Audio Fundamentals.
We’ve delved into audio principles before, but lest newcomers feel lost, let’s take a quick reprise through the key points. These are the concepts you’ll need to understand when approaching audio and explaining it to your customers.

Pitch and frequency.
When you pluck a guitar string, the string quickly swings back and forth. The rate at which this happens is known as the pitch, and pitch is measured in Hertz (Hz), meaning the number of complete swings back and forth per second. Frequency is the number of Hz, and human hearing generally spans from about 16Hz to 20kHz. Musical instruments and other sound sources are able to create sounds beyond this spectrum, only we can’t hear them. One of the ways in which audio compression schemes reduce file size is to eliminate this inaudible material.

DAC, ADC, and codecs.
Digital-to-analog conversion (DAC) and analog-to-digital conversion (ADC) are essential parts of PC-based audio. All sound is inherently analog. It has to be, right? A string of 0s and 1s hitting your eardrum doesn’t do much. Similarly, PC software can’t make much sense out of pressure waves in the air. Even “digital speakers” preserve the digital signal stream up until the data hits their onboard processor, whereupon the information is converted to analog for reproduction in the speaker cones.
“The audio codec provides DAC and ADC functions,” explains Keith Kowal, audio marketing manager for VIA, “so it can take a digital audio stream and convert it to analog and then also convert analog back to digital. This is important because it determines the quality of the audio you hear. How cleanly that conversion takes place impacts your quality, whether for recording or playback.”
There are many codec architectures on the market. By far the most common is Audio Codec ‘97 (AC’97), although there are different implementations of AC’97 as well as alternate codecs. AC’97 is an Intel specification that actually bridges a codec with the AC-Link control found within the southbridge. The codec could be within a chip on the motherboard (VIA and Realtek are prominent vendors of such chips) or just software performing the same function, although these tend to produce lower quality results.

Sampling.
If you were to draw a waveform on an X-Y axis, that would be analog. In the world of digital data, though, there are no smooth, flowing lines. Instead, you get to plot dots. Now imagine that you have a sound wave that lasts for five seconds. Four times per second, you are able to take a “snapshot” of that wave’s current state as measured in frequency and amplitude (volume). Plotted on your paper, the result would look like rising and falling stairs. Think of this as 16-color video on your monitor. You can probably make out the general shape, but the image looks terrible. However, the more times per second that you “sample” that source audio, the closer your graph comes to looking like a smooth waveform and thus the better the digital sound will be. Part of the ADC process involves taking those digital snapshots of an analog source.
There are two components of sampling, the sampling rate and the bitrate. The frequency refers to how many samples are being taken per second. The general rule is that to avoid noise you sample at the highest frequency in the audio spectrum of the source material plus 10 percent. Since human hearing tops out at 20kHz, that yields a 44.1kHz sampling rate, which is the standard used in CD-Audio.
The bitrate refers to the amount of data captured during each snapshot. As you’d expect, more data per sample yields finer quality. Early sound cards used 8-bit sampling. Starting with the Sound Blaster 16 and extending to today, the norm has been 16-bit sampling, which again is used with CD-Audio. In the last couple of years, the market has started migrating toward 24-bit, which is the standard used in DVD-Audio.

Wattage, RMS, and THD.
Remember how with hard drives you have burst data rates and sustained transfer rates? Nobody really cares about the burst rate because that’s not how the device gets used in the real world. The same idea applies in speakers. Common knowledge holds that more wattage (power) means more volume. But peak wattage is sort of a useless spec. You can scream at the loudest level your voice can manage, but you won’t be able to keep it up for long. When it comes to speakers, you want the RMS, or root mean square, measurement of sustained power over time. Now, as you would expect, the amount of distortion you will hear from speakers (there is always some, whether you can hear it or not) is proportionate to the amount of power being used. This is why you should look for total harmonic distortion (THD) ratings given relative to frequency range and RMS power. The FTC requires this information, although very few vendors will supply it in their product information. A THD of .1% is acceptable to most ears. Be sure to look for THD ratings in sound cards, too.

SNR.
Signal-to-noise ratio is a measurement of audio signal strength relative to background hiss as measured in decibels. Higher numbers are better because you want maximum volume without a detectable amount of noise. Most people in regular use environments have a hard time telling the difference in SNR ratings past about 100 dB. The best audio cards deliver performance in this range, with Creative’s Audigy 2 ZS currently holding the crown at 108 dB.

3D audio, multichannel, and matrixing.
The idea of “3D” audio is a bit misleading as there are currently no models in which sound is designed to originate from above or below the listener. (Yet. We’d bet money that with lateral schemes now approaching the point of saturation, the Z-dimension will be next up—literally.) A better term is “positional” audio, which refers to manipulating sound characteristics in order to simulate sound coming from any point on the lateral plane surrounding the user. The amount of computation required to pull this off convincingly is formidable, and unless your sound adapter performs hardware processing through the use of a dedicated audio processor at least a 1GHz processor is recommended for the software load. Also, while there have been many APIs that sought to perform the task, Creative’s proprietary family of EAX standards, DirectSound3D, and Sensaura’s 3DPA have emerged as today’s chief rivals. To support the widest possible base of titles emerging from developers, the sound adapters you sell should support both platforms.
Creating realistic positional audio also hinges upon the number of speakers used. Obviously, you can’t reproduce sounds coming from behind the listener without rear speakers. (Altec Lansing and others have attempted to do so, however, through the use of pscyhoacoustic technologies that bounce sound signals off of the walls, but these have generally been panned in the market. However, one interesting newcomer is Nirotek (www.niro1.com), a related company to the venerable Nakamichi, which has a 1.1 surround system for home theaters. A 1.1 desktop system is due shortly for desktop PCs, and we expect to have a review sample written up in the Easy Upsell column during the next quarter.) Dolby helped make 5.1 (four satellites, one front-center channel, and one sub-woofer) the standard in modern surround sound. Most DVD movies are now coded in Dolby Digital 5.1, as are an increasing number of PC games. CD-Audio and all MP3s are strictly two-channel stereo, but DVD-Audio is 5.1, Dolby Digital EX is 6.1 (the second optional rear speaker is merely a mirror of the first rear), and WMA (Windows Media Audio) version 9 will even handle 7.1.
In cases where the source audio contains fewer channels than the number of speakers being used, the content can be matrixed, or “upmixed,” to fill in the channels. Matrixing is a huge deal for vendors like Creative Labs who want to turn users’ libraries of stereo content into a surround sound experience. A year ago, matrixing software was a relatively expensive affair. Now, though, it is becoming a checklist item available to mid-range and even low-end users. As we’ll discuss later, though, the quality achieved by matrixing is debatable.

Normally, we’re not big fans of licensing programs. After all, what’s in a name? In the case of THX, though, the answer is “quite a bit.” THX first appeared in 1983 with “Return of the Jedi.” The effort grew from a frustration in George Lucas when he discovered that the sound and visual experience being reproduced in theaters did not match what he’d crafted in the studio. THX emerged as a technology program designed to ensure that the content he made would be absorbed by the public exactly as he’d intended. Soon enough, other content developers started adopting his methods and eventually the THX certification program was born. Not only does THX apply to movie theaters and home theaters but also DVDs, multimedia products, and even car audio.
“Before THX,” says Mark Tuffy, director of advanced technology for THX Ltd., “what we found in talking to a lot of game developers was that because the content was going back and forth between studios, they wouldn’t be seeing or hearing the content correctly. For example, one of Microsoft’s biggest beefs at the GDC conference last year was that developers and publishers didn’t use the full video bandwidth of the Xbox. In other words, they didn’t use the blackest black and whitest whites. It all got compressed in the middle. So you got games that didn’t look as lifelike as they could be. Now, people are talking about lifelike the visuals are in THX games. That’s because our program is based around showing people exactly what your content looks like. You’re providing people with a clean palette on which to paint the picture. Before, the palette might have been cream instead of white and a bit dark. Suddenly, you can see and hear all this detail you were missing. If you’re working in an environment that has defects, you’re continually trying to adapt to those defects, so the content is always being adversely manipulated.”

THX certification doesn’t necessarily aim to replicate a cinema experience. Rather, it examines the type of content that is applicable to a given device and its environment. For instance, it would make no sense to certify a 2.1 speaker system for playback of 5.1 content. When a 2.1 speaker system is THX certified, it is tested to see if it meets a barrage of quality requirements pertinent to content such as high-bitrate stereo music.
“We look at the real-life performance of the kind of material you’re going to be using,” says Tuffy. “I would not say to anyone that you can take a set of our partners’ PC speakers and replace a THX home theater with them. But I would say that given the price point, if you were to play back a DVD over that audio system, you get a good percentage of a THX system for a fraction of the cost.”
Three of the four speaker systems picked as the industry’s best by PC Magazine last year were THX-certified. (The fourth company is currently in partnership talks with THX.) Every exceptional speaker system we’ve enjoyed enough to want on our own multimedia systems in the last two years has either been THX-certified or made by a company with THX-certified units in its lineup. After all, once a company masters the art of making THX-class products, these lessons can be applied to other products without sending them through the certification process.

For resellers, the advantage of a certified product is in the assurance of quality that goes with it. Sure, we’ve seen FCC- or WiFi-certified products that lacked proper shielding or receptivity. But we’ve never heard a set of THX speakers in any environment that didn’t deliver a best-in-class experience, although there are varying degrees of excellence within THX-certified products.
“The whole point,” says Tuffy, “is for products to meet or exceed our quality requirements. We’re not trying to have 15 sets of speakers that sound exactly the same.”
While certified models will inevitably carry slightly higher price points, resellers shouldn’t underestimate or ignore the potential of the THX brand to sell itself.