M-AUDIO Dman PCI - Manual (Page 25)

Appendix A - About Digital Recording
If your past experience has been limited strictly to analog recording, you're in
for a big treat! If you have digital recording experience already, you can skip
over this section (our feelings won't be hurt).
As taught in high school science class, sound is composed of waves of chang-
ing pressure (level) and frequency (pitch). Analog recording captures these
waves in their entirety and records them as variations in magnetic flux (tape)
or variations in depth (good old-fashioned vinyl). On playback, analog record-
ing adds noise to the recorded sound at relatively high levels (tape has back-
ground hiss, vinyl has surface noise); it therefore causes a low signal-to-noise
(S/N) ratio. Because this noise level is comparatively high, the dynamic range
(the ratio of softness to loudness) of the recording is decreased.
On the other hand, digital recording samples the sound waves (typically 44,100
times per second - i.e. a rate of 44.1 kHz) and records the sounds as digital
data. This numeric data can be stored on a hard or floppy disk, DAT tape (in
some software packages), CD, or any other common data storage media.
During digital playback, no media noise is added by the recording medium.
This results in a much greater S/N ratio and greatly increased dynamic range
over traditional analog methods. The end result is cleaner, quieter recordings.
The only noise present in the digital realm is any noise introduced by the A/D
or D/A converter and this varies according to the quality of the converter and
the PC board design. Using quality components and design practices as in the
DMAN PCI, the only way to make a noisy digital recording is to manually add
the noise yourself!
DMAN PCI supports recording and playback in many audio formats. When
selecting an audio format, you are usually trading off sound quality versus the
amount of disk space the recorded audio will consume. In general, sound
quality increases as the Sample Rate increases, as the number of bits used
increases, as stereo is used instead of mono, and when the audio is not com-
pressed (i.e., its format is PCM not ADPCM or A-Law). Sample Rate is the
parameter most directly related to frequency response. For CD audio, 44.1kHz
is the standard, and should be chosen for most audio applications. Choose a
sample rate that is at least twice the frequency (this is called the Nyquist value)
of the highest audio frequencies you want to record. For example, if your
source material has a frequency range of 30 Hz to 10 kHz, such as voiceovers,
and its appropriate for your application, you can record at a sample rate of 22
kHz. Also, if storage space is an issue, this rate will take up half of the disk
space that 44.1kHz will (plus, voiceovers can be done in mono). You can see
from this example that 44.1kHz has been chosen for CD audio because it is a
little more than twice the highest frequency in the range of human hearing.
Bits determine the resolution of the data. The more bits, the higher the resolu-
25