An apparatus provides a quick view of security events by transformation of surveillance video streams stored over many hours. Images captured at one or more video surveillance cameras are transformed into a summary of activities by triggering on contents within the views. Selected key images are bookended by anticipatory or consequential frames for better comprehension. Variable masking scopes the date-times of interest to a display user. A frame selection process is triggered by video encoding and compression indicia to eliminate duplication and inactivity. Hours of H.264 encoded video are reduced to seconds of surveillance animation. Motion indicia in video frames and pixel blocks steer the short summary. Timespans and camera views of inactivity are purged from the summary. To meet time management budgets, the report is transmitted to a display/selection apparatus as a mix of still images and video activity glimpses.