Audio Output:

An emerging technology that has recently demonstrated commercial
viability is that of synthetic audio generation.  Speech
synthesizer devices have the capability to convert ASCII text
files to synthetic speech.  These systems are memory intensive,
requiring up to 10-25K of memory per second of speech.  Though
the technology is becoming more sophisticated, the limited
inflection of the current commercial systems produces poor sound
quality.  

There are ROM based synthetic speech systems that have a limited
vocabulary of several hundred words and require no memory
overhead.  These systems possess instant response capabilities
without requiring any delay for file conversion.  The
programmable ROM environment enables the system operator to
record predefined messages and to modify speech patterns.

There is also technology capable of storing compressed audio on a
single frame of a videodisc.  The system, which accommodates up
to 10 seconds of audio per frame, requires that an audio decoder
device be installed in the computer.  As much as 40 seconds of
audio can be stored in a memory buffer for playback during the
display of a still image frame from the videodisc.  The system is
capable of accommodating a total of 100 hours of sound with 1800
images at a display rate of 20 seconds per image.  

The current limitations of each these audio technologies make a
hybrid system an appropriate strategy for a large information
management application with audio capabilities.  A standard
videodisc soundtrack would be the most cost effective audio
strategy for real time (30 frames per second) motion segments. 
Depending on the system design parameters, digitized sound stored
on a dedicated hard disk would be suitable for audio segments
subject to change, while ROM based synthetic audio would
accommodate repetitive or generic requirements.

<cfile65>   Data Output