Various methods, for example image recognition, speech-to-text algorithms, and natural language processing, can be used to capture location references in linear and non-linear media content. The methods differ in terms of the technologies, procedures or media, such as the audio track or video images. Our investigation, which is based on the metadata of a video, reveals that georeferencing media content is possible, and that, using examples from the ARD Mediathek, the genre and thus the content of a video can influence the results.