Following a platform studies approach to video games, this article explores Deep Learning Super Sampling (DLSS), a cluster of machine learning-based upscaling and anti-aliasing techniques designed to make graphically intensive games run faster and in higher resolutions. Its inclusion as an optional setting in video games influences how players sense time.
Following a platform studies approach to video games, this article explores Deep Learning Super Sampling (DLSS), a cluster of machine learning-based upscaling and anti-aliasing techniques designed to make graphically intensive games run faster and in higher resolutions. Its inclusion as an optional setting in video games influences how players sense time, in terms of increased frame rates and temporal artefacts, and the visual instabilities, flickerings and ‘ghosted’ images that this technology creates. To offer a closer look at the visual manifestations of temporal manipulation at play in contemporary upscaling technologies, this article focuses on the video game Control (Remedy Entertainment 2019). While Control is not mechanically concerned with time management or manipulation, this article argues that the game thematically addresses the takeover of futures by medial pasts in a way that rhymes with the temporal manipulation inherent in DLSS’s use of machine learning models.