Introduction

Computer speaker systems have brought a new element to the computing experience.  Though they may be a form of analog technology, they are not yet archaic or outdated.  Instead, speakers serve to merge the digital world to the sensory world that we live in.  Ultimately, they have served as a catalyst for multimedia computing—what good would an MP3 or webcast be with just subtitles?

Our society has been bred on television as a medium.  Integrated video and audio is a standard set in our childhood that is now fueling a bandwidth hunger via the Internet.  As a result, computer speakers are as important as the monitor in our computing environment.  The growth of computer audio has also resulted in more choices in size, styling, and format of speaker systems.  The challenge is to find a high quality speaker system that will heighten the computer experience through quality and power, instead of an ill chosen system that holds back the entire computer.

The Decibel

Before jumping directly into the technical nitty gritty, some explanation of the measurement standards used in audio is necessary.  For everyone that knows the difference between dBm and dB SPL, and especially how they relate to the audio industry, feel free to skip directly to the specification breakdown.  Audio specifications are not exactly transfer rates or IP addresses, so some background can be helpful.

The key unit in audio is the decibel.  Technically one tenth of a Bell (yes, it is named after Alexander Graham Bell, and carries the capitalization when abbreviated), it is a logarithmic scale that compares two power quantities.  Logarithmic, besides being a hard word to spell, is a technique used when relatively large quantities, such as 10 and 1000 for example, are compared to each other.  A logarithmic scale compare the difference based on the ratio of the second quantity to the first instead of the difference between the two as in a linear scale.  Figure 1 demonstrates the logarithmic curve of the decibel scale; note that it does not follow a straight line.  In the dB scale, a change from 1 watt to 100 watts would be measured the same as a change from 10 watts to 1000 watts because it compares the first value to the second.  Both changes would be measured as 20 dB (see the equations below).

Figure 1: The Logarithmic Decibel Curve

Because the dB is a measurement based in comparison, it can be used in various applications.  In audio, the two primary comparisons are made with power and force.  The electric power in a system typically is measured against a reference level of one milliwatt (.001 watt).  Though it may seem arbitrary, this level comes out of standards for interconnections and the corresponding powers dating back to the early era of radio, first set forward in 1940.  The equation that dictates decibels measuring power is dB = 10 * log (P₁/P₂).  If you work the math, this means that twice the power results in a 3 dB difference, so specifications dealing with audio signal power are dictated by this rule of three- 3 dBm = 2x the power.

The electrical analogy of a speaker’s air movement (what makes the sound), is voltage in a circuit.  As the human senses are logarithmic, it is convenient to continue to use the decibel scale when describing speaker volume.  However, because the analogy is to voltage and not power, the Bell equation must be slightly modified.  Power is related to voltage by the equation P = E²/R.  Substituting this into the earlier equation, and moving the exponent to outside the log results in the equation for comparing Sound Pressure Levels (SPL): dB SPL = 20 log (E₁/E₂).  The reference level (0 dB SPL) used for the scale is the threshold of hearing for a youngster (before the heavy metal music that starts to kill his or her ears).  The threshold of pain is about 120 dB.  As this equation has been slightly modified, twice the SPL equals a 6 dB difference.  Therefore, the sound pressure levels in speakers are dictated by a rule of six: 6 dB SPL= 2x the sound pressure.  As a note, the human body perceives about 10dB SPL to be twice as loud—again, the body acts as a logarithmic scale.

Well, enough of the math review.  Let’s dive into the specifications and find out what’s really going on in the speakers.

Speaker Specifications: How to read between the lines

Before listening to any speaker system, the manufacturer’s specifications provide a basis for comparison.  Through a variety of tests to determine the sonic quality and electrical characteristics of the speakers, manufacturers determine the numbers intended to reveal the performance of their systems.  However, results can, and often are, presented in an ambiguous manner that appears to enhance the specs of a system.  In order to reveal these possibilities, careful examination of each specification is necessary.

·         Drivers:  The actual speakers used in the cabinet.  The physical components of a speaker are fairly simple.  The cabinet, or enclosure, is the wood or plastic box that contains the drivers, or speakers, and the electronics involved in making the sound, such as an amplifier and crossover networks.  This specification indicates the size and type of transducers (speakers) used.

The number of drivers is very revealing as to the nature of the speaker system.  When there are more speakers, they can be tailored to fit particular sound spectrums as each component handles just part of the frequency range. In a two way system, low and mid frequencies are reproduced by the woofer while a tweeter reproduces the highs.  A three way system is when a subwoofer reproduces the lowest frequencies, a mid range woofer powers the intermediate frequencies, and the tweeter provides the high frequencies.  When a transducer handles a smaller frequency band, it can handle higher levels of power; sound is energy and if they are reproducing a more limited range of energy, they can handle higher levels.

In these multi speaker units, a crossover network is used to split the sound spectrum and route the parts to the appropriate speaker.  The crossover can also control the relative output levels for the speakers, actually balancing the sound across the frequency spectrum.  Most importantly, they are set so as not to overwork the speakers and shorten their life spans by sending high energy frequencies to speakers not designed to handle them.

Here is a breakdown on the components often used in speaker systems.  Through a combination of these components, a speaker is able to reproduce the full (or almost full) spectrum of sound.

o Tweeters are used for high frequency sound, usually sounds over 1.5 kHz.  A smaller transducer can reproduce high frequency sound.  Compression drivers and piezoelectric transducers can also serve as tweeters in a speaker system.

o Woofers are used for low frequency sound reproduction up to around 1.5 kHz.  Because low frequencies have a longer wavelength, woofers must be larger in order to move the air volume necessary to reproduce these lower frequencies.  As a result, the larger a woofer, the more power it can carry, and the louder the bass will be.  In addition, for all the rumbling low frequency noise, the body senses the energy as much as feels it.  When used for both low and mid frequency sound, woofer will produce sounds up to around 1.5 kHz

o Sub woofers are use to reproduce the lowest frequencies, usually up to 500 Hz.  These provide the rumble element to the sound system as they move the most air and shake the body as much as the eardrums.  The larger the sub woofer driver, the more energy it can handle, and the more impressive it can be.

·         Frequency Response:  How much and how accurately the speakers can reproduce the sound spectrum.  The response is measured using an analyzer that reads the speaker’s reproduction of a standard audio test signal.

Human hearing spans the frequency response of 20 Hz to 20 kHz, and in an ideal world, a speaker would reproduce all of these frequencies.  Unfortunately, when indicating the frequency response of speakers, manufacturers often fail to tell the whole truth.  If they do not indicate how accurate the speaker response is, specs can often look far better than the speaker actually sounds.

The simple description of frequency response of 20 Hz to 20 kHz would seem ideal; however, this is a true statement even if the sound at 20 Hz is 40 dB SPL lower than the sound at 1.2 kHz.  This means that the lowest bass frequency is a hundred times less powerful than an average midrange frequency— i.e. the speaker may reproduce all of those frequencies but nowhere near at the same level.

A much more clear method of specifying frequency response involves giving a tolerance, the range within which the speaker produces all of the frequencies within its frequency response range.   For example, a frequency response of 20 Hz to 20 kHz +/- 3 dB indicates a much superior speaker to the one mentioned earlier with a bass roll off of 40 dB.  Basically, the latter speaker maintains its level all the way into its lowest frequency, while the former just goes away (rolls off) in its lower frequencies.  The lowest bass frequency is at most only half of what a typical mid range frequency is reproduced at.  Without indicating the tolerance on the specifications, companies can create extremely misleading specifications. 

Figure 2: Frequency response chart for anonymous speaker - click to enlarge

Specifications given in figure 2 could indicate that this speaker responded from 35 Hz to 20 kHz—it reads on the diagram across that entire spectrum.  However, the red box indicates what frequency range would be offered with a tolerance of roughly +/- 5 dB, or a range of roughly 50 Hz to ~17kHz.  Both specifications would be true, but only the second one would be very honest.

If a speaker company will not indicate a tolerance in their specifications, it raises the question of what they are trying to hide.  In addition to indicating how the ends of the frequency spectrum, the tolerance indicates the accuracy of the spectrum.  The tolerance indicates that at no point in the spectrum are there extreme spikes or discrepancies that would not show up in a spec without a tolerance reading.  Each speaker has its own frequency response that gives it “color,” a speaker’s individual tonal qualities or sound.  This characteristic determines how the frequency response shapes the sound to make it better by emphasizing certain frequencies.

·         Impedance: Technically, impedance is the opposition to current flow in an alternating current circuit.  Its significance in audio is a way to ensure that components work together correctly; the audio input needs to use up the current from an output circuit.

Typically, input impedances are higher than the output impedance of the circuit they are connected to.  So, a powered speaker system is taking in a line level input and should have high impedance.

·         S/N Ratio: The difference between the nominal program level, or speaker volume, and the noise floor, or underlying hiss and static in an electronic circuit.  The larger the S/N ratio, the better.

Higher S/N will be very noticeable.  The sound will be cleaner with less noise during playback, and when the speakers are turned on but not in use, quieter with less residual hiss.  For critical listeners, this specification should carry some weight.  If your music, or games, have a broad dynamic range (the difference between the loudest and quietest portions of the program), it is very important to insure that the S/N ratio is larger than this dynamic range.  If is not, then either the quietest portion will be lost beneath the hiss of the noise floor, or the loudest portions will overload the speakers and cause clipping and distortion.  Either way, the listening environment will not be satisfactory and deduct from your overall computing experience.

·         Output Power: How much power the amplifier provides to the speaker system.  The amplifier is often integrated into one or all of the speakers in a configuration known as “powered speakers,” as opposed to having an independent amplifier like in a home stereo system.  A larger output power will directly lead to a louder speaker system.  Also, in a multichannel system, the principal stereo speakers should be more powerful than the surround speakers, which are just used for spatial effects.  Finally, in a sub woofer, higher powers will lead to a punchier system with more presence.  When examining speaker’s handling of power, sometimes the speaker’s efficiency is presented.  As a speaker is a transducer, its job is to convert forms of energy, from electrical to physical movement of the air.  The efficiency measure how much energy is actually transformed instead of lost to heat.  Given two equal speakers, one with a higher efficiency will attain higher volume levels at lower power and will perform better than the less efficient speaker.

The trick with output power is how it is measured.  When not indicating the standard used in measuring, any specification talking about output power can be deceitful.  The most common, and fairly honest, standard is to use the RMS (root mean square) testing method.  This technique takes the average of an alternating current signal, such as the power sent to speakers, and gives a specification that would be equivalent to the power dissipated in a comparable direct current circuit.  If a company chooses to instead list peak power, this doesn’t always reveal the full truth of the speaker.  Sound is produced by alternating current with ever-changing levels, and the RMS value gives a much better idea of what the system can really do, instead of the instantaneous peaks in power.  Some other methods, such as describing “music power” or “program power” will vary based on the signal used to create the specification and can prove to be more deceptive, yielding higher ratings than the actual performance in normal listening conditions.

Another specification bundled into output power is the Total Harmonic Distortion, or THD.  This is a measurement of the purity of the audio signal.  Like digital artifacts in imaging, electronic equipment can introduce distortion in the form of harmonics, or frequencies not present in the source, but reproduced as integer multiples of the source frequency (i.e. a 100Hz source produces an output with 100, 200, 200 Hz and so on).  Besides the loss of quality, the human ear is bothered by these imperfections.  The distortion can be presented in either of two ways—by presenting the difference in signal between source and harmonics in dB, or by giving a percentage to indicate the ratio between the harmonics and the original source.  Therefore, the lower of either two numbers, the higher quality of the audio.  Given a typical reading of so and so watts at 0.1% THD, this means that the artifacts are going to be 1/1000^th of the program level, or 30 dB lower.  Translated into speaker pressure, that’s 1/64^th of the program source in perceived loudness.  What does this all mean?  If the program’s dynamic range is greater than 30 dB, then the harmonics are going to be evident in the program as they fall within the dynamic range.  However, the dynamics would be present within the range only under the loudest portions, which would drown them out.  This makes 0.1% THD reasonable, but a lower value would mean a better speaker.

When these specifications are tested, there is a quick way and there is a thorough way; the two don’t go together.  When checking distortion, it is possible to examine just one frequency and how the system reacts to it, or to look at the full spectrum and present the system’s overall distortion.  THD across the entire spectrum will indicate a more accurate picture of the speaker’s quality.  And seeing as sound is much more than just a test tone, this gives a more accurate picture of how well the speakers will actually sound.  Also, computer manufacturers aren’t subjected to the requirement to present all of their specifications from stereo mode, like on home receivers.  So their measuring environment may not resemble an actual listening environment; instead it may be optimized for good specs.

Speaker Formats: From simple to surround sound

The expansion of speaker technology for home theater has trickled down to the computer industry, bringing a variety of new choices for gamers and audiophiles alike.  The advent of surround sound has opened new options for experiencing sound.  Even the concept of adding remote subwoofers has had computer desks shaking for some time, and new virtual positioning technologies have even empowered the older concept of stereo speakers.  When considering what speaker format to buy, it is important to consider how you use your computer.  Each format has its own benefits that must be weighed against their costs; the more fancy the technology, the larger the price tag.

When determining which format to purchase, it is important to examines what your needs in a speakers system are.  By generalizing your needs into one of four categories, it can be easier to classify what speakers are right for you.  Four convenient categories are: basic computer sound, music playback, gaming, and computer based home theatre.  Basic computer sound includes only the alert noises or simple sounds that the computer uses to interact with you, as well as vocal reproduction from webcasts.  A simple sound system is sufficient.  Music playback is the next level of speakers.  At a minimum, stereo should be required.  Adding a subwoofer will extend the frequency response and improve the tonal quality of the sound.  Higher quality speakers will also more accurately reproduce the sound.

Gaming systems and computer based home theatre move into the surround sound environments.  Gaming manufacturers have begun to integrate positional audio into their systems, and when combined with a surround sound system, create an immersive gaming experience.  Computer based home theatre, especially as new DVD drives allow for movie playback at the computing station, offers the opportunity to have a cinema like experience from your computer.  These systems will be able to decode the surround signals embedded in the audio tracks of movies and reproduce these at the desktop.

·         Stereo:  A basic standard, stereo speakers do an excellent job of reproducing music and basic audio at a very affordable price.  Setup is easy— just two speakers to plug in, and sit between them.  The stereo image lends a sense of direction to the sound and allows for basic effects.  For music playback or basic computing, stereo speaker are efficient and well priced.  They are sufficient for many games, though will not provide the most immersive listening experience. It is competent for basic gaming but not well suited to those seeking an in depth experience.

·         2.1 Speaker system: A cute way to say stereo plus a subwoofer, this is a format that expands on the concept of stereo speakers.  It adds a third speaker, usually a remote powered sub woofer that reproduces only the lowest frequencies.  The sub widens the frequency response of the system by providing the lows that small computer speakers can’t provide.  Well suited to music playback, the added frequency response ensures more accurate reproductions of everything from rock to techno to classical.  Also, a 2.1 system provides an entry-level game system that will still shake the table top some with added bass. 

·         Surround sound: Multiple companies have developed methods to create surround sound, or the conceptive of creating an immersive environment of sound that lends both direction and positioning to the audio image.  Dolby* Laboratories has developed many of the various surround technologies used in home theatre that are now making their way into the computer realm.  In addition, Creative Labs has offered its own technology called EAX.  Surround sound technology heightens the gaming environment as designers have grasped this new creative freedom to prey on another of our senses with positional audio.  The introduction of surround sound technology to computers has also continued to merge our home theatre systems with our home computer.  Keep in mind that surround sound can only be produced off of encoded signals; however, any stereo source will play through the front stereo speakers without any problems.  Even for those that don’t choose a surround sound system, any Dolby encoded audio source will always play back over the most basic stereo system without problems, even without a decoder.

o       Dolby Surround:  The first surround technology developed based on commercial theatre system.  It consists of the playback of three channels—stereo left and right and a mono surround channel with a limited frequency response that would feed speakers behind the listener.  This was the first Dolby format available to the consumer that allowed for the playback of Dolby encoding used for cinematic playback years before.  It allows access to three of the four cinematic channels encoded in the stereo signal; it does not produce the center channel for vocal reproduction that is localized to the source.

o       Dolby Surround Prologic: The next step in the Dolby evolution was to access the center channel in cinematic surround sound.  The consumer gained this access with Dolby Prologic, which provided the stereo left and right, mono surrounds, and mono center channel.  The advantage of the center channel is that vocal frequencies are reproduced in a speaker ideally placed at the visual source, localizing the vocals in a movie to the actual faces on the screen.  The Prologic design is based in processing that “steers” the sound to the center channel if it falls within the vocal frequency range and it is present in both the left and right front channels.  However, most computing speaker systems are setup around a near field listening system—that is, the speakers are already close to the listener as well as the visual source.  So the directivity issues that Pro Logic addresses in larger systems with separated speakers is rather not an issue at the computing station.  Though if the computer were going to be used by multiple people by at once, the center channel could be useful as the sweet spot between the speakers can not be shared by all viewers/listeners.

o       Dolby Digital: This technology is centered around the 5.1 Surround standard.  It consists of six discrete channels- stereo left and right front, stereo left and right surrounds, center channel, and the low frequency channel.  The low frequency channel is the “.1” channel-- it requires one tenth of the bandwidth of the other channels because it is only reproducing one tenth of the frequencies of a full range channel.  The other 5 channels are all full range channels that allow for extensive creative use.  They combine to place the sound in a three dimensional environment around the listener.  The low frequency channel feeds a sub woofer that provides the rumble to complete the viewing experience.

o       Dolby Virtual Surround or Dolby Multimedia Surround:  This approach to surround sound is based in psychoacoustical processing.  Basically, the theory is that if an individual with two ears (sound sources) can accurately place where a sound is coming from, than there should be an algorithm that allows two speakers (again, sound sources) to create a three dimensional sound image.  Loading a speaker with directional drivers and shooting sound in more than one direction combined with electronic processing creates the surround image from only two speakers sitting on the desktop. However, the sweet spot, or ideal listening zone, is smaller due to the nature of the processing.

o       A3D:  Aureal’s role in the computer surround sound market comes under their product name A3D.  Based in the processing concepts used in virtual surround sound, A3D serves as a 3D sound solution based in either a two, or with A3D 2.0, a four speaker setup. However, A3D recommends using headphones to take full advantage of the spatialization effects.  This makes sense under the theory of virtual surround sound—two point sources close to the ears should be just liking reproducing what the ears would hear in a 3D environment.  Processing algorithms give the audio a position within a three dimensional listening field by computing how the sound would react to its surroundings.  A3D 1.0 established these hardware based algorithms; the newer 2.0 standard added on some features such as environmental audio, occlusions, and reflections.  Occlusions are how a sound is perceived when passed through a surface (closed window or brick wall) while reflections calculate how the sound is perceived as bouncing off of such surfaces.  The 2.0 standard also includes wavetracing; the graphics data describing the environment (the picture of the walls) is used to realistically calculate how that room would sound.  This data is taken directly from the data sent to the graphics card.  The A3D processing technique is based in the sound card, and requires some processing time, which may result in a performance hit.  However, this internal processing sends out the audio signal to any speakers connected to them, creating the 3D audio environment.

o       Creative Labs EAX:  EAX, or Environmental Audio Extensions, is based on algorithms developed by Creative Labs that have been implemented in their sound cards, speakers systems, and their Personal Digital Entertainment (PDE) products.  It is based in a 4.1 surround sound concept, absent the center channel for vocal range directivity.  It also is based on the concept of environmental audio, or standard processing algorithm designed to make sound characteristic of an environment.  Gaming manufacturers use the capabilities to add environmental audio to their games, such as an underwater or arena effect, through programming control over processing done by the sound card.  The user can also customize and apply these environments.  As a standard designed more for the PC gaming industry, it is geared towards computer multimedia in a near field listening situation.  The speaker format that supports this standard is based in stereo front and rear channels, as well as an added subwoofer.

Let your ears be your guide

When finally sitting down to purchase a speaker system, the previous information can only serve as tools.  The final decision is personal—it is all about what sounds good to the listener.  However, making these decisions is a process that requires careful development.  Some basic guidelines can be extremely useful in deciding which units to buy.

1.      What is needed?

Decide carefully what your needs are as a user.  Is listening to MP3s, webcasts, or CDs your most common activity?  Or are you the ambitious gamer that rides the leading edge of technology in search of the most immersive gaming experience possible? Or finally, is your computer becoming your home theatre system, and you want it to sound like one?  Surround sound carries a price, and stereo can sound great for much less.

Also, examine the possibility of adding a sub woofer.  The smaller speakers used for short range listening often do not carry the frequency response of larger speakers that would be overkill.  However, the frequency range can be extended with the sub woofer that will put some punch back into music and basic games.  Also, the lows can be run at a higher volume without raising the high-pitched frequencies that are more likely to damage hearing.

2.      How do the specifications look?

Carefully examine what the manufacturers say about their own speakers.  Also, look at what they don’t say.  Frequency response can be one of the most deceptive specifications when presented without a tolerance.  Also, the signal to noise (S/N) ratio can prove to be important as it describes how clean the sound will be.  Make informed decisions as to which speakers are worth listening to, but don’t necessarily eliminate all of the others.  Keep an open mind, and open ears.  Remember to get speakers that are magnetically shielded if they will be near the monitor because otherwise they will distort the screen.

3.      Listen, Listen, Listen

Try to find a store where you can actually get your hands on the speakers.  Try and stand (or even better, sit) with the speakers in a position like they would be in when at your computer.  Play your favorite song—one that you know how it should sound.  Then just listen.  Make sure that everything is there that you want to be.  The highs are warm and crisp, the lows are present, and everything seems balanced.  Then listen to another set of speakers and decide if it is better or worse.  It’s worth taking some time because the speakers will be with you for a while.

Good luck in what speakers you decide are best for you.  There are many options out there to choose from.  Look for more articles as we begin to pursue some of the best computer speakers on the market today.