Abstract: A video sequence may include a modality corresponding with an embedded pattern. At least one state machine detects the modality in accordance with difference signals. A signal generator generates the difference signals responsive to decision windows that define regions of interest in the video sequence. The modality may correspond with an embedded film source or other pattern types in the video sequence. Where the state machine detects more than one pattern, a single pattern is selected according to a predetermined priority. The video sequence may contain both static patterns and embedded film source patterns. The state machine discerns the presence of the embedded film source patterns notwithstanding the presence of the static patterns.