Typical application scenarios in the area of rich-media management, such as the continuous digitisation of the media production processes, the search and retrieval tasks in a growing amount of information stored in professional and semi-professional audio-visual archives, as well as the availability of easy-to-use hard- and software tools for the production of rich-media material in the consumer area, lead to an increasing demand for a meaning-based management of digital audio-visual assets. This “semantic turn” in rich-media analysis requires a semantic enrichment of content along the digital content life cycle and value chain: The semantic enrichment of content can be achieved manually (which is expensive) or automatically (which is error-prone). In particular, automatic semantic enrichment must be aware of the gap between meaning that is directly retrievable from the content and meaning that can be inferred within a given interpretative context. Each solution has its benefits and drawbacks. Our paper discusses the relevance of semantic analysis of rich-media in certain application scenarios, compares two possible approaches, a semi-automatic and an automatic approach, and presents a case study for an automatic solution. Following the observations of the case study, we come up with recommendations for the improvement of the semantic enrichment by an manual annotation step.