Decode and stream video frames from file

Use This When

• Testing vision pipelines with recorded video for reproducible experiments
• Processing archived surveillance footage or dashcam recordings offline
• Creating video analysis workflows for batch processing of video datasets
• Developing computer vision models where file-based input simplifies debugging

What It Does

• Decodes video files from configured directory in lexicographic order using FFmpeg and streams VideoFrames at original FPS timing
• Supports optional resizing to target width/height via FFmpeg scaling
• Controls playback with sleep timing to simulate real-time or enable loop playback
• Handles live video path changes for dynamic video source switching

Works Best With

• This component → detect-objects → track-object → visualize for offline analysis
• Video datasets → this component → action recognition or event detection pipelines
• Integration with calculate-optical-flow or pose estimation for motion analysis
• Feeding output-browser-video or output-video-file for processed video generation

Caveats

• High resolution or high FPS video can overwhelm downstream without frame dropping
• Sleep-based timing may drift over long videos; not frame-accurate for precise synchronization
• Mutable video path allows runtime switching but may cause discontinuities in processing
• Repeat mode loops indefinitely; disable for single-pass batch video processing