<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Within an hour-long TV episode, viewers will typically experience somewhere between 600 and 900 edit points. This means that viewers will experience a unique composition featuring a unique combination of spatial phenomena 600 to 900 times. On top of this, these shots must lead into one another, allowing for changes from the start of a shot to the end of the shot and confronting the “temporal contrast” that comes from having edit points at all. During any season of a show’s thousands of shots, production encounters specific technical imaging and perceptual failures and finds ways to work around them to get the image “good enough” for a general audience. However, these images remain valuable even after production is complete, and in fact the archive of ten episodes per season or more represents a treasure trove of naturally occurring perceptual phenomena. This paper presents guidelines and examples for harvesting challenging images from real production and ways in which these images can be analyzed to further a better end-user experience</i> .