A multi-media search and indexing system for automatic selection of scenes
or sounds recorded in media for replay in other context by analyzing the
media for a set of frequency ranges of interest for replay. The media is
analyzed to select desired frequency ranges of interest within the media,
the selected frequency ranges being indicative of a scene or sound of
interest. The granularity or length of the selected frequency ranges is
selected for analysis. A number of analysis intervals are determined in
the selected frequency ranges. The media is filtered for the selected
frequency ranges and the audio level is measured for each interval. An
audio clip level is established for each frequency range as indicative of
a scene or sound of interest. The interval frequency range and audio level
are recorded for subsequent replay.