound when you do extreme transposition." Celemony Melodine (Mac, PC) and Serato Pitch 'n Time (Mac) provide transparent results when performing non-real-time pitch shifting. Ableton Live (Mac, PC) and Prosoniq Timefactory (Mac, PC) also work well for non-real-time shifting. For real time pitch shifting, Waves Soundshifter (Mac, PC) is a good sounding DirectX plugin.

4. Filter

In the late 18th century a scientist named Wolfgang von Kempelen invented a mechanical speaking machine. This tabletop device consisted of a bellows, an oscillating reed, a rubber mouth, nostrils, and numerous mechanical controls allowing manual operation by a single person. With some practice one could produce words and short sentences by pumping the bellows, simulating the lungs, and using levers to change the shape of the artificial vocal tract. By no means was it able to accurately emulate the human voice, but its sound was clearly that of a human being. Up until the early 20th century other scientists and engineers continued this research, building numerous mechanical voice machines, making only minor improvements on Kempelens model. In 1939, at the New York World Fair a scientist named Homer Dudley presented the Voder, the first electrical speech synthesizer, a significantly less mechanical attempt at voice emulation. By the 1950's, speech synthesizers had adopted the system of using an electrical tone fed through a filter to create artificial speech.

	MHC Vox FX

A filter that is used to approximate the resonances inside an acoustic body, in our case the vocal tract, is called a formant filter. If we compare the human voice to a synthesizer, the larynx would be the oscillator and everything that makes up the vocal tract would be the formant filter. In order for these filters to make vowel sounds they must be fed with harmonically rich waveforms such as a sawtooth, pulse, or white noise. Most synthesizers do not include formant filters but they can be constructed by using an assortment of 3 or more bandpass filters in parallel. Either way works fine; I prefer the power and ease of using a plugin, such as MHC Vox FX (Mac, PC), which allows user defined programmability and also morphing between parameter settings.

It is best to select source sounds that are full-bodied, with lots of texture and high end, so that the filter's shape is well defined. For creature sounds one really powerful trick is that you can run normal object sounds through the filters, like creaking doors, air bursts, or scraping sounds, giving them a vocal character that blends well with animal sounds. Sometimes formant filters can make animal sounds more artificial sounding which is useful for sci-fi effects, if you want something that sounds a bit more android. Automating the frequency and bandwidth of individual formants over time will simulate the mouth changing shape. The sound of this can lean towards synthetic if the adjustments are dramatic. Since the formants are resonant, the sound may leap out unrealistically at times, so it is helpful to use a dynamics compressor or limiter before or after the filter, in order to tame the peaks.

5. Convolve

When creating creature sounds, it is necessary, especially in film production, to use source sounds that are as dry as possible, meaning free of ambient reflections so that the ambience can be added later. For example, we might not want to be using dolphin clicks recorded at an indoor aquarium as a design ingredient, because it could impart a particular ambience that would not blend naturally with the rest of the sounds and scenery. But for some creatures, acoustic reflections are an integral part of the voice. For example, if we are designing sounds for a being that lives underwater, an ethereal ghost, a creature that communicates telepathically or perhaps even a large robot, the reflective nature of the sound could be so important to the end result, that it should be designed into the voice. If our creature is wearing a helmet, or some type of mask or if it has a tube shaped appendage for a mouth then that might call for an effect simulating the voice as it travels through or around these items; the reflection and obstruction of sound. We could use standard reverbs and equalizers for tasks like this, but one can achieve much more powerful and realistic results with convolution software.

Convolution reverbs use what's called an "impulse response" (IR) to produce realistic reverb effects. Usually the IR is made by playing a short burst of white noise or a starter pistol shot into a reverberant space while making a stereo recording of it. If we played this IR back, it would, of course, just sound like nothing more than a blast with a bit of reverberation. With the help of convolution software we can use this IR to process another sound. The convolution reverb multiplies the frequency content of the IR with the frequency content of the input signal (the sound we want to process) which in turn makes the input signal sound like its happening within the same acoustic space that we recorded the IR in.

Making an impulse response of an acoustic space is similar to making a 3D model of a house using photographs. We walk around the building taking a photograph of each side which we then cut out and glue together to reconstruct the building. When we are finished, we have a very realistic looking scale model of the house, complete with its own lighting conditions and fine geometric details. If we took a movie camera and panned around the model of the house, it would generally look no different to the viewer than the actual house. We've captured the way light bounces of the house and stored it in a convenient little model that we can take anywhere. Maybe though, there are some imperfections in the photographs such as a tree or a telephone wire that we couldn't avoid when taking the picture. Maybe one side of the house is a little out of focus, or unnaturally dark.

	Altiverb with a vacuum cleaner hose impulse loaded.

The quality of the IR recordings is very important, especially for realistic effects. It is usually necessary to have an extremely neutral and flat frequency response in the recording gear. The noise floor of the microphone must be extremely low and the background should be entirely silent, as anything that goes into the recording will end up in the reverb. Also, the IR will sound drastically different according to microphone placement. Unless we are doing something unusual, sometimes it is better to simply use commercial impulse libraries, which are usually excellent for reverbs, especially since some of them were made in spaces that the average sound designer couldn't access such Yankee Stadium, a volcano in Hawaii, or a forest grove in New Zealand.

The space that the impulse is recorded in can be as small as a paper cup. Not only that but we could also make the IR by recording sound as it travels through water, fabric, around corners, coming out of a tiny hole, through a crowd of moving people, or through electronic devices such as vintage radio speakers. Since the software simply multiplies the spectra of the two signals together, we can use any other sound effects for impulses, allowing us to achieve effects similar to morphing and pitch shifting. When this effect is used in non-standard ways the results can be very unpredictable.

It is important to know that there will always be some smearing of attack transients, therefore the output signal will seem a bit smoothed out. Also, any frequencies that are prominent in both signals will be louder in the output and frequencies that are not prominent will be attenuated. In a way, the signals filter each other, because the only sounds that get through are frequencies that they have in common. If you are loading in particularly dissimilar sounds, on occasion, you will not hear anything at all. Likewise, if the signals are very similar, or the same, the output may be so loud that it is distorted. For experimental convolution, the best results are achieved when the two waveforms share some frequencies but not others.

I recommend Altiverb (Mac, PC), SIR Convolution Reverb (PC), and Waves IR. The things that matter most with convolution reverbs are CPU load, the quality of the impulses, and the programmability. The Altiverb impulse library is particularly good and includes a photograph for most of the impulses, giving a clearer idea of how they were made. The Synesthesia library from Virtuasonic is an interesting set of impulses. When processing lion roars with them I happened upon some sounds that were strikingly reminiscent of the reverberation effects heard during the Mines of Moria scenes in Fellowship of the Ring.

Dimension of Personality

Special effects are clearly a major part of films these days. Hearing the impressive effects in Star Wars when I was younger is partly what lured me into the field. Yet, even though today’s SFX are much more dramatic and use state of the art technology, I feel more detached and disinterested by today’s blockbuster films. The personalities of the characters seem to be pushed into the background, making them feel less believable.

	HSSSSSSSSS!

There is a simple precept in entertainment for some characters to be corrupt and powerful, and for others to play the role of good and be in need of defense. We are so familiar with this good vs. evil template that once the script introduces this prescribed formula, our curiosity about each character's motives is pretty satisfied. A famous quote from Alien Resurrection is when Ripley says, "There's a monster in your chest. These guys hijacked your ship, and they sold your cryo tube to this... human. And he put an alien inside of you. It's a really nasty one. And in a few hours it will burst its way through your ribcage. And you're gonna die. Any questions?" That pretty much describes what the aliens' intentions are in full detail. The aliens are simply evil and they will kill the good guys. Then we move our attention onward, to unsolidified aspects of the story, like which characters are going to live. This is an example of a creature with a one dimensional personality, with little room for anything but nasty sounds. The sound designer makes a loud nasty screeching sound for them and everyone is happy.

In Spirited Away, a film by Hayao Miyazaki, a spirit called Kaonashi, meaning "no face", appears as a mysterious masked man, partly translucent, silent and passive in nature. The main character, Chihiro, invites him into a bath house, the place where she works. Kaonashi follows her around, offering gifts, which she refuses. When he offers things, he makes the quiet sound of someone helpless and innocent, perhaps even hurt, which naturally makes us believe that Koanashi is benevolent. He offers gold nuggets to other people in the bath house, staff members, which they all gladly take, and in turn he eats them. Koanashi uses this friendly sound to lure people in so that they give him food for the gold, which he always devours. If they do not give him food he actually feasts on whomever took the gold from him. After spending more time in the bath house he slowly becomes enraged, grows to massive proportions, and devours everything in sight. The bath house staff does not fight him, but continue to treat him as a guest, although more cautiously. Kaonashi begins to speak in a shrill comical voice that he wants to eat Chihiro. Unafraid of Kaonashi, Chihiro tries to help him, believing that the bath house is making him crazy. She eventually gets him out, and he slowly returns to the silent, benign form.

This is an example of a creature that has multiple personalities, and two completely different voices, one almost silent, one very loud, which enhances the depth of its character. It could have had one voice sitting comfortably between nasty and nice but with a dual voice, it's personality gains complexity, it's motives are more ambiguous, and it is less predictable.

The Loudness Trend

In her book Sounding of the Soul, Mary Lynn Kittelson stated "With myths being trumpeted forth with over-whelming dramatics, overpowering high tech effects and high volume sound systems, there is less gently imaginative space to imagine the characters and actions and settings."

The structures of most films and video games are so familiar to us that our attention often cannot find anything new to focus on except the latest special effect, which is never enough. The sound of today’s films is fantastic but there is this surge forward to fill more and more of the sound field so that every pocket of silence is gone.

The Exorcist (1973) used restraint to frighten us, a brooding quietness. In this way, our imaginations ran wild, creating a fear in our chests that has never been surpassed in any other film to date. The director, William Friedkin says:

"What we did back then was very instinctive; you put the dialogue in the way it feels right, you play the effects and the backgrounds for whatever emotions you want to get out of them. But today the dialogue is all in your face, the sound effects are in your face, the music’s right up there and they just beat you over the head with them all for hours. That’s completely different to THE EXORCIST which was actually often a very silent movie – a loud silence."

Rather than using a dramatic sound effect for ever single action of the creature, consider the impact that silence can have on the scene. Also, if you want any particular effect to jump out, then dial back the volume on other effects. What many filmmakers do not understand is that the way to make a film louder, is not to push everything up to full, but to actually pull back. This subtractive method will allow more headroom in the signal, allowing dramatic events to jump out far beyond everything else.

Epilogue

I hope that the information provided here will be inspirational. Remember to keep the audience engrossed and create creatures that draw attention to themselves rather than provide something safe and familiar. Instead of only researching your favorite films, research science, nature, and biology to create your beings. Experiment. If you do this, you will come up with an immense amount of new ideas, unprecedented techniques that you can bring into your projects and the industry. Being inspired by other sound designers is a powerful propellant, but dig in and find out what makes their work so inspirational to you and use that knowledge as a launch pad for innovation. If you steer clear of the common course and come up with your own recipe, you might truly leave your mark in the field of sound design.

George Sanger, when speaking about John Williams' contribution to Star Wars, and how it continues to inspire similar soundtracks, said:

"Begin by imitating his compositions, but go on to learn how to create his feelings. When you imitate his orchestration, continue, and learn how to be evocative of emotions. When you imitate his greatness, learn how it progresses; learn how it leaves its own roots behind."

Research Materials:

Papageorge, John. Interview. Radio Interview with Ben Burtt. Silicon Valley Radio. 1996

Lord of the Rings, Special Extended DVD Edition: The Fellowship of the Ring. Dir. Dir Peter Jackson. New Line Home Entertainment. 2001

Lord of the Rings, Special Extended DVD Edition: The Two Towers. Dir. Dir Peter Jackson. New Line Home Entertainment. 2002

Lord of the Rings, Special Extended DVD Edition: The Return of the King. Dir. Dir Peter Jackson. New Line Home Entertainment. 2003

Star Wars. Dir. George Lucas. Perf. Mark Hamil, Harrison Ford, and Carrie Fisher. Paramount, 1975.

Brunelle, Ray."The Art of Sound Effects." Rumblestiltskin Sound FX Design. (1999): n.pag. Online. Internet. Available. web.archive.org

Benedikt, Reinhard G."Prosoniq Morph - Behind the scenes." Bernsee.com 1 Nov (2003): n.pag. Online. Internet. Available. www.dspdimension.com

"King Kong: Mammoth Post Power Yields Epic Results" Avid (2006): n.pag. Online. Internet. Available: www.avid.com

Sasso, Len. "Fun with Formants." Electronic Musician. 1 Jun. (2006): n.pag. Online. Internet. Available: emusician.com

Traunmüller, Hartmut. "Wolfgang Von Kempelen's Speaking Machine and its Successors." History of speech synthesis, 1770 - 1970. (2000) n.pag. Online. Internet. Available: www.ling.su.se

Gordon. "Granular Synthesis." Granular Synthesis. 16 Oct (1996): n.pag. Online. Internet. Available: homepages.kcbbs.gen.nz

Kemmler, Alex. "Acting on Impulse." Electronic Musician. 1 Jun. (2006): n.pag. Online. Internet. Available: emusician.com

Cited Materials:

Alien Resurrection. Dir. Jean-Pierre Jeunet. Writ. Joss Whedon (screenplay). 20th Century Fox, 1997

Deshimaru, Taisen. "What the Bleep Do We Know!?" What the Bleep!? (2004): n.pag. Online. Internet. Available: www.whatthebleep.com

Kittelson, Mary Lynn. The Sounding of The Soul. Einsiedeln: Daimon, 1996

Sanger, George Alistair. The Fat Man on Game Audio: Tasty Morsels of Sonic Goodness Indianapolis: New Riders, 2003

Sharp, Daryl. "Extracts from The Jung Lexicon: A Primer of Terms & Concepts." Compiler Press (2002): n.pag. Online. Internet. Available: www.compilerpress.atfreeweb.com.

"The Sound of Silence." The Exorcist, the Film, the Phenomenon, the Aftermath. (2000): n.pag. Online. Internet. Available: theexorcist.warnerbros.com

Wally, David. "Aural Surgery." USC Trojan Family Magazine (2001): n.pag. Online. Internet. Available: www.usc.edu.

Yewdall, David Lewis. Sound Design Myths and Realities. New York: Focal Press, 2003

[back]