Browse Prior Art Database

Sequence Index for Temporal Data Descriptions

IP.com Disclosure Number: IPCOM000119935D
Original Publication Date: 1991-Mar-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 81K

Publishing Venue

IBM

Related People

Bonsall, GW: AUTHOR [+5]

Abstract

In a system for representing digitally-encoded time-based information, such as encoded audio and video, it is desirable that the information be randomly accessible at specific time intervals to provide the processing application the greatest flexibility. For example, digital video information is typically captured at a specific scan rate ('n' frames/second). In order to play back this information in random fashion, the beginning of each frame should be accessible. Digital audio information is also captured at a specific scan rate ('m' samples/second). Typically, these samples are grouped into a collection representing a period of elapsed time (i.e., 1/10 second). Each of these collections should be randomly accessible.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Sequence Index for Temporal Data Descriptions

      In a system for representing digitally-encoded time-based
information, such as encoded audio and video, it is desirable that
the information be randomly accessible at specific time intervals to
provide the processing application the greatest flexibility. For
example, digital video information is typically captured at a
specific scan rate ('n' frames/second). In order to play back this
information in random fashion, the beginning of each frame should be
accessible. Digital audio information is also captured at a specific
scan rate ('m' samples/second). Typically, these samples are grouped
into a collection representing a period of elapsed time (i.e., 1/10
second). Each of these collections should be randomly accessible.

      One system for assuring random access is to assign a fixed-byte
length for a specific time period. That is, for some pre-determined
time period, 'X', a contiguous sequence of 'Y' bytes are reserved.
Therefore, to access the 'i'th interval of time-based information,
the ((i-1)x(Y))+1 byte is accessed. Since this technique uses
fixed-length records for the representation of a specific time
interval:
     Excess storage is required as each time interval need not
require the whole amount set aside by 'Y'. The remainder must be
filled with null data.
     Compression and encoding techniques are restricted to those that
can guarantee that they can produce a digital representation that can
fit into the fixed space 'Y'.

      An alternative approach is to provide variable length records
for the temporal units, with header information in each unit to point
to the next temporal unit. This technique eliminates the deficiency
of the fixed-length scheme, but requires accessing of (i-1) temporal
units to determine the location of the 'i'th temporal unit. This adds
processing time to determine the location of individual units.

      The current invention provides a Collection Index that
accompanies the digitally-encoded time-based data. The Collection
Index is an arra...