How Sound Cards Work

The sound you hear all around you is produced when a physical object-such as the strings of a guitar or violin, the brass surface of a bell, or the vocal cords of a bird or a human being-moves within a range of frequencies between about 15 and 18,000 times per second (15 Hertz–18 kilohertz). This vibration causes the surrounding air to vibrate, which transmits the sound to the membranes in your eardrums and causes them to vibrate at the same frequencies.

This is called analog sound, because your eardrum vibrates at the same rate and volume as the object that produced the original sound. Analog sound travels to your ear or through an electronic circuit in a continuous stream.

In a computer, recorded sound is digital, because that continuous stream of vibrations has been broken down into thousands of separate numeric values every second. Digital sound must be converted back to analog and played through a speaker, earphones or some other similar device before you can hear it.

A computer's sound controller performs these activities:

• It receives digital data from the CPU and converts it into an analog signal.

• It amplifies the analog signal and sends it to two or more speakers.

• It receives analog sound directly from a microphone or from some other source and amplifies it if necessary.

• It converts the analog signal to digital form and sends the digital data to the CPU.

Some sound controllers use separate analog-to-digital converters (ADCs) and digital-to-analog converters (DACs). Lower-quality controllers perform both kinds of conversion in a single circuit called a CODEC (COder/DECoder).

In order to convert analog sound to digital form, the processor on the sound card slices the analog audio stream into tiny pieces, and it measures the volume and frequencies of the sound in each piece. The quality of a digital recording depends on two elements: the number of slices per second (the sampling rate or sample rate) and the number of digital bits the processor uses to store each sample (the bit depth).

A sound controller can play existing files, and it can also instruct the CPU to create new sound recordings and store them on the computer's disk drive.

The combined sampling rate and bit depth is often described as the bit rate of a recording; the quality of the recording improves as the sampling rate and bit depth increase. For example, the bit rate of a commercial audio CD is 16 bits and 44,100 samples per second, or 16-bits/44.1 kHz, but many recording studios and sound archivists create and store their digital master recordings at 24-bits/96 kHz, or even 24-bits/192 kHz. Of course, there's definitely a point of diminishing returns where most listeners can't hear any improvement, but a higher bit rate also allows an audio engineer to perform more precise processing on a recording to reduce noise or add special effects.

If you're acquainted with the way your computer handles video, the rest of a sound controller's operation might sound familiar: a specialized processor called the digital signal processor (DSP) controls the conversions between analog and digital sound, and exchanges instructions with the computer's central processor. The CPU treats the sound controller as one more source and destination for data to and from active programs.

CROSS-REF 

See Chapter 10 for more information about understanding video in a computer.

Some programs, such as Windows Media Player and RealPlayer, act as readers for sound files and incoming data streams. When you want to listen to an audio file stored on your disk drive, a CD, or a streaming Internet service, one of these programs reads the data in the file or stream and sends it to the sound card. Other programs can send sounds directly to the sound controller, without the need for a separate program in between. For example, when America Online produces its well-known "You've got mail!" message, the audio goes directly to the sound controller.