New technology from Microsoft may make pouring through hours of audio recordings and videos for the right content a thing of the past.
Launched on Sept. 10, Microsoft Azure Media Services Indexer leverages technology from the company’s research division to extract “keywords from a video file as metadata that can be used for categorization of audio-video in a custom application, such as whether the content pertains to entertainment, politics or neural-network research,” wrote Microsoft spokesperson Athima Chansanchai in a company blog post. The company is offering a public preview of the technology during IBC2014 (International Broadcasting Convention), currently being held in Amsterdam.
The tech powering the Azure Media Services Indexer hails from a Microsoft Research project called MAVIS, short for Microsoft Audio Video Indexing Service.
MAVIS has already spent time in the field, according to a separate Inside Microsoft Research post by author Rob Knies. “It has been used to index the digital archives for the U.S. state of Washington and also has been deployed by the U.S. Department of Energy, the British Library, NASA’s Jet Propulsion Laboratory, and The Washington Post,” he wrote.
Two years ago, Microsoft enhanced MAVIS with a new algorithm for improved speech recognition accuracy and error reduction. Additionally, the algorithm cut processing time by 30 percent compared with the top speech-recognition algorithms of the time.
Now, those innovations are being made widely available to Azure customers.
“We provide the ability for customers to build a search experience that would allow a deep search for spoken words inside audio and video archives, an effective way of browsing archives,” said Behrooz Chitsaz, Microsoft Research director of IP Strategy, in a statement. Further, the technology “can deliver a machine-generated draft description for video and audio content, sufficient for a basic understanding of the clip in question,” he said.
In addition to making audio and video content searchable, the technology could yield new sources of information for businesses.
“There’s more and more audio-video, whether it’s social audio-video, whether it’s broadcast, or in the enterprise,” stated Chitsaz. “We’re giving people the ability to unlock the information in their audio and video content.”
The indexer is part Microsoft’s cloud-based Azure Media Services suite for media streaming organizations. Azure Media Services provides “everything you need to quickly build great, extremely scalable, end-to-end media solutions for streaming on-demand video to consumers on any device,” stated Scott Guthrie, then the corporate vice president for Azure, during its launch last year. He now serves as executive vice president of Microsoft’s Cloud and Enterprise group.
“For example, you can easily build a media service for delivering training videos to employees in your company, stream video content for your Website or build a premium video-on-demand service like Hulu or Netflix,” he boasted. Now, it encompasses a variety of offerings, including media processing and encoding, mobile video delivery and digital rights management.