The Acoustetron II from Crystal River Engineering, Inc. performs real-time spatialization of multiple real-time audio sources. For each audio input, the system produces Left and Right outputs, which are mixed and played through conventional headphones, nearphones, or speakers. The processing creates the perception that the source is positioned at any specified location in three-dimensional space. In addition, the Acoustetron II provides a wave file editing option, which converts the sound server into a digital sound recording, mixing, and editing environment.

The Acoustetron II is a stand-alone 3D sound server system that can be controlled via a communication protocol (a RS232 serial connection by default) from any client computer that is capable of implementing the communication protocol.

The system consists of four signal processing cards housed in an industry standard 486DX4 PC host controller. Each card holds a Motorola DSP56001 clocked at 40 MHz, and high-resolution stereo A/D and D/A converters, with input and output sampling rates of up to 44,100 samples per second.

The spatialization software included with the Acoustetron II comprises both a software library and several demo programs. The library routines provide automatic detection of the Acoustetron sound server, and translate high-level commands describing source and listener positioning, etc., into the low-level format needed by the system.

Chapter 1 "Introduction" describes the audio spatialization process and its implementation on the Acoustetron II.

Chapter 2 "Getting Started" reviews hardware and software requirements for using the Acoustetron II, explains what comes with the Acoustetron II (standard and optional) and how to install it, and tells you how to obtain technical support and service.

Chapter 3 "Using the Acoustetron II" examples how to test your Acoustetron II, and how to write your own programs for controlling the Acoustetron II.

Chapter 4 "CRE_TRON Function Reference" describes the function calls that are used to program your Acoustetron II.

Chapter 5 "Glossary" contains a list a often-used terms relating to AudioReality 3D sound.

The spatialization processing of two separate sound sources is illustrated in Figure 1. Each of the two sources is pre-processed through a "Distance Model" filter, which simulates atmospheric loss as a function of the specified distance from that source to the listener. The output of this filter is then applied to both Left and Right filters, which together render the source as if it were coming from a certain direction relative to the listener’s head. Finally, the Left and Right outputs from the individual sources are mixed together.

The Distance Model and the Left and Right filters can be changed dynamically, as often as every 23 msec, or about 44 times per second.

The Left and Right filters are generally known as "Head Related Transfer Functions," or HRTFs. The filter characteristics were obtained from actual measurements on a human head. The Acoustetron II uses Finite Impulse Response (FIR) filters (also known as non-recursive filters).

The Acoustetron II offers dynamic selection of two different kinds of sound sources: external mono inputs and 16-bit sampled mono sound files in wave file format.

The only way to enable/disable the External Input is by controlling whatever is connected to the physical input jacks on each DSP card (turning it on or off, or attaching/disconnecting it) - see chapter 3 "External Inputs" for more details. In addition, a monophonic sound file can be digitally mixed (i.e., by the DSP itself) with the external sources.

Since the Acoustetron II is a stand-alone system, the only hardware requirements are a serial port a on the client system that you intend to use to control the Acoustetron II audio server. When using the ethernet option the client will require a standard ethernet card.

The Acoustetron II is currently available with client software libraries and serial drivers for MS-DOS, IRIX, SOLARIS, and HPUX operating systems. Controlling the Acoustetron server via ethernet will require the client system to support RPC (Remote Procedure Call) .

The main demo application software for UNIX systems, audioClient, requires an X Windows environment with Motif to operate.

The input of the Acoustetron II's DSP cards is not pre-amplified for direct microphone input. When digitizing live sound, use a high quality microphone coupled with a microphone pre-amplifier.

Headphones are an important part of a virtual acoustic system. Your Acoustetron II system has been optimized for use with supra-aural, "diffuse-field equalized" headphones, i.e. headphones which enclose your entire ear, as opposed to headphones that are placed inside the ear canals. For optimal spatialization results, we strongly recommend usage of high-quality headphones.

The reference headphones for use with Crystal River Engineering HRTF models, are Sennheiser HD540 II's (contact Crystal River for availability).

Depending on your virtual audio application, the difference between acoustically open or acoustically closed headphones might be important. Open headphones (such as the Sennheiser HD540 II's) do not provide a tight seal between the ear and the environment. Such headphones are useful in applications where the user has to be able to hear sounds from the surrounding environment (operator's voice, warning signals) during the simulation. Acoustically closed headphones try to suppress all sounds other than the ones delivered by the headphones. They are useful in completely immersive, virtual environments, or to cancel noisy surroundings (trade shows, noisy work spaces).

If you plan to use electro-magnetic head-tracking devices in conjunction with your Acoustetron II , use headphones with as few metal parts as possible, in order to avoid electro-magnetic interference between headphones and tracking sensors. Try to avoid tracking systems which use transmission frequencies between 20-20,000 Hz (the audible region).

In applications where unobtrusive equipment is important, nearphones can be used instead of headphones. Nearphones are two speakers (left and right signal), placed near (within 25 inches) the user's ear. An example of where nearphones are applicable would be a simulator cab with a projection screen and two speakers mounted next to the user's seat. The user can get in and out of the cab by simply sitting down in a chair. To achieve optimal results the user's head should not move out of the range of the speakers, or turn more than 45 degrees in any direction. The closer the speakers are placed to the user's ears, the better the resulting spatialization.

If speakers cannot be placed close to the listener, a quad speaker - left front/back, and right front/back - setup is recommended. In this case, the left output of the Acoustetron II would be wired to both left speakers, the right output to both right speakers. The listener should be placed near the center of the square formed by the four speakers.

The term multimedia speakers refers to a stereo speaker setup where the speakers are built in to a monitor or located close to the sides of a monitor. In such a setup the user's position is assumed to be right in front of the monitor, forming a more or less fixed geometry between the listener and the two speakers. Special processing of the left and right audio signals is applied to enhance the 3D effect in such a setup. Please note that multimedia speaker processing is sweet spot (see glossary) limited.

This category includes all other forms of stereo speaker setups. A speaker layout that does not fit the nearphone, quad, or multimedia categories, is unlikely to produce a convincing 3D effect, and is therefore not recommended.

An Acoustetron II base system consists of the following items:

Follow these steps to setup and test your Acoustetron II:

Your local demo program does not produce sound and does not start properly (no spinning tumbler on the screen):

Your local demo program does not produce sound but there is a tumbler:

The server runs in local demo mode, but does not respond to the client:

If you are having difficulties with the operation of your Acoustetron II, be sure to review the Installation procedure described earlier.

If you can’t solve your problem, you should contact technical support at Crystal River Engineering, at the address or phone numbers listed in the inside front page of this manual. Please be sure to have available as much as possible of the following information:

Before returning faulty equipment or media for service, you first need to obtain authorization from Crystal River Engineering or from your distributor.

Please report suspected or confirmed software problems to Crystal River Engineering, at the e-mail address or phone numbers listed in the inside front page of this manual. It is essential that you include a complete description of the problem, in sufficient detail that we can reproduce it.

On startup, the Acoustetron II system performs a number of self-tests, and then boots up the server. It is useful to talk about two different modes of operation for your Acoustetron II: development or playback . During development of an application, the server should be connected to its monitor and keyboard, in order to make several options and run-time information accessible to the developer. Once an application is developed, and the Acoustetron II is used for run-time playback only, it can be run stand-alone without the need for a monitor or keyboard.

On boot-up of the system the following menu is displayed for a short while:

Please make a choice (you have 5 seconds before the default gets selected, press b key to backup to previous menu)

1. run Acoustetron II server in default mode

2. run Menus

3. exit to DOS

If the ethernet option is installed the menu will be slightly different:

Please make a choice (you have 5 seconds before the default gets selected, press b key to backup to previous menu)

1. run Acoustetron II server in ETHERNET mode (default)

2. set Acoustetron II IP address

3. change ethernet connection type (TP, COAX, BNC)

4. run Acoustetron II server in SERIAL mode

5. run Menus

6. exit to DOS

If you don't type anything, choice number 1 gets selected, and the Acoustetron II server starts up waiting for communications from the client, ready to render sounds. Once the server is running, a few keyboard commands can be issued to control it:

To change the serial link baudrate for the default startup mode, quit to the DOS file system, change to the C:\CRE\ATRON directory (type 'cd \cre\atron'), and edit the file GO.BAT (type 'edit go.bat'). The lines starting with atronbmp.exe and atronars.exe contain /b384 to select a baud rate of 38400. Change the value to /b96, /b192, or /b1152 for 9600, 19200, or 115200 baud respectively. Save the file and quit the editor.

If you select "exit to DOS" in the main menu, the Acoustetron II will quit and allow you to access its DOS file system. Typing 'start' at the command prompt from any directory will start up the menus again.

If you select "run Menus", the following menu will appear:

Please make a choice (you have 5 seconds before the default gets selected, press b key to backup to previous menu)

1. run Acoustetron II server

2. run local demo program

3. run local test program

4. run wave file development program

(only if option installed)

Selection '2' will start up an AudioReality demo sequence.

Selection '3' will start up the BMP1TEST program (see the chapter on "Test programs" below for description).

Selection '4' will start up the optional wave file editing package (see separate documentation for this option).

Finally, choice '1' allows you to start up the server using via serial communications at a non-default baud rate:

Please select the baudrate at which you want to run the server (you have 5 seconds before the default gets selected, press b key to backup to echo previous menu)

1. run server at 9600 baud

2. run server at 19200 baud

3. run server at 38400 baud

4. run server at 115200 baud

To use your Acoustetron II with an existing application, simply connect the client and server systems and connect all audio equipment (headphone amp, cables, headphones). The monitor and keyboard are not necessary for operation, since the system can simply be powered up, and is ready to be accessed from the client after a dual beep has been emitted by the Acoustetron II.

Once hardware and software installation is complete, you may test the Acoustetron II system in a number of ways:

The demo sequence can be started from both the server and the client end. It serves as a verification of server functionality.

A test program available on your client system. It will start up eight sounds that are pre-installed on your Acoustetron II system, and will spin them around the listener's head randomly until the program is stopped (CTRL-C).

A very simple client example program that moves a sound around in space. This program is a good place to look for a simple code sample to get you started on writing your own Acoustetron II applications.

This program is available on the server only, and can be selected from the startup menu on the Acoustetron II ("local demo test program"). It allows you to test the Acoustetron II locally, independent of a client system that is connected to it.

The program places a listener at the origin of a virtual world, and revolves several audio sources around the user's head. Sounds can be created externally or through wave playback. BMP1TEST includes help screens and displays graphical feedback.

When started, the program will display several status messages as it auto-detects for DSP cards and downloads program code and filter coefficients to the card. Once the program starts up, you should see a continuously updated status message showing the current position of source #0:

#0 (aaaa) < xxxxx, yyyyy, zzzzz> 0.0 dB

where aaaa is the azimuth angle (in degrees) to the source and xxxxx, yyyyy, zzzzz specify its front–back, left–right, and top–bottom coordinates (x,y,z) (in inches), respectively. Initially, the source moves about in a 20-in. horizontal orbit, at ear level (z=0) (refer to Figure 4 for a picture of the coordinate system).

The help screen (displayed when pressing the 'h' key) lists a number of single-keystroke commands which move the orbit up or down, or closer or further away, make it slower or faster, change its direction, or pause its movement. You can also raise or lower the output level for the current source, and you can cycle through all available sources (by typing TAB).

If you supply a "live" source connected to the IN jack at the rear of a card (be sure to make this connection with the power off), the Left channel will be spatialized as source #0, and the Right channel will be source #1.

Another help screen appears when the 'F1' key is pressed. It explains the playback of wave files.

The following sample application is provided on UNIX platforms:

If running X Windows with the Motif widget set, the audioClient program can be used to verify the functionality of your Acoustetron II. Once started on the client, it presents a graphical user interface, and a two dimensional graphical representation of the listening space, that allow you to load wave files, play them, and move the sounds and the listener around in a space that includes sound reflecting walls.

The Acoustetron II server can be controlled using the following (optional) "environment variables" on the client system:

The TRONCOM variable needs to be set if you want to operate your Acoustetron II on a different setting than serial port 1, at 38400 baud. The syntax is as follows:

setenv TRONCOM x@yyy,zzz

where x is the serial port number (TTYDx or COMx), yyy the baudrate divided by 100, and zzz the time-out period (the amount of time the client will wait for a response from the server on an init() call).

The TRONDEV variable is optional and only needed if your client system's serial port has a different descriptor than the defaults. For example,

setenv TRONDEV /dev/ttya

When using the ethernet option the TRONDEV environment variable is not used and the TRONCOM has no default. It must be set to the IP address of the Acoustetron II server. For example,

setenv TRONCOM 206.119.89.168

or

setenv TRONCOM <hostname>

CRE_TRON is a 3D audio programming interface that was developed by Crystal River Engineering to facilitate the creation of interactive three dimensional AudioReality sound spaces.

The goal of the CRE_TRON API is to allow a user or developer to build up a sound space using the concepts of physical reality without having to know about the underlying algorithms, implementation or audio hardware.

This API implements the concept of a sound space in the form of easy-to-understand objects. Objects include sound emitting sources, sound reflecting surfaces, and sound receiving listeners. Sounds get created by sources, such as a ringing phone, propagate through space, bouncing off passive objects such as walls, and finally reach a listener's ears, where they are received and interpreted.

The C function calls listed below are used to write programs to control the Acoustetron II. They are described in detail in the "CRE_TRON Function Reference". A good place to start programming is by expanding on the demo.c example code in the CRE/TEST directory.

The function calls that allow a programmer to interact with the Acoustetron II can be grouped into the following categories:

cre_init (driver, head, sources, mode);

cre_update_audio ();

cre_close (driver, head);

cre_locate_head (id, hloc);

cre_locate_source (id, sloc);

cre_amplfy_source (id, dB);

• Propagation medium specific functions, declared in CRE_TRON.H:

cre_define_medium (prm, pts, data);

cre_send_midi (src, midistr);

cre_open_wave (wavefile, mode);

cre_ctrl_wave (src, wave, cmd, data);

cre_close_wave (wave);

ars_amplfy_surf (int surf, float dB);

ars_apply_matl (int matl);

ars_locate_box (float x, float y, float z,

float sizeX, float sizeY, float sizeZ);

ars_lock_box (int mode);

• 44,100 Hz: the sample rate of CDs. Advantage: highest possible quality for audio and spatialization. Disadvantage: highest level of computation.

• 22,050 Hz: the sample rate common in multimedia titles and video games. Advantage: the lower bandwidth and storage requirements increase the number of possible concurrent sounds. Disadvantage: both audio and spatialization quality are reduced (most spatialization information is encoded in high frequencies which get cut off when going from 44kHz to 22kHz).

Please note that independent of driver sample rate, both 22kHz and 44kHz wave files can be played back.

#define SourceID 0

#define HeadID 0

#define Sources 2

#define WaveFile "TEST.WAV"
#define PanLimit 100.0

void main(void)
{
float step = -0.05;

float SrcLoc[6] = { 10.0, PanLimit, 0.0, 0.0, 0.0, 0.0 };
float HeadLoc[6] = { 0.0, 0.0, -10.0, 0.0, 0.0, 0.0 };
wavFt *wave;

/* initialize two Tron sources, with verbose report */
if (cre_init(Atrn_BMP3, HeadID, Sources, 0) < Ok) return;

{
/* amplify source 0 - default is GAIN_dB_OFF (inaudible) */
cre_amplfy_source(SourceID, GAIN_dB_ON);

/* open WAV file and load wave form using all buffers */
if (!(wave = cre_open_wave(WaveFile, 0))) {

printf("\nwave load error.");

return;

}

/* play open wave form as SourceID with repeat loop */
cre_ctrl_wave (SourceID, wave, WaveCTRL_LOOP, NULL);
/* locate listener once (not moving) */

cre_locate_head (HeadID, HeadLoc);

while(!kbhit()) {
SrcLoc[AtrnY] += step; /* move source location */
if ((SrcLoc[AtrnY]<-PanLimit) || (SrcLoc[AtrnY]>PanLimit))
step = -step; /* reverse panning direction */
   /* set new location as location of source 0 in space */
cre_locate_source(SourceID, SrcLoc);
cre_update_audio(); /* flush all changes to DSP */
}
/* stop wave form playback and detach from SourceID */
cre_ctrl_wave(SourceID, wave, WaveCTRL_STOP, NULL);

cre_close_wave(wave); /* close waveform */
}
cre_close(Atrn_CLOS, HeadID); /* close Tron */
}

The example program listed above is a "minimal" program which initializes the Acoustetron II, loads and plays a wave file, turns on a single source to be used, positions a listener in space, moves a source between two points in space, and "displays" the sound space by updating the hardware.

The environment in which spatialized sounds can be experienced is described by a three-dimensional coordinate system. Within this coordinate system, six-dimensional vectors are used to specify the position and orientation of the listener’s head and of all sound sources. The inputs to the Acoustetron II (external or wave forms) are mapped to the corresponding locations in the coordinate system relative to the listener’s location.

Each Acoustetron II can playback and spatialize up to eight monophonic sound files at a time when sampling at 44,100Hz, or 16 wavefiles when sampling at 22,050Hz. In order to play multiple sound files simultaneously, each sound file must be assigned to a different spatialization source.

Before a sound file can be played, it has to be loaded onto the Acoustetron II, either by using the downloadAtron utility from the client, or by loading it via the Acoustetron's floppy drive to the c:\cre\waves directory. Once a wave file is loaded, it can then be used in an application using the cre_open_wave() and cre_ctrl_wave() commands.

If you wish to play the same sound file simultaneously on multiple sources, you must open that sound file multiple times as well.

The maximum length of a sound file is limited only by the size of the hard disk, since files get automatically played back from disk when playback memory is exhausted. If you hear stuttering (gaps) in the wave file playback, the system bandwidth has reached its limit, and too many files are being started at the same time.

If a sound file’s name ends with ".WAV", it must be formatted as a "RIFF" file format (per Microsoft Multimedia specifications). WAVE files can be created on the server using the Wave File Editing option available for the Acoustetron II, or using other sound recording and editing tools and downloading the sounds from the client.

The Acoustetron II can spatialize up to eight sounds that are connected via live inputs. The external inputs can take their data from microphones, CD players, or other sources of analog audio. The physical connectors can be found on the rear panel of the Acoustetron II. Each of the four DSP cards has an 1/8 inch stereo input connector labeled IN. The left and right lines of these connectors map to spatialization channel id's in the following way:

In addition, the AUX connector on each board can be used to mix in stereo signals with the spatialized sound output of the Acoustetron II (however, mixing of spatialized and non-spatialized sounds is not recommended).

The Acoustetron II includes an "atmospheric absorption" model which attenuates higher frequencies a greater amount than lower frequencies. The degree of attenuation depends on the distance through which the sound travels in the atmosphere—the further it travels, the greater the relative attenuation. As a result, distant sounds have a lowpass-filtered, or "muffled" characteristic.

This model is controlled by a "distance" parameter. For sounds which are close to the listener, as compared with the absorption distance, the relative high-frequency attenuation will be slight. Conversely, sounds whose range is equal to or greater than this distance will incur correspondingly more high-frequency attenuation.

As sound waves radiate from a point source, their power spreads over an ever-increasing volume of propagation medium. This spreading reduces the sound pressure level as the sound propagates from the source position to the listener position.

The effect of this "spreading loss roll-off" is governed by an exponential curve which scales a sound’s apparent power by the inverse of the sound’s distance raised to an exponent. In a perfect model, this exponent is 1.0, but anechoic simulations may need some adjustment to sound "right."

The waveform structure and typedef are provided for the application developer to have detailed information on open wave forms, for the purpose of computing useful estimates such length of play. All members are read-only.

typedef struct wavFs {

const char *fname; /* host soundfile filename */

void *pSignal; /* pointer to the signal buffers */

struct wavFs *next; /* linked-list pointer for user */

struct wavFs *synch; /* synchronize with this signal */

int diskBased; /* TRUE if file still open */

double sampleRate; /* in Hz; assumes 44.1kHz */

short numChannels,/* 1 = mono; 2 = stereo */

sampleSize, /* in bytes: 1=8-bit, 2=16-bit */

frameSize, /* in bytes: samplesize*channels */

waveId, /* remote serial identifier */

sourceId; /* value = -1, if not attached */

float pitchFactor;/* pitch shift factor */

unsigned long

numFrames, /* total frames in file */

remFrames, /* # remaining frames NOT loaded */

selFrames, /* length of signal selection */

startFrame; /* beginning of signal selection */

unsigned long

loopStart, /* loop start, in samples */

loopEnd, /* loop end, in samples */

loopCount; /* loop count */

} wavFt;

MEMBERS:

Currently, the following members are not being used:

sampleRate numFrames loopStart

diskBased fname loopEnd

sampleSize loopCount

rad_table[3] = rad_table[4] =

{0.0,-20.0,-60.0} {0.0,-5.0,-20.0,-60.0}

• 3D sound: refers to the fact that sounds in the real world are three-dimensional. Human beings have the ability to perceive sound spatially, meaning that they can figure out where a sound is coming from, and where sounds are in relation to their surroundings and in relation to each other. There are three main pieces of information that are essential for the human brain to perform these functions:

• ITD, or Interaural Time Difference, means that unless a sound is located at exactly the same distance from each ear (e.g. directly in front), it will arrive earlier at one ear than the other. If it arrives at the right ear first, the brain knows that the sound is somewhere to the right.

• IID, or Interaural Intensity Difference, is similar to ITD. It says that if a sound is closer to one ear, the sound’s intensity at that ear will be higher than the intensity at the other ear, which is not only further away, but usually receives a signal that has been shadowed by the listener’s head.

• Finally, the trickiest part of spatialization is the fact that a sound bounces off a listener’s shoulders, face, and outer ear, before it reaches the ear drum. The pattern that is created by those reflections is unique for each location in space relative to the listener. A human brain can therefore learn to associate a given pattern with a location in space.

Since 3D sound consists of two signals (left and right ear) it can be rendered on conventional stereo equipment, preferably headphones (because of the clean separation of the two signals). The 3D sound produced by a direct path AudioReality system is combined with sound reflections (wavetracing) to create a very high level of realism and immersion in a sound space.

• ambient channel: a way of displaying sounds as coming from everywhere - all around the listener. This is useful for background music or ambiance sounds such as rain.

• atmospheric absorption (pp. 3, 27, 42): the attenuation of sounds as they propagate through a medium. For example, in air the high frequency components of sound attenuate faster than the lower frequency components.

• AudioReality: binaural, immersive, interactive, real-time 3D audio technology by Crystal River Engineering (a trademarked term).

• auralization: the process of rendering audio by physically or mathematically modeling a soundfield of a source in space in such a way as to simulate the binaural listening experience at any given position in a modeled space.

• binaural: two audio tracks, one for each ear (as opposed to stereo, which is one for each speaker). Binaural sounds are what we hear in everyday life.

• Convolvotron: the world’s first multi-source, real-time, digital spatialization system built by Crystal River Engineering for NASA in 1987.

• direct path: the direct path from a sound source to a listener’s ears (as opposed to reflections off of surfaces). The direct path allows a listener to tell where each sound is coming from, 360 degrees both in azimuth and elevation. This is the main concept of any 3D sound system.

• Doppler effect (p.57): the change in frequency of a sound wave due to the motion of a sound source or of a listener. For example, if a car moves past a listener while sounding its horn, the listener will hear a sudden drop in pitch as the car passes.

• extended stereo: a term that summarizes a number of techniques that involve processing of traditional stereo sounds with the goal of making them appear to originate from a range which extends beyond the physical speaker locations. The effect is often limited to a planar arc in front of the listener with everything at the same elevation. Extended stereo effects tend to be incompatible with headphone listening and to only have the intended effect if the listener is located at a particular spot in relation to the speakers (see "sweet spot").

• Foster, Scott: the founder of Crystal River Engineering and inventor of the Convolvotron. Often confused with Scott Fisher, his friend and founder of Telepresence Research.

• gain (pp. 5, 31, 35): the amplification or attenuation of a sound source, usually measured in dB (decibels). 0 dB means no amplification and no attenuation. A positive value amplifies a source, a negative value attenuates it.

• HRTF (p. 3): HRTFs, or Head Related Transfer Functions, are a set of mathematical transformations which can be applied to a mono sound signal. The resulting left and right signals are the same as the signals that someone perceives when listening to a sound that is coming from a location in real-life 3D space. HRTFs are the core concept behind AudioReality, since they contain the information that is necessary to simulate a realistic sound space (see spatialization). Once the HRTF of a generic person is captured, it can be used to create AudioReality sound for a large percentage of the population (most people’s heads and ears, and therefore their HRTFs, are similar enough for the filters to be interchangeable).

• IID: Interaural Intensity Difference, see "3D sound".

• ITD: Interaural Time Difference, see "3D sound".

• listener (pp. 1, 19, 23, 51): an object in a sound space that is sampling ("listening to") sound, usually a head with associated HRTF characteristics.

• materials (p. 32): by absorbing sound energy at different frequencies, the material of which an object is made effects the way the sound reflects off and transmits through the object. A carpeted room sounds very different from a glass room. An object’s material characteristics can be measured empirically by recording known sounds as they bounce off of materials.

• medium (pp. 27, 42): see "atmospheric absorption" and "transmission loss".

• mono/monophonic: refers to a single audio signal, usually rendered on a single speaker. Mono sounds appear to originate from the speaker, or from the center of a listener’s head in the case of headphones.

• MIDI (p. 56): MIDI, or Musical Instrument Digital Interface, is a standard control language that is used for communication between electronic music and effects devices.

• psychoacoustics: an area of psychology that studies the structure and performance of human auditory perception.

• quadraphonic sound (pp. 7, 44): refers to four audio signals, usually rendered on four separate speakers. Quadraphonic sounds appear to originate from somewhere in-between the four speakers. The inconvenience associated with the amount of equipment necessary to produce quadraphonic sound, coupled with the fact that it is not compatible with conventional stereo equipment (and therefore headphones), makes quadraphonic sound an unpopular choice.

• radiation pattern (pp. 45-47): each sound-emitting object can optionally radiate sound in a certain pattern (rather than uniformly all around it). For example, a head should emit sounds in the direction that its nose is pointing.

• reflection (pp. 31-34): a sound reflection off of a surface. It gives a listener information about the listening environment and the location and motion of sound sources. See "surfaces".

• refraction: sounds get refracted as they travel around the edges and through openings of objects.

• reverberation: or reverb, refers to the sum of all sound reflections in a listening environment.

• sample rate (p. 20): the number of samples per second at which a sound is processed (usually ranges from 8kHz to 50kHz (CD quality is 44.1kHz, or 44,100 samples per second).

• source (pp. 25, 35, 45, 52): refers to an object in 3D space that emits sound. The actual sound signal that it sends out can be a live signal, a wave file, a MIDI voice, or any other audio signal. A 3D sound device often gets rated on how many different sources it can independently position at any one time. Realistic sound spaces can be created with as few as four concurrent sources, very complex spaces can have dozens of separate sounds at a time.

• speaker arrays: an installation of multiple speakers in a certain pattern, usually designed to create a sound field within the space defined by the speakers. Examples are stereo speakers, or quadraphonic speakers.

• stereo/stereophonic: refers to two audio signals, usually rendered on two separate speakers. Stereo sounds appear to originate from somewhere between the two speakers, or between the ears of a listener in the case of headphones.

• surfaces (pp. 31-34): sounds not only travel to a pair of ears on a direct path, but they also bounce off of objects in the world. Most natural listening environments contain at least a sound reflecting ground plane, such as a floor. Therefore, reflecting objects are necessary to make virtual environments sound natural and realistic. They help listeners navigate and enhance the overall effect of immersion in a virtual environment. Almost as important as reflections, is the absence of a reflection. For example, the brain can tell the change in a sound space when A reflection is removed by opening a door or a window.

• sweet spot (p. 7): the location where a listener has to be placed to get the optimal effect when listening to a specific speaker setup.

• transmission loss (p. 32): sounds get absorbed as they travel through objects such as walls (similar to atmospheric absorption in the case of traveling through a medium). Transmission loss models are needed to realistically simulate sounds outside a window or in the next room.

• update rate (p. 58): the number of times that a specific instance of a sound space gets re-computed and updated per second. Each time any object moves (most often the listener), the space needs to get updated. The higher the update rate, the faster objects can move without creating audio artifacts, such as clicking. Audio update rates generally range from a minimum of 20Hz to 100Hz. Video update rates are usually in the same range (TV signals are updated at 30Hz).

• wave file: a digital sound file stored in the Microsoft RIFF file format.

• wavetracing: the idea of tracing sound waves as they emit from a source and bounce around an environment (walls, objects, openings). The resulting sound reflections are rendered to a listener to create a more convincing 3D effect, as well as a more immersive, familiar, and realistic sound space.

WARNING: This equipment generates, uses, and can radiate radio frequency energy, and if not installed and used in accordance with this instruction manual, may cause interference to radio communications. It has been tested and found to comply with the limits of a class A computing device pursuant to Subpart J of Part 15 of FCC rules, which are designed to provide reasonable protection against such interference when operated in a commercial environment. Operation of this equipment in a residential area is likely to cause interference in which case the user, at her own expense, will be required to take whatever measures necessary to correct the interference.