New HD Cloud video processing technology speeds large-scale transcoding tasks
October 26, 2009
Targeting video-on-demand services and the fast encoding of large film libraries,
HD Cloud, a video technology service provider developed by Diversion Media, has unveiled a service called HD Accelerate that can transcode a two-hour movie file in multiple separate video formats in less than 10 minutes. It uses a special parallel encoding technique developed by HD Cloud that enables an unlimited number of output formats.
The company plans to launch the service to large-scale video-on-demand providers and film studio libraries in the coming months.
Separately, the company also announced that HD Accelerate has performed 4500 video encodes at 100Mb/s each in a 20-minute period using 900 processors. HD Cloud handles large video files and high bit-rate content (up to 1080p and 100Mb/s).
Since its launch, HD Cloud has begun working with feature-film distributors that require scalable “in-the-cloud” transcoding for on-demand feeds. The company’s HD video encoding service is now being used by WeatherChannel.com, “New York Magazine” and Zappos.com (through a partnership with Magnify).
HD Accelerate uses a series of proprietary video scaling algorithms for enhanced video transcoding throughput, particularly for on-demand video services. It has performed 900 simultaneous encodes, and recent enhancements to HD Accelerate enable greater scalability and capacity for its cloud computing-based service. HD Cloud currently performs terabytes of transcoding sessions for customers, applying metadata to completed files and automatically distributing files to destinations ranging from syndication Web sites, content delivery networks (CDNs), local archives, content management systems (CMS), and local drives.
The HD Accelerate concept has been combined with another HD Cloud innovation, a security strategy in which proxy files are wiped from the server cloud once files are encoded, adding another layer of protection for customers.
comments powered by Disqus.