Deriving the Throughput of Known Multimedia File Types

Deriving the Throughput of Known Multimedia File Types

      Digital multimedia files may contain sound, video, and other
types of flows that have vastly different resource requirements in
terms of their read throughput (or rate).  For example, a typical
rate for a production-level video (PLV) DVI file may be 150 kilobytes
per second, and a 8-bit/22.05 KHz stereo WAV file has a rate of 44
kilobytes per second.  These rates may be calculated from the data
stored within the file.

      The read throughput values must be known by a multimedia server
if the server is to ensure the QoS playback of a file.  However, it
is impractical for a multimedia server to be knowledgeable about the
types of every multimedia file that is stored upon it.  It is much
more efficient for the server to store the QoS parameters for each
file in a uniform manner, such as a set of QoS extended attributes,
regardless of file type.  By doing so, the server complexity is
reduced, as well as the time spent determining the rate of the file
for playback.

      This disclosure describes a utility that is capable of
analyzing different multimedia file types (namely, DVI* (AVS), Indeo*
(AVI), Voice (VOC) and RIFF-WAVE (WAV)), and extracting the rate.
This rate may then be attached to the multimedia file as a "read
throughput" QoS extended attribute.

The throughput analysis tool will support the following multimedia
file types:

     AVS - DVI files (both PLV and real-time video (RTV) files) [1]

     AVI - Indeo, Video for Windows**, and Ultimotion*** files

     VOC - Creative Labs Sound Blaster* voice files [2]

     WAV - Microsoft*...