Azure Media Indexer has been updated with set of new features, improvements and bug fixes. The recent update provides a general high level recognizability score for each job to enable users to determine whether or not the provided media file was recognized by our language models. This is in addition to the display of scores on the confidence of each word.
“This can prevent the accidental inclusion of unusable output files from Indexer due to poor audio quality etc,” mentions Adarsh Solanki,Program Manager, Azure Media Services.
Azure Media Indexer Version 18.104.22.168 includes a demo Java code contributed by Andrew Weiss.
The recent update has fixed a issue where users were having issues running Indexing jobs when input asset names or configuration XML files contained special characters. Hence, all regular Unicode characters are supported.
Moreover, the SMI and TTML output files now accurately report MIME types as application/smi+xml and application/ttml+xml in addition to certain minor performance improvements.
Microsoft Azure product team is working hard to provide support for more languages. The Spanish language build is currently in a private preview with many more updates planned for the upcoming year.
Please refer to the official blog post if you would like to learn more about Azure Media Indexer.