• Audio track has high information content in a produced video

• Segment audio based on significant changes without speech recognition

• Audio primitives:

                Speaker changes

                Speech-Music changes

• Index on key sounds, word spotting and speech recognition