Restoring Video Bitstream Integrity by Resolving Field Reversal and Mixed Pulldown Errors

• Published in: IEEE transactions on broadcasting Vol. 59; no. 1; pp. 188 - 196
• Main Authors: Elangovan, P., Harding, P. R. G.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.03.2013; Institute of Electrical and Electronics Engineers
• Subjects: Applied sciences; Correlation; Detection, estimation, filtering, equalization, prediction; Exact sciences and technology; Field reversal; Image processing; Image restoration; Information, signal and communications theory; interlaced; Measurement; mixed pulldown; progressive; Reliability; Signal and communications theory; Signal processing; Signal, noise; Streaming media; Telecommunications and information theory; Video sequences; Performance evaluation; Motion estimation; Image processing; Interlacing; Video signal; Artefact; Frame rate; Algorithm; Implementation; Image sequence; Signal processing; progressive; Field reversal; Real time processing; interlaced; Visual control; mixed pulldown
• ISSN: 0018-9316; 1557-9611
• DOI: 10.1109/TBC.2012.2219213

The quality of the decoded video is affected by errors occurring at various stages of the production chain. In this paper, the errors occurring in the editing layer, due to the coexistence of different video standards in the broadcast market are addressed. The problems investigated are field reversal and mixed pulldown. Field reversal is caused when the interlaced video fields are not shown in the same order as they were captured. This results in a shaky video display, as the fields are not displayed in a chronological order. Whereas, mixed pulldown artefacts occur when the video frame-rate is up-sampled and down-sampled. These errors destroy the integrity of the bitstream and continue to corrupt the videos with which it is edited. The current solution is to visually inspect the video stream. This paper investigates these errors and describes a new technique that eliminates the visual inspection process and automatically restores the bitstream integrity. The methods are based on the principles of motion flow estimation and image correlation. The new algorithms were designed for real-time implementation and outperform existing techniques with superior performance in quality and speed.