|
|
|
| Reference Series |
Table of Contents For This Issue
|
|
| |
| How Computers Work, Part I | |
|
August 2001• Vol.5 Issue 3 Page(s) 42-47 in print issue | |
Tuning In To Audio Learning How Sound Works In Your PC |
If you don't recognize terms like polyphony, signal-to-noise ratio, and wavetable synthesis, then you must not be very familiar with sound card technology. A good sound card can make your monitor shake while you're playing some of your favorite tunes or blasting
bad guys in your favorite game. But few people know what's really going on when a sound card does its thing. Sound cards are more powerful than they have ever been. They can process 3-D audio and all kinds of amazing sound effects. Sound cards can do this because most new cards have a DSP (digital signal processor) built into them. The digital signal processor is like a mini CPU. It processes information and gives the sound card instructions. Your computer’s CPU could handle these tasks, but a digital signal processor frees up your computer’s CPU to handle other things. The signal-to-noise ratio is perhaps the most important specification to look for when buying a sound card. The signal-to-noise ratio refers to the cleanliness of the sound. The higher the signal to noise ratio, the less background noise will hiss out of your speakers. Most sound cards have a signal-to-noise ratio around 85 or 90dB (decibels). However, you can find a sound card with a signal-to-noise ratio as high as 97dB. A few years ago, sound cards used FM Synthesis to create sounds. Instead of using actual recordings of real sounds and instruments, the sound card would try to approximate a sound using mathematical formulas. The result was not very good; the audio sounded like it was generated by a computer. That’s when wavetable synthesis entered the equation. Wavetable synthesis was a big step up in terms of sound quality. Sound cards with wavetable synthesis used samples of real sounds instead of trying to create them on their own. The results were much better. The sound card could now create much more realistic audio.
Your computer’s sound card is equipped to play hundreds of sounds right out of the box, but your sound card is not limited to that large orchestra of sound effects if it supports DLS (downloadable sounds). DLS isn’t something that home users purposely use. Rather, it allows software designers to customize their own sounds and write them into the software they create. Your sound card can recognize and use the original sounds, thanks to the DLS standard. DLS makes a sound card very versatile. Adding to the versatility of sound cards are APIs (Application Programming Interfaces). APIs are standards that allow software designers to affect sounds in different ways. For example, they can create three-dimensional effects or increase the reverb of a particular sound. Microsoft’s audio APIs, DirectSound and DirectSound 3D, are by far the most popular. Any sound card on the market should, at the bare minimum, support the Microsoft DirectSound standard. There are a couple of other APIs that have made a splash in the world of PC audio, but we’ll talk more about these later. Another term you may come across with sound cards is Full Duplex. This means that you can play and record sounds at the same time. Think of a video conferencing session. You can talk into the mic and hear what someone else is saying at the same time. The sound card will process both signals simultaneously. Nearly every sound card made today should be a full duplex sound card.  How A Sound Card Works.
Before a sound card can start beeping and serenading you, it needs to be told that it’s supposed to make a sound. An application usually makes the request, although it can come from other elements. For example, when you maximize a window or open a dialog
box, your computer will tell the sound card to play the sound associated with that event. According to Dave Sparks, Systems Architect for Creative Labs, sound cards create sound by going through the following steps. The request is made through one of the several APIs we discussed earlier. The request is passed through the API to the sound card. The request tells the sound card where to find the sound in the computer’s system memory, or RAM. The sound card finds the sound and applies any effects or other elements that it’s supposed to add to the sound. Once the sound has been mixed, it goes to the DAC (digital to analog converter). The DAC converts the sound to analog signals. The sound card then ships the sound to the speakers or headphones. The sound has to pass through the DAC, because prior to that, the sound is merely a collection of digital ones and zeros, called bits. If the sound requires effects to be added, the digital signal processor kicks in. Sparks says the effects can range from 3-D spatialization to interesting effects that “can make it seem as if you are inside a cave, a heavily carpeted hallway, or a large stadium.” 3-D sound works by tricking your brain into thinking that sounds are coming from several different directions. According to Sparks, one way to create this illusion is to create a set of virtual speakers by manipulating the sounds coming from a two speaker set or a pair of headphones. Virtual speakers can be created by using HRTF (Head Related Transfer Function technology). HRTFs make you think sound is coming at you from different directions by using a couple of techniques. Part of the trickery involves the direct path sound takes to your ears. Let’s say a sound comes from your left. According to Sparks, “you get a direct path into your left ear, but the right side is in the shadow, so that affects how you perceive different frequencies.” Sparks says that because the sound is coming from the left, you’ll hear less of the higher frequencies in your right ear, which helps you locate the source of the sound. Another HRTF technique is to cause a short delay in the sounds coming from each speaker. If a sound comes from your left, it takes the sound longer to reach your right ear than your left ear. And the volume makes a difference too, Sparks says. “It’s going to be softer, because it’s being blocked by the head.” Another method for creating 3-D sound, according to Seth Dotterer, Director of Marketing for Voyetra Turtle Beach, is to use a four-speaker configuration. Dotterer says it’s easy to create the illusion of a waterfall splashing behind you while someone talks in front of you. “The sound card splits the signal. It sends the first stream to the front with a slight pan to the back. And it sends the waterfall to the back with a slight pan to the front.” Of course, Dotterer recognizes that a four-speaker setup isn’t practical for everyone, and agrees that using HRTFs are also a good way to create the illusion of 3-D sound. Both Dotterer and Sparks agree that the current quality of 3-D audio is pretty good but not perfect. Sparks gives an example. “If I turn my head ever slightly…I disrupt the impression, because my expectation is I’m going to hear a slight change in the sound when I turn my head.” However, the computer has no way of tracking your head’s movements and therefore can’t adjust the sound accordingly. Dotterer says, “I think there will be some advances in the near future to sort of tweak out an individual’s usage. . . . But for a general standard right now, it’s very good.” Historical Development.
Early sound cards, as one might expect, didn’t exactly deliver the type of quality sound that would blow your hair back. They were only 8-bit cards that sounded more like an AM radio at best, according to Voyetra’s Dotterer. But it didn’t take long for sound
card manufacturers to improve the sound quality of their cards. For example, Turtle Beach released the Multisound sound card in 1992, which, according to Dotterer, was the first 16-bit sound card. He says this “moved sound into the CD quality realm,
so 1992 was really the start of serious sound.”Creative Labs released its first Sound Blaster sound card in November of 1989. Although not the first add-in audio card for PCs, it is a significant milestone. The Sound Blaster line of products popularized audio for the PC. A year and a half later in May of 1991, Creative Labs released the Sound Blaster Pro, which became the industry stereo standard for multimedia PCs. The Sound Blaster standard was by far the most commonly supported audio industry standard for many years. The main prerequisite when shopping for sound cards back then was Sound Blaster compatibility. However, the Sound Blaster standard would not dominate the PC audio scene forever. When Microsoft developed Windows 95, it made a strong effort to create universal multimedia standards for software developers. Computer game designers, usually the first software designers to adopt new multimedia technologies, accepted the new standard. Sound card manufacturers did too. Thus, the DirectX API was born. The big advantage of Windows 95 and DirectX was that it made installing and configuring multimedia devices, and the software that used them, easier than ever. Because of the DirectX standard, computer users no longer had to deal with IRQs (interrupt request lines) and DMA (direct memory access) settings. Software and hardware makers had a common standard that took advantage of the new Windows 95 operating system. In other words, there was much rejoicing. There was also much to celebrate as wavetable synthesis began to replace FM synthesis in sound cards. As we mentioned earlier, wavetable synthesis uses recordings of real sounds, a big improvement over the old FM Synthesis standards. According to Creative Lab’s Sparks, FM synthesis was still more common than wavetable synthesis as recently as two years ago. However, that has changed significantly. It would be very difficult to find a new sound card today that used FM Synthesis rather than wavetable synthesis. The release of Windows 95, the implementation of wavetable synthesis, and the development of DirectX, gave multimedia computing a big boost. In fact, multimedia applications looked and sounded better than they ever had in the past at a consumer level. However, sound cards continued to improve. The next major change in sound card technology occurred when manufacturers began selling PCI (Peripheral Component Interconnect) sound cards, as opposed to ISA (Industry Standard Architecture) sound cards. This transition began to occur about two years ago. PCI sound cards have several advantages over their ISA counterparts. PCI sound cards can transfer data over the PCI bus several times faster than ISA sound cards can transfer data through an ISA bus. Furthermore, PCI cards can use the system’s memory resources, whereas ISA cards had extra memory built into them. Because PCI cards use the computer’s memory, manufacturers don’t have to build that memory into the sound cards, which reduces manufacturing costs. This has helped to keep the cost of relatively high-quality sound cards at a low price. DSPs also began to appear on sound cards around this time. With a DSP on board, sound cards had the power to create 3-D sound and apply an impressive arsenal of effects to sounds. This was accomplished without using the computer’s CPU. This gave software designers the chance to strut their stuff without overtaxing a computer’s resources. Current Status.
Today’s sound cards bear the fruit of labors past. The use of audio in PC computing continues to grow rapidly. Computer audio was once confined almost entirely to games and a few reference titles that hyped their multimedia content. Audio is used in many other ways today. Voice recognition software is reaching the point where it is becoming more functional and practical, and voice dictation is reaching into the business side of PC audio. Let’s also not forget a sound card’s musical applications. With the proper sound card and a good pair of speakers, your PC audio system could rival your home stereo. Many sound cards now support four or more speakers so you can enjoy Surround Sound or even Dolby Digital sound through your PC. This extra audio muscle can take PC games to a new level, and audio files will sound much better as well. Also, DVD drives have become very common on PCs these days, and DVD movies practically beg to be played through at least four speakers. But you can do more with your sound card than listen to your personal computer. You can use it to create your own audio files. Sound cards usually come with software utilities with which you can record and edit your own music. These software utilities are primarily used to create MP3 (MPEG Audio Layer 3) files. MP3 has been a very popular format for audio files the past few years. The amount of traffic that MP3-sharing sites like Napster have generated is testament to that. MP3 is popular because it lets users record CD quality audio without eating up a lot of valuable disk space. You can also download MP3 files from several sites on the Internet. You can turn your PCinto a jukebox by storing your entire CD and MP3 collection on your computer. Of course, games are still the arena in which PC audio shines brightest. There’s little need for cool 3-D sound or special audio effects in business applications. Put that power to good use in a game, however, and the results can be more than music to your ears. 3-D audio relies heavily on APIs and a DSP. We briefly mentioned the development of Microsoft DirectX (http://www.microsoft.com/directx) APIs as a turning point in the advancement of sound cards. Specifically, Microsoft created the DirectSound and DirectSound 3D standards for PC audio. These two standards are by far the most commonly found APIs in sound cards. But there are a couple of other standards that you’ll often find in sound cards.
Creative Labs (http://www.creativelabs.com), a company that has been very influential in the realm of PC audio over the years, also has its own proprietary API known as EAX (Environmental Audio eXtension). This API takes a different approach to rival A3D. EAX places its emphasis on sound effects, much like a Hollywood studio. For example, you can modify audio so that environmental sounds have the effect of emanating from a cave or carpeted hallway. Of course, EAX can also play 3-D audio. The DSP is the essential element that makes these APIs work. The two DSPs most commonly used now are the SoundFusion DSP from Cirrus Logic and the EMU10K1 processor from Creative Labs. Several vendors use the Cirrus Logic DSP in their sound cards, including Voyetra Turtle Beach and Hercules. But you'll find Creative Lab's EMU10K1 processor only on Creative Labs sound cards. When Creative Labs initially released the EAX API, only Creative Labs sound cards could use it. Today, there are several sound cards from other vendors that have EAX support.
An Ear Toward The Future.
Sound cards have improved so much over the years that it seems like there’s little room left for them to grow. Sound cards already process sound very rapidly and are able to play an incredible chorus of sounds at one time. Sound cards can play audio at CD
quality, and higher signal-to-noise ratios mean that the sounds you record will be crystal clear. It seems the main area for improvement in sound cards will be in the field of 3-D audio. Although 3-D audio is at a pretty impressive stage, more can be done to enhance its realism. Creative Labs’ Sparks says that work is being done on head tracking devices that will compensate for a person’s head movement and maintain the illusion of 3-D audio throughout. (Creative Labs also sells its own line of computer speakers.) But, he says “there are latencies involved in doing that, and basically it’s not fast enough now.” One possible threat to the future of sound card technology is the development of USB (universal serial bus) speakers. USB speakers can process all the digital audio information internally, so a sound card isn't necessary when using a set of USB speakers. USB speakers haven't been very popular with consumers to this point. Early USB speakers were fraught with problems, including difficulty in setting up the speakers and interrupted audio because the USB bus was overtaxed by other USB peripherals. There is a chance that the release of USB 2.0 will once again generate more interest in USB speaker technology. And there are still some speaker vendors selling improved USB speaker systems, including Altec Lansing. Sparks isn’t exactly worried about USB speakers replacing sound cards, however. He says, “USB adds overhead to the host processor . . . and if you have any USB peripherals on it that have high data rates, for example a joystick, they tend to interrupt the audio.” Of course, Sparks works for one of the leading manufacturers of sound cards, so we might expect him to have a poor outlook on USB speakers. In the end, however, we tend to agree with him. The data transfer rate of PCI is 133 megabytes per second (MBps), whereas USB only pokes along at 12Mbps. Clearly, PCI sound cards can transfer more data more quickly over a PCI bus than USB speakers can over a USB bus. And the point Sparks makes about using USB peripherals while using another USB device is well taken. In our experience, USB speakers are merely a novelty right now. They are not a serious innovation that will change the face of PC audio today. USB speakers have a long way to go in their quest to replace sound cards. PC audio is too important these days to be taken for granted. With that being the case, it’s certainly helpful to know more about this technology. Understanding sound cards means you can learn how to get more out of yours.  by Michael Sweet View the graphics that accompany this article. (NOTE: These pages are PDF (Portable Document Format) files. You will need Adobe Acrobat Reader to view these pages. Download Adobe Acrobat Reader)
|
