Methods apparatus and computer programs products for creating a multimedia
presentation having video frames and video segments with synchronized audio are
disclosed. Audio-video data is separated into an audio stream and the video sequence.
The video sequence is divided into video segments, with each video segment comprising
a group of frames. For each video segment an audio significance measure is calculated
using the audio stream. Using the audio significance measure, the video segment
is selected in its entirety or, alternatively, one or more slide frames are extracted
from the corresponding group of frames. The selection of video segments or slide
frames can also be performed on the basis of an additional video significance measure.
The audio stream and selected video segment and slide frames are synchronized,
then synchronously reproduced as the multimedia presentation.