Most videos that end up being broadcast, or becomes viral, contain, in a single image frame, both perpetrator and victim. This is the case in most videos of police brutality since the video recording of Rodney King, the black motorist beaten up by the LAPD, emerged in 1991. But for every shot that includes a beater and a beaten or a shooter and a shot person, there are many more that include only one or the other, or just audio, or things that happened before and after the incident. Their relation to other images and the main incident is not obvious. We are less apt at viewing and understanding incidents and processes that slip between images. These images are therefore often regarded as trash. searching through this image flotsam in areas of conflict we pick these up, synchronise and re-assemble them with other images to virtually (sometimes physically) re-enact incidents in space. Virtual models are useful to locate and trace the movement of each camera’s cone of vision in time and space. Whereas recent debates in the field of photography and visual cultures were concerned with the spectators’ relation to single images and photojournalistic trophy shots, with questions regarding the image’s ability to capture the pain of others; today, the sheer number of images and videos generated around incidents means that to view images is to construct and compose. The ‘the architectural-image complex’ is a dynamic time/space relation between images that replaces the thematic classification systems of traditional archives. Transition between images is not based on the dialectics of montage but in spatial navigation within an image-space - a space that is at once virtual and photographic.