While adoption of file-based workflows provided more flexibility with the basic paradigm of file processing, it has also added complexities during media transformations. Improper handling of these complexities can lead to perceived video quality issues for the end consumer. The issues are due to errors caused by media capturing devices, encoding/transcoding devices, editing operations, pre- or post-processing operations, etc. A significant majority of video issues nowadays are due to the loss or alteration in coded or uncoded video information, resulting in the distortion of the spatial and/or temporal characteristics of the video. These distortions in turn manifest themselves as video artefacts, termed hereafter as video dropouts. Detection of such video quality (VQ) issues in the form of dropouts are gaining importance in the workflow quality checking and monitoring space, where the goal is to ensure content integrity, conformance to encoding standards, meta-data fields and most importantly, the perceived quality of the video that is ultimately delivered. This end video quality can certainly be measured and verified using manual checking processes, as was traditionally the case. However, such manual monitoring can be tedious, inconsistent, subjective, and difficult to scale in a media farm.

Automated video quality detection methods are gaining traction over manual inspection as these are more accurate, offer greater consistency, have the ability to handle large amount of video data without loss of accuracy and moreover, can be upgraded easily with changing parameters and standardizations. However, automatic detection of video dropouts is complex and a subject of ongoing research. The source where the artefacts are introduced has a bearing on the way the artefact manifests itself. Automatic detection of the variety of manifestations of video dropouts requires complex algorithmic techniques and is at the heart of a “good QC tool”. This paper discusses various kinds of video dropouts, the source of these errors, and the challenges encountered in detection of these errors.