Video and multimedia data have become the predominant information on the World Wide Web. To cope with the ever growing amount of multimedia data on the web search engines have to open up the media content for search and retrieval. Automated multimedia analysis technologies such as, e.g. automated speech recognition, video OCR, or visual concept detection help to open up large scale multimedia repositories although the achieved analysis results often are error prone and unreliable. Semantic analysis considers the multiple (mostly text-based) metadata streams from automated analysis and constructs a semantic context to enable understanding the media content. Thus, semantic analysis enables the improvement of metadata reliability by evaluating the plausibility of the semantic assumptions. In addition, semantically annotated multimedia data enables semantic and exploratory search to open up new ways of accessing multimedia repositories.