Introduction - Computer Animation: Algorithms and Techniques - page 22

form of “note x on” and “note x off” where x is any of the standard musical notes. There are also control
commands to set pitch, loudness, etc. Some devices can also send out MIDI commands as they are
played in order to record the performance as keys are depressed and released. MIDI supports up to
16 channels and devices can be daisy-chained so that, for example, a device can be set up to respond
only to a particular track.
Digital audio sampling
Sounds are pressure waves of air. In audio recording, the pressure waves are converted to some rep-
resentation of the waveform. When it is recorded digitally, the wave is sampled at certain intervals—
the sampling rate—with a certain number of bits per sample—the sample size—using a certain number
of tracks. The sampling rate determines the highest frequency, called the Nyquist frequency, which can
accurately be reconstructed. A voice signal can be sampled at a much lower rate than CD-quality music
because the highest frequency of a voice signal is much lower than the highest frequency that might
occur in a piece of music. The number of bits per sample determines how much distortion there is in the
recorded signal. The number of tracks is how many independent signals comprise the music—one for
mono, two for stereo, more for various “surround sound” recordings or for later editing. A voice signal
and AM radio are sampled at approximately 10 K samples per second with 8-bits per sample using one
track. CD-quality music is sampled at 44.1 K samples per second with 16-bits per sample using
two tracks.
Similar to digital imagery, the digital recording can then be compressed for more efficient storage
and transmission. The compression can either be lossless (the original signal can be reconstructed
exactly) or lossy (some of the original signal is lost in the compression/decompression procedure—
usually the very high frequencies). General data compression techniques can be used, but don't do
as well on audio data as the compression schemes that use the fact that the file contains audio, referred
to as being perceptually based . Compression of speech can use techniques that are less sensitive to
preserving all of the frequency components of the sound. Common audio data compression techniques
include MP3, MPEG-4 ALS, TTA, and FLAC.
1.5 A brief history of computer animation
1.5.1 Early activity (pre-1980)
The earliest computer animation of the late 1960s and early 1970s was produced by a mix of researchers
in university labs and individual visionary artists [ 24 ][ 25 ][ 37 ]. At the time, raster displays driven by
frame buffers were just being developed and digital output to television was still in the experimental
stage. The displays in use were primarily storage tubes and refresh vector displays. Storage tubes retain
an image indefinitely because of internal circuitry that continuously streams electrons to the display.
However, because the image cannot be easily modified, storage tubes were used mainly to draw com-
plex static models. Vector (calligraphic) displays use a display list of line- and arc-drawing instructions
that an internal processor uses to repeatedly draw an image that would otherwise quickly fade on the
screen. Vector displays can draw moving images by carefully changing the display list between
refreshes. These displays were popular for interactive design tasks.
During this time period, static images were often recorded onto film by placing a camera in front of
the display and recording an image of vectors on the screen. In this way, shaded images could be