Hardware

The job of the audio circuitry in your computer is to set the air in motion, making sounds that you can hear to alert you, to entertain you, and to amaze you. Computers have had sound circuitry from the very beginning. But in the beginning of computers, as in the beginning of life on earth, things were primitive, about the audio equivalent of amoebic blobs bobbing along en masse in black swamps. From humble beginnings, however, computer sound systems have evolved to parity with the stuff in all but the best and most esoteric stereo systems.

From the standpoint of a computer, sound is foreign stuff. Indeed, it's something that happens to stuff—air—while the computer deals with the non-stuff of logical thoughts. Video images are much more akin to computer electronics—at least the photons that you see are electromagnetic. Sound is purely mechanical, and that makes the computer's job of dealing with it tough. To make sound audible, it somehow has to do something mechanical. It needs a transducer, a device that transmits energy from one system to another—from the electrical computer to the kinetic world sound.

Basic Sound System

Although the audio capabilities of some computers rival the best stereo systems, the least common denominator among them is low indeed. The basic sound system that you're assured of finding in all computers is exactly the same primitive design that IBM bolted into its original Personal Computer.

To be charitable, the basic computer sound system wasn't designed for high fidelity. In fact, it was conceived as a beeper. Its goal was to generate pure, if harsh, tones to alert you to events occurring in your computer (for example, the beep code of the BIOS). After all, in 1981 computers didn't sing opera.

This basic sound system has three components—a tone generator, an amplifier, and a loudspeaker—all of which must be called rudimentary because there's nothing lower on the scale. When all worked together, you could make your computer beep as easily as typing Ctrl+G in DOS (this won't work in Windows). The frequency and amplitude of the tone was predetermined by the designers of the first IBM Personal Computer. You were lucky to get any noise at all, let alone a choice.

Clever programmers quickly discovered that they could easily alter the tone, even play short ditties with their data. As programmers got clever, they found they could modulate the primitive sound system and indeed make the computer sing. Considering the standard equipment, you can make your computer sound surprisingly good just by adding the right driver to Windows.

Tone Generator

The fundamental tone-generation circuit is the oscillator, the same as the clock that generates the operating frequency of your computer's microprocessor. The only difference is that the tone generator operates at lower frequencies, those within the range of human hearing (once they are translated into sounds).

The first computers used one of the channels of their computer's 8253 or 8254-2 timer/counter integrated circuit chips as the required oscillator. Modern computers integrate the same functions into their chipsets.

No matter the implementation, the circuits work the same. The timer develops a train of pulses by turning a voltage on and off. The timing of these cycles determines the frequency of the tone the circuit produces.

The computer timer/counter chip starts with a crystal-controlled fixed frequency of 1.19MHz and divides it down into the audio range. A register in the timer chip stores a 16-bit divisor, by which value the timer reduces the oscillator frequency. Loading the highest possible value into the divisor register (65,535) generates the lowest possible tone the basic computer sound system can produce, about 18 Hz, low enough to strain the limits of normal hearing were the computer's speaker able to reproduce it. Divisor values above about 64 produce tones beyond the upper range of human hearing.

Because of the circuit design of the computer, these tones are produced as square waves, which means they are not pure tones but are rich in overtones or harmonics. Musically they sound harsh—exactly what the doctor (or engineer) ordered for warning tones.

In this basic operating mode, the dynamics of the signal are limited. The output of the timer/oscillator chip is set at a constant level—the standard digital signal level—so the sound level produced by the speaker does not vary. All the sounds produced by the computer's motherboard have the same level. Some tones generated by the computer timer sound louder than others primarily because they are more obnoxious. They are made from the exact right combination of frequencies to nag at the aesthetic parts of your brain. That's about all they were designed to do. Listen long enough and you'll agree that the computer's designers succeeded beyond their wildest dreams at creating obnoxious sound.

Using a technique called pulse-width modulation, programmers discovered they could use even this primitive control system to add dynamics to the sounds they generated. Pulse-width modulation uses the duty cycle of a high-frequency signal coupled with a low-pass filter to encode the loudness of an analog signal equivalent. The loudness of a sound corresponds to the length of a signal pulse of a high-frequency carrier wave. A soft sound gets a brief pulse whereas loud sounds are full-strength square waves. The low-pass filter eliminates the high carrier frequency from the signal and leaves a variable-strength audio signal (the modulation).

Amplifier

The output of the tone-generator chip is too weak for sounding a speaker at a listenable level. The chip simply cannot supply enough current. The standard way to boost signal strength is with an amplifier. The basic computer sound system uses a simple operational amplifier. Modern systems incorporate this circuitry into the basic motherboard chipset. In any case, even with the boost, the signal amounts to only about 100 to 200 milliwatts, not enough to shake a theater with THX.

The standard computer design also adds a low-pass filter and a current limiting resistor between the driver and the loudspeaker. The low-pass filter eliminates frequencies higher than normal hearing range (and some in the upper ranges that you probably can readily hear). Computers often use higher, inaudible frequencies as pieces that can be put together to make audible sounds, and the low pass filter prevents artifacts from these high frequencies from leaking into the speaker. In other words, it smoothes things out.

A resistor (typically about 33 ohms) in series with the loudspeaker prevents the internal loudspeaker of a computer from drawing too much current and overloading the driver circuit. A resistor also lowers the loudness of the speaker because it absorbs some power as part of the current-limiting process. Although some circuit-tinkerers bypass this resistor to make their computers louder, doing so risks damage to the driver circuit.

Loudspeaker

The actual noisemaker in the computer's basic sound system is a small, two-to-three-inch dynamic loudspeaker. To most computer designers, the internal loudspeaker is an obligatory headache. They have to put one somewhere inside the computer no matter how inconvenient. Usually the speaker gets added almost as an afterthought. After all, all you have to do is hear it. You don't have to hear it well.

Because computer designers make no effort at optimizing the quality of the sound of the basic speaker, it is usually unbaffled, meaning it is open to the air on both sides (which ruins whatever ability it has to produce bass notes). Its small size and acoustics prevent it from generating any appreciable sound at low frequencies, anyway. Its physical mounting and design limit is high frequency range as well. Outside of replacing the loudspeaker with something better, you cannot do anything to break through these limits.

In most computers, the speaker connects to the motherboard using a simple, short two-wire twisted-pair cable. The motherboard usually, but not always, includes a four-pin loudspeaker connector. In many computers, one of the pins is removed for keying the polarity of the speaker connection. One matching hole in the speaker cable connection often is blocked. Only two of the four pins of the motherboard connector are active, the two at the ends of the connector. The center one or two pins aren't connected to anything. Figure 25.1 shows this connection.

In most computers, the loudspeaker is electrically floating. That is, neither of its terminals is connected to the chassis or any other computer wiring except for the short to which it is soldered. When the speaker is electrically floating, the polarity of its connection is irrelevant, so the keying of the connection is essentially meaningless. In other words, if the speaker connector on your computer is not keyed—if all four pins are present in the motherboard connector and none of the holes in the speaker connector are plugged—don't worry. The speaker will operate properly no matter how you plug it in.

Drivers

Other than beeping to indicate errors, the basic sound system in a computer has a life of leisure. If left to its own devices, it would sit around mute all the while you run your computer. Applications can, however, take direct hardware control and use the basic sound system for audible effects. For example, text-to-speech converters can use the basic sound system with pulse-width modulation techniques to simulate human speech.

Under all versions of Windows, only the native tone-generating capabilities of the basic sound system get used, again just to beep warnings. Microsoft and some others have developed speaker drivers that allow the built-in basic sound system to play more elaborate noises, such as the "Windows sound" when the operating system starts. These drivers let you play WAV files only. They do not synthesize sounds, so they cannot play MIDI files (discussed later) nor do they work with most games.

Some business systems still lack high-quality sound (perhaps as a way for managers dissuading employees from playing computer games on company time). A speaker driver can sidestep the issue, adding some more advanced capabilities to the tiny internal speaker.

The prototype of these drivers is called the Windows speaker driver and was developed by Microsoft strictly for Windows 3.1. The speaker driver has not been updated for more recent versions of Windows. In fact, the history of the speaker driver is even more checkered—it was included with the beta test versions of Windows 3.1 but not the release version. During the development of Windows 3.1, Microsoft found that this driver sometimes misbehaved with some programs. To avoid support headaches, Microsoft elected not to include the driver in the basic Windows 3.1 package. The driver is included in the Microsoft Driver Library and Microsoft does license it to developers to include with their products. It remains available from a number of sources, as are other, similar drivers.

These drivers will work under Windows 95 and some later versions, although Microsoft offers no explicit support of such operation. The official Microsoft strategy is to require a real sound system in your computer if you want to hear anything beyond simple beeps. If you're used to simple beeps, these drivers sound amazingly convincing. They are not, however, a substitute for a real sound system.

Sound Systems

Clearly the basic sound system in computers is inadequate for the needs of modern multimedia. Getting something better requires additional circuitry. Traditionally, all the required electronics get packaged on a single expansion card termed a soundboard. Higher quality audio has become such a necessity in modern computers that most new notebook machines include all the circuitry of a soundboard on their motherboards or an audio riser board (which simply extends the motherboard another level). Many new desktop computers and replacement motherboards also make all the functions of a soundboard an integral part of their designs. The physical location of the circuits is irrelevant to their normal operation. Electrically and logically, they are equivalent.

To cope with the needs of multimedia software and the demands of human hearing and expectation, the sound system needs to carry out several audio-related functions using specific hardware features. Foremost is the conversion of digital sound data into the analog form that speakers can shake into something that you can hear using a digital-to-analog converter. In addition, most sound systems sample or record sounds for later playback with a built-in analog-to-digital converter. They also create sounds of their own using built-in synthesizers. Sound systems include mixer circuits as well that combine together audio from all the sources available to your computer—typically a microphone, the output of the sound system's digital-to-analog converter (which itself combines the synthesizer, WAV files read from disk, and other digital sources), the analog audio output of your computer's CD player, and an auxiliary input from whatever audio source tickles your imagination. Finally, the sound system includes an amplifier that takes this aural goulash and raises it to ear-pleasing volume.

Soundboards may include additional functions, one of which is required by the various multimedia computer specifications—a MIDI interface. This additional connection lets you link your computer to electronic musical instruments—for example, allowing your computer to serve as a sequencer or, going the other way, connecting a keyboard to control the sound system's synthesizer. Some makers of soundboards want their products to act as single-slot multimedia upgrades, so they include CD drive interfaces on their soundboards. With the trend for incorporating audio onto motherboards, however, MIDI has been left behind. Few computers with built-in audio offer built-in MIDI as well. In general, to integrate your computer with external musical keyboards and synthesizers, you'll have to add a dedicated MIDI interface on an expansion board.

Soundboards can be distinguished in several ways. The most important of these divisions are the three C's of soundboards—compatibility, connections, and quality. Compatibility determines the software with which a given soundboard will work. The connections the board supports determines what you can plug in—usually MIDI and CD devices. Quality influences how satisfied you will be with the results, essentially whether you will be delighted or dismayed by your foray into multimedia.

Compatibility

Compatibility often is the more important, because if your software can't coax a whimper from your sound system, you won't hear anything no matter what you plug in or how well the circuitry on the board might be able to do its job. Compatibility issues arise at two levels: hardware and software. More practically, you can regard these levels as DOS and Windows compatibility (or games and Windows, if DOS is foreign to your nomenclature).

For the most part, compatibility refers to the synthesizer capabilities of a sound system. For your games to make the proper noises, it must be able to tell the sound system exactly what to do, when and whether to belch and boom. Your software needs to know which ports to access the functions of the sound system. With today's sound systems, this compatibility is ensured through Windows and the driver software associated with your sound system.

If you want to play old games that sidestep Windows, compatibility becomes an issue. Most old games and other software require compliance with two basic de facto industry standards: Ad Lib and SoundBlaster.

Beyond synthesis capabilities, the reigning standard for computer sound systems is Audio Codec '97, a set of specifications published by Intel that documents an audio architecture originally designed around two separate chips—a controller that handles all digital operations and an analog chip that turns the computer signals into audio. The two are connected by a five-wire serial connection termed the AC link.

The heart of the Audio Codec '97 design are two digital-to-analog converters (DAC) capable of operating at a 48 kilohertz sampling rate to generate a pair of stereo audio channels. The specification requires that the controller be able to simultaneously process four signals: two inputs and two outputs. Each channel must be able to translate the signals to or from at least six sampling rates to the 48 kilohertz basic rate of the system. These include 8.0, 11.025, 16.0, 22.05, 32.0, and 44.1 kilohertz. To maintain true hi-fi quality, the specification requires a signal-to-noise ratio of 90 dB.

The DACs are fed by a mixer that accepts inputs from all the digital sources in the computer, including the synthesizer section of the Intel codec chipset. For inputs, the system includes a pair of analog-to-digital converters that also operate at 48 kilohertz as well as an optional third, matching channel dedicated as a microphone input. Nearly all computers sold today meet or exceed the requirements of this standard.

Control

Sound systems provide important control functions for music-making and audio playback. The mixer circuitry in the sound system serves as a volume control for each of the various sources that it handles.

An audio mixer combines several signals into one—for example, making the distinct signals of two instruments into a duet in one signal. Most audio mixers allow you to individually set the volume levels of each of the signals that they combine.

Mixers work by summing their input signals. Analog mixers work by adding together their input voltages. For example, an analog mixer would combine a 0.3 volt signal with a 0.4 volt signal to produce a 0.7 volt signal. Digital mixers combine digital audio signals by adding them together mathematically using their digital values. The results are the same as in analog mixing, only the type of audio signal differs.

In your computer, however, the difference is significant. The sound system performs analog mixing in real time in its onboard circuitry. Most computers use their microprocessors to do digital mixing, and they usually don't make the mix in real time. For example, your computer may mix together the sounds from two files and let you later play back the combination.

All sound systems incorporate mixer circuitry that lets you combine all the analog signals the board works with. The resulting audio mixture goes to your speakers and the analog-to-digital converter on the board so that you can record it into a file.

To control the relative volume levels of the various mixer inputs, most soundboard-makers provide mixer software that gives you an onscreen slider to adjust each input interactively.

Quality

In sound systems, quality wears more than one face. Every board has designed-in capabilities and, likewise, limits that control what the board might possibly do. Lurking beneath, however, is the quality that a given board can actually produce. Dreams and reality being as they are, most sound systems aspire higher than they perform. This difference is hardly unexpected and would not be an issue were not the difference so great. A sound system may have specifications that put it beyond Compact Disc quality and yet perform on par with an ancient AM radio crackling away in a thunderstorm.

In general, most sound systems list their "quality" capabilities in the range of the digital signals they manage. That is, a given sound system may support a range of sampling frequencies and bit-depths. Nearly any modern sound system worth considering for your computer will list capabilities at least as good as the 44.1KHz sampling and 16-bit resolution of the CD medium. Newer boards should accommodate the 48KHz sampling of professional audio and the DVD system.

In terms of digital quality, the CD rates should be good for a flat frequency response from DC to 15 kilohertz with a signal-to-noise ratio of about 96 decibels. Commercial sound systems miss both of those marks.

The shortfalls arise not in the digital circuitry—after all, any change in a digital signal, including those that would degrade sound quality, are errors, and no modern computer should let errors arise in the data it handles. Rather, the analog circuitry on the sound system teases, tortures, and truncates the signals that travel through it. Both frequency response and dynamic range suffer.

Most of the time, man-handling audio signals makes little difference. When sounds play through typical computer loudspeakers, you probably won't hear the deficiencies. Most inexpensive computer speakers (which means most computer speakers) are so bad that they mask the signal shortfalls. Listen through headphones or through a good stereo system, however, and the problems become readily apparent.

Many computer sound systems shortchange you on frequency response. Certainly the sampling rate of a digital system automatically imposes a limit on the high frequencies that the system can handle. At the other end of the sound spectrum, there should be no such limitation. A digital system should easily be capable of encoding, storing, and reproducing not only the lowest frequencies that you can hear but also sounds lower than you can hear and even levels of direct current. Less expensive sound systems do not dip so low, however. In fact, many boards cannot process low-frequency sounds within the range of human ears, often missing the fundamental frequencies of bass notes and the stomach-wrenching rumbles of computer game special effects. These limits arise in the analog circuitry on the board from the coupling of AC signals through the various amplifier stages.

Most common analog systems require coupling capacitors between amplifier stages (and in the output) to block both supply voltages and direct current errors from interfering with the amplified audio signal. The size of the capacitor (its rating in microfarads, a unit of measure of electrostatic capacity) determines the low-frequency limit of the overall system. Larger capacitors are more expensive and harder to place on compact circuit boards, so manufacturers of budget products are apt to skimp on capacitor size. Better audio circuits use larger capacitors that pass lower frequencies or direct-coupled designs that totally eliminate their need (and push response down to DC).

The size of any capacitor in the amplifier output is particularly critical because, as the power level increases, the required capacity to pass a given frequency increases. Consequently, the first place the low-frequency limit of a sound system suffers usually is in its power amplifier, the part that provides a high-level signal suitable for unpowered speakers. It is not unusual for low-priced sound systems to cut off sounds below 150 Hz, which means few bass frequencies get through. Because the low frequencies are not present as the speaker jacks of such sound systems, plugging in better amplified speakers or even your stereo system will do nothing to ameliorate the low frequency deficiencies. Line-level outputs, as opposed to speaker outputs, are more likely to preserve low frequencies, and they should be preferred for quality connections.

The other signal problem with inexpensive sound systems is noise. The noise level in any digital system is constrained by the bit-depth, and in the typical 16-bit system, the noise floor is pushed down below that of many stereo components. Listen critically to the sounds from many sound systems and CD-ROM drives, however, and you'll hear squeaks and peeps akin to the conversations of Martians, as well as more mundane sounds such as swooshes and hisses. Most of these sounds are simply extraneous signals intercepted by the sound system, mixed with the sounds you want to hear and amplified to offend your ears. Not only do these arcane sounds interfere with your listening enjoyment, but they may also make their imprint on the sounds you digitize and store, forever preserving the marginal quality of your poor choice of sound system for posterity. Most sound systems keep the level of these noises below that which you can hear through inexpensive speakers, but listen to soft music from your CD drive, and you may find your peace shattered by Martian madness.

Better sound systems incorporate shielding to minimize the pickup of extraneous noises. For example, the circuit traces on the boards will be shielded by ground planes, extra layers of circuit traces at ground potential covering the inner traces of multilayer printed circuit boards. Only the best sound systems can digitize analog signals without raising the overall noise level well above the 92 to 96 dB promised by good CD drives.

If you plan on using a sound system for transferring analog audio to digital form (for example, to back up your old vinyl phonograph records to CD), you will want a sound system that guarantees low noise and low-frequency response extending down to 20 Hz.

Digital Audio Output Systems

To avoid the problems inherent in using low-level audio signals inside noisy digital machines, the makers of computers and audio equipment are increasingly moving the sensitive analog circuitry out of the computer itself. For example, professional recording systems may put both analog input and output circuits in an external box connected through a digital port. Consumer-level devices such as USB speakers and MP3 players adopt the same strategy, although for different reasons.

The underlying idea is to eliminate the need for almost any analog circuitry inside your computer and cut the interference from the digital signals, which may be a thousand times louder than the audio, thus making crosstalk (the leakage of an undesired signal into the desired signal) almost inevitable. The result is that the low-level noises you can hear through your speakers when your hard disk rattles into action (as well as any strange sounds that appear when your computer engages in heavy-duty processing) disappear. Your computer becomes a truly hi-fi sound system.

USB Speakers

USB speakers look like ordinary loudspeakers, usually a pair of small desktop speakers or a three-piece system with a subwoofer and two satellites. The only difference between USB speakers and conventional speakers is the connection. As you might guess, the USB speakers use an all-digital USB link with your computer instead of an analog connection. The necessary analog-to-digital conversion circuitry is built in to the speakers.

In theory, USB speakers can eliminate the need for a soundboard or sound system in your computer, at least if you only want to play back sounds. In addition, the speaker-maker can tailor the D-to-A converter and amplifier to exactly match the speaker. It also eliminates a confusing jack from the back of your computer and integrates all your computer's wiring into the USB system. Because the signals used by USB speakers remain digital at all times, both inside your computer and through the cable that transports them into the speakers, they are not subject to noise or the degradation that analog signals suffer. The speakers reproduce the sounds encoded in your audio files in pure, unadulterated form.

In theory, USB speakers should sound better than ordinary analog speakers. But the sound quality of any speaker system mostly depends on the transducers rather than the signals. Small speakers cannot produce room-filling bass notes, for example.

USB speakers have shortcomings, too. Digital audio eats a big chunk from the available USB bandwidth—as a practical matter, once you account for overhead, the speaker connection swallows up about one-third of a USB version 1.0 channel's bandwidth, limiting the number of other devices that can share the USB port. Load the port too heavily, and the sound is likely to drop out for long, irritating gaps. With the introduction of the USB 2.0 interface and its wide bandwidth, such problems are unlikely (but both speakers and your computer's USB port should follow the USB 2.0 standard).

Remember, too, the speaker connection is only one function of a sound system. USB speakers do nothing to assume the other sound system functions. For example, they provide no microphone input. Although you could, in theory, link your microphone to your computer with a USB connection, too, USB mics are as rare as unicorns or dot-com success stories.

Although USB speakers have strong theoretical advantages over conventional analog speakers, in current applications the benefits accrue mostly to the computer-maker. Designing a computer system that relies on USB speakers shifts the costs of the analog circuitry to the speaker manufacturer.

On the other hand, using the high speed of USB 2.0, some multimedia developers are creating entire external sound systems for computers. These have the advantage of moving the low-level audio circuitry out of the noisy system unit. In theory, external sound systems can delivery better overall audio quality.

MP3 Players

One of the truly revolutionary products of digital audio technology, the MP3 player is the ultimate portable music machine. Because it has no moving parts, it holds two key advantages over other portable music systems. No discs or tapes to shake, rattle, or roll ensures that jogging or even leaping hurdles won't cause your music to warble, skip, or entirely stop. No motor to move any parts means that an MP3 player doesn't waste energy battling friction and spinning its wheels during every song you play. The MP3 player therefore needs less power for a given playing time.

At heart, a portable MP3 player is little more than a USB speaker with memory. It downloads digital code, typically through a USB port, and decodes the digits into analog format that you can hear through headphones or speakers. The difference is that USB speakers play back in real time, but the MP3 player remembers and plays on demand.

An MP3 player requires prodigious memory, enough to hold all the songs you want to hear. That's why the "MP3" in the name is so important. Without MP3 compression, the typical player would be able to store little more than a single song. The monotony would make you want to cut any exercise short. With MP3 technology, however, you can put an hour of music in the palm of your hand—or even a day's worth, if you opt for a "jukebox" player with a built-in hard disk drive.

MP3 players also incorporate a control system that, at an invisible level, allows the player to manage its memory and keep tabs on the songs it stores. On the visible level, it lets you organize your music and play the selections you want.

Memory and compression determine how much music you can fit into any given player. Most MP3 players have built-in memory, although many supplement their native endowments with memory cards or Memory Sticks. The minimum in any current player is 32MB of native RAM. At the typical rate of music downloaded from the Internet, 64 kilobits per second, that's enough for about an hour of music—67 minutes. More memory increases playing time commensurately; high data rates for better quality playback cuts playback time. Table 25.5 is a matrix showing playing time for various combinations of memory and quality.

Table 25.5. MP3 Playing Times Versus Memory Capacity and Data Rate

	Data Rate
Memory	64Kbps	128Kbps	256Kbps	320Kbps
16MB	33 min	17 min	8 min	7 min
32MB	1 hr 7 min	33 min	17 min	13 min
64MB	2 hr 13 min	1 hr 7 min	33 min	27 min
128MB	4 hr 27 min	2 hr 13 min	1 hr 7 min	53 min
256MB	8 hr 53 min	4 hr 27 min	2 hr 13 min	1 hr 47 min
512MB	17 hr 47 min	8 hr 53 min	4 hr 27 min	3 hr 33 min
1GB	35 hr 33 min	17 hr 47 min	8 hr 53 min	7 hr 7 min

Transducers

The bridge between the electronic world of audio (both analog and digital) and the mechanical world of sound is the acoustic transducer. The microphone converts sound into audio, and the loudspeaker converts audio into sound.

Microphones

All sound systems have microphone inputs to enable you to capture your voice into the digital medium. You can use digital transcriptions of your voice to annotate reports, spreadsheets, and other files or incorporate them into multimedia presentations. With a suitable sound system, you can even connect high-quality microphones to your computer and make digital recordings of music, edit them, and write them to CDs to play in your stereo system.

The job of the microphone is simple: to translate changes in air pressure into voltage changes. The accuracy of the microphone's translation determines the quality of the sound that can be recorded. No microphone is perfect. Each subtly distorts the translation, not making the results unidentifiable but minutely coloring the captured sound. One side of the microphone designer's art is to make these colorations as pleasing as possible. Another side of the art is to attempt to make the microphone work more like the human ear, tuning in only to what you want to hear, rejecting unwanted sounds.

Technologies

Engineers can use any of several technologies to build microphones. The microphones you're most likely to connect to a sound system to capture your voice are dynamic. A dynamic microphone acts like a small electrical generator or dynamo, using a moving magnetic field to induce a current in a coil of wire. To detect changes in air pressure, a dynamic microphone puts a diaphragm into the path of sound waves. The diaphragm is typically made from lightweight plastic and formed into a domed shape, or something even more elaborate, to stiffen it. The diaphragm connects to a lightweight coil of wire called a voice coil that's wrapped around a small, usually cylindrical, permanent magnet. The voice coil is suspended so that it can move across the magnet as the diaphragm vibrates. The moving coil in the permanent magnetic field generates a small voltage, which provides the signal to the microphone input of your sound system.

Most microphones used for recording music today use a different operating principle. Called condenser microphones (or sometimes, capacitor microphones), they modify an existing voltage instead of generating a new one. In a classic condenser microphone, the diaphragm acts as one plate of an electrical capacitor (which, in the days of vacuum tubes was often called a condenser, hence the name of the microphone). As the diaphragm vibrates, the diaphragm capacitance changes, which in turn modifies the original voltage.

Directionality

Microphones are often described by the directionality, how they respond to sounds coming from different directions. An omnidirectional microphone does not discriminate between sounds, no matter what direction they come from. This type of microphone hears everything the same in a full circle around itself. A unidirectional microphone has one preferred direction in which it hears best. It partially rejects sounds from other directions. Most unidirectional microphones are most sensitive to sounds directly in front of them. Sounds in the preferred direction are called on-axis sounds. Those that are not favored are called off-axis sounds. The most popular unidirectional microphone is called the cardioid microphone because of the heart-like shape of its pattern of sensitivity (kardia being Greek for heart). Hypercardioid microphones focus their coverage more narrowly while maintaining the basic cardioid shape. Bidirectional microphones are, as the name implies, sensitive to sounds coming from two directions, generally the front and rear of the microphone, which resembles the numeral 8. Consequently, bidirectional microphones are sometimes called figure-eight microphones. This design is chiefly used in some special stereophonic recording techniques. Figure 25.2 illustrates the major types of microphone directional patterns.

The inexpensive microphones that accompany cassette tape recorders and some sound systems are typically omnidirectional dynamic microphones. If you want to minimize external noises or office commotion when annotating documents, a better unidirectional microphone will often make a vast improvement in sound quality.

Electrical Characteristics

The signals produced by microphones are measured in several ways. The two most important characteristics are impedance and signal level.

Microphones are known as low impedance (from 50 to 600 ohms) and high impedance (50,000 and more ohms). Some microphones have switches that allow you to change their impedance. Plugging a microphone of one impedance into a circuit meant for another results in low power transfer—faint signals. Nearly all professional microphones and most other microphones now operate at low impedance, as do most microphone inputs. If your microphone has an impedance switch, you'll usually want it set to the low (150 ohm) position.

The signal levels produced by microphones are measured in millivolts or dB (decibels) at a given sound pressure level. This value is nominal. Loud sounds produce higher voltages. Most microphones produce signals described as –60 to –40 dBv, and they will work with most microphone inputs. If you shout into any microphone, particularly one with a higher output (closer to –40 dB), its output level may be too high for some circuits to process properly, particularly those in consumer equipment—say, your computer's sound system. The high level may cause distortion. Adding an attenuator (or switching the microphone with output level switches to a lower level) will eliminate the distortion.

Microphone signals can be balanced or unbalanced. Balanced signals require two wires and a ground; unbalanced, one wire and a ground. Balanced signals are more immune to noise. Unbalanced signals require less sophisticated electronic input circuitry. Most sound systems use unbalanced signals. Most professional microphones produce balanced signals.

You can often convert a balanced signal into an unbalanced one (so you can connect a professional microphone to your sound system) by tying together one of the two signal wires of the balanced circuit with the ground. The ground and the other signal wire then act as an unbalanced circuit.

Connectors

Both inexpensive microphones and sound systems use the same kind of connector, known as a miniature phone plug. Better quality professional microphones with balanced signals typically use XLR connectors (named after the model designation of one of the original designs) with three pins for their two signals and ground. In these connectors, pin 1 is always ground. In balanced circuits, pin 2 carries the positive signal; pin 3, the negative. When used in unbalanced circuits, pins 1 and 3 are usually connected together.

Phone plugs have two or three connections. The end of the plug is called the tip, and the shaft of the connector is called the sleeve. Some connectors have a third contact in the form of a thin ring between the tip and the sleeve. This ring is called the ring. Figure 25.3 illustrates a typical phone plug.

In unbalanced audio circuits, the tip is always connected to the hot or positive signal wire, and the sleeve is connected to the shield or ground. With balanced signals, positive still connects to the tip, negative connects to the ring, and the shield or ground goes to the sleeve. In stereo connections, the left channel goes to the tip, the right channel to the ring, and the common ground or shield goes to the sleeve.

Loudspeakers

From the standpoint of a computer, moving air is a challenge as great as bringing together distant worlds, the electronic and the mechanical. To make audible sounds, the computer must somehow do mechanical work. It needs a transducer, a device that transmits energy from one system to another—from the electrical computer to the kinetic world of sound. The device of choice is the dynamic loudspeaker, invented in 1921 by Kellogg Rice.

The dynamic loudspeaker reverses the dynamic microphone design. An electrical current activates a voice-coil (a solenoid or coil of wire that gives the speaker its voice) that acts as an electromagnet, which is wrapped around a permanent magnet. The changing current in the voice-coil changes its magnetic field, which changes its attraction and repulsion of the permanent magnet, which makes the voice-coil move in proportion to the current change. A diaphragm called the speaker cone is connected to the voice-coil and moves with the voice-coil to create the pressure waves of sound. The entire assembly of voice-coil, cone, and supporting frame is called a speaker driver.

The art of loudspeaker design only begins with the driver. The range of human hearing far exceeds the ability of any driver to reproduce sound uniformly. Accurately reproducing the full range of frequencies that you can hear requires either massive electronic compensation (called equalization by audio engineers) or using multiple speaker drivers, with each driver restricted to a limited frequency range.

Commercial speaker systems split the full audible frequency range into two or three ranges to produce two-way and three-way speaker systems. Modern systems may use more than one driver in each range, so a three-way system may actually have five drivers.

Woofers operate at the lowest frequencies, which mostly involve bass notes, usually at frequencies of 150 hertz and lower. Tweeters handle the high frequencies associated with the treble control, frequencies that start somewhere in the range of 2000 to 5000 hertz and wander off to the limits of human hearing. Midrange speaker drivers take care of the range in between. A crossover divides the full range of sound into the individual ranges required by the specialized speaker drivers.

The term subwoofer is also used to describe a special, auxiliary baffled speaker system meant to enhance the sound of ordinary speakers by extending their low-frequency range. Because the human ear cannot localize low-frequency sounds, you can place this subwoofer anywhere in a listening room without much effect on stereophonic imaging. The other, smaller speakers are often termed satellite speakers.

Baffles and Enclosures

The low-frequency range is particularly difficult for speaker systems to reproduce. The physics of sound require that more air move at lower frequencies to achieve the same pressure changes or loudness, so larger speakers do a better job generating low frequency sounds. But the packaging of the speaker also influences its low-frequency reproduction. At low frequencies, the pressure waves created by a loudspeaker can travel a substantial distance in the time it takes the speaker cone to move in and out. In fact, when frequencies are low enough, the air has time to travel from the high pressure area in front of the speaker to the low pressure area behind an outward-moving speaker cone. The moving air cancels out the air pressure changes and the sound. At low frequencies—typically those below about 150 Hz—a loudspeaker in free air has little sound output. The small size and free-air mounting of the loudspeakers inside computers severely constrain their ability to reproduce low frequencies.

To extend the low-frequency range of loudspeakers, designers may install the driver in a cabinet that blocks the air flow from the front to the back of the speaker. The cabinet of a speaker system is often termed a baffle or enclosure. Strictly speaking, the two terms are not equivalent. A baffle controls the flow of sound, whereas an enclosure is a cabinet that encircles the rear of the speaker.

Not just any cabinet will do. The design of the cabinet influences the ultimate range of the system as well as its ability to deliver uniform frequency response. As with any enclosed volume, the speaker enclosure has a particular resonance. By tuning the resonance of the enclosure, speaker system designers can extend the frequency range of their products. Larger enclosures have lower resonances, which helps accentuate the lowest frequencies speaker drivers can produce.

Most speaker enclosures use one of two designs. Acoustic suspension speaker systems seal the low-frequency driver in a cabinet, using the confined air to act as a spring (which in effect "suspends" the speaker cone in its resting position). A ducted speaker or tuned-port speaker or bass reflex speaker puts a vent or hole in the cabinet. The vent both lowers the resonance of the enclosure and, when properly designed, allows the sound escaping from the vent to reinforce that produced by the speaker driver. Ducted speakers are consequently more efficient, producing louder sounds for a given power, and they can be smaller for a given frequency range.

Although tuning a speaker cabinet can extend its frequency range downward, it can't work magic. The laws of physics stand in the way of allowing a speaker of a size that would fit on your desk or bookshelf or inside a monitor from reproducing bass notes at levels you can hear. For most business applications, reproducing low frequencies isn't necessary and may even be bothersome to coworkers in adjacent cubicles when you start blasting foreign agents and aliens.

Subwoofers and Satellites

A subwoofer extends the low-frequency capabilities of your computer's sound system for systems that need or require it. The distinguishing characteristic of the subwoofer is that it is designed to supplement other speaker systems and reproduce only the lowest audible frequencies, typically those from 20 to 100 hertz. Because these low frequencies are essentially nondirectional (your ear cannot tell where they are coming from), a single subwoofer suffices in stereo and multichannel sound systems.

The classic speaker system puts all the drivers for various frequency ranges in a single cabinet to produce a full-range speaker system. One major trend in speaker design is to abandon the full-range layout and split the cabinetry. Designers put the midrange speakers and tweeters into small cabinets called satellite speakers and rely on one or two subwoofers to produce the low frequencies. Practical considerations underlie this design. The small size of the satellites allows you to place them where convenient for the best stereo imaging, and you can hide the nondirectional subwoofers out of sight. Figure 25.4 shows a satellite-subwoofer combination.

Passive and Active Systems

The speakers you add to your computer typically come in one of two types: active or passive. The difference is that active speakers have built-in amplifiers whereas passive speakers do not.

Most speaker systems designed for stereo systems are passive. They rely on your receiver or amplifier to provide them with the power they need to make sound. Most computer speakers are active, with integral amplifiers designed to boost the weak output of the typical sound system to the level required for filling a room with sound.

The amplifiers in active speakers are like any other audio amplifiers, with output power measured in watts (and in theory matched to the speakers) and quality measured in terms of frequency response and distortion. The big difference is that most active speakers, originally designed for portable stereos, operate from battery power. If you plan to plug active speakers into your desktop computer, ensure that you get a battery eliminator power supply so you can plug them into a wall outlet. Otherwise, if you're serious about multimedia, you'll be single-handedly supporting the entire battery industry.

Most sound systems produce sufficient power to operate small passive speaker systems. Their outputs are almost uniformly about four watts because all use similar circuitry to generate the power. This level is enough even for many large stereo-style passive speaker systems. Most active speakers work with these higher-powered sound systems and, in many cases, deliver better (if just louder!) sound through their own amplifiers.