Fraunhofer IDMT presents its latest technologies for A/V analysis and media asset management
At this year’s International Broadcasting Convention IBC taking place September 15–19, in Amsterdam, Fraunhofer IDMT will be presenting a further updated set of its A/V analysis and media management tools. Complementing previous presentations on automatic metadata extraction and quality control, this year’s presentations include new and unique tools to automatically detect relationships and the processing history within A/V content sets, as well as an Open-Source framework for multi-modal analysis and recommendation.
At previous events, Fraunhofer IDMT demonstrated various technologies for automatic metadata extraction, such as speech-music discrimination, music annotation and similarity, video shot and object detection. This year, presentations will include various tools for analyzing content relationships and datasets, as well as managing complex analysis workflows.
The exhibit further contains A/V segment matching, which can be used for a surprising number of purposes. Patrick Aichroth, head of Media Distribution and Security at Fraunhofer IDMT explains: “Our software can be used to detect partial duplicates. Depending on our clients’ requirements, this can be used to remove redundant material and support automatic propagation and curation of metadata and rights clearance. However, it can also be applied to analyze radio and TV streams, to track ads or to support archive search for material that was recorded at the same event.”
A/V segment matching is perfectly complemented by another Fraunhofer IDMT technology: “Audio phylogeny is like family research for audio - it is about identifying the relationships within a set of similar items”, describes Aichroth. "The software, to our knowledge the first of its kind, is able to automatically detect the best-quality original version and coding history within a set of audio duplicates, which can be very obliging e.g. for duplicate removal."
Fraunhofer IDMT also offers a set of unique “audio forensics” tools that can be used to automatically detect previous encoding steps and parameters, identify recording devices, and localize editing of previously encoded material, all via baseband analysis. The tools can be used to detect and avoid audio coding problems, optimize audio processing and support metadata annotation and search.
Finally, presentations will include an overview over “MiCO”, an Open-Source framework for multi-modal semantic search and recommendation, which can be used to integrate flexible and complex analysis workflows into existing platforms.
Fraunhofer IDMT will also demonstrate its recently-integrated automatic QC functionalities for easyDCP, the professional movie post-production solution, at the easyDCP exhibition from Fraunhofer IIS at the same booth.
If you are interested in any of the aforementioned topics, please visit our “A/V Analysis and Search” exhibit at Fraunhofer booth 8B80.