Make your multimedia more discoverable and accessible
- Automatically generate standard caption files
- Search your library using deep neural network-based lattice indices
- Choose from a rapidly growing selection of languages
- Use custom vocabulary adaptation to recognise domain-specific speech content
- Perform jobs in parallel and easily integrate them into your existing workflow
- Extract spoken keywords to aid tagging and recommendations
Highly accurate audio search results
Azure Media Indexer automatically makes your media deeply searchable without the need for manually applied metadata. Using deep neural network (DNN)-based speech recognition technology from Microsoft Research, Media Indexer converts digital audio into natural language and automatically extracts meaningful metadata from your media.
Innovative custom vocabulary adaptation
With its custom vocabulary adaptation, Media Indexer consistently outperforms industry-standard speech transcription technology. Indexing medical lecture content? Submit custom words such as “aneurysm” alongside your indexing job, and watch as Media Indexer scours the Internet to include related words such as “haemorrhage” or “embolism” to its internal dictionary, dramatically increasing accuracy.
Auto-generated closed captions
Reduce the effort required to make your multimedia accessible by passing your content through Media Indexer. Use the output caption file (in your preferred format) to provide closed captions for your users.
Extract keywords from speech
Media Indexer can generate keywords from speech content in your multimedia and produce an XML file containing the frequency and time offset of each spoken keyword and other valuable data. Use the file to perform speech analytics, tag your content or power a recommendation engine.