Ironically, the key to successful audio and video search is text. Text is the navigation “currency” of the web – so multimedia must be linked to text, and then published to the web. As text, audio and video content is more “discoverable” and more readily “consumed” which makes it far more valuable to advertisers.
At the heart of EveryZing’s solutions is our core speech-to-text technology – the fruit of $100 million of government-funded research by BBN Technologies (inventor of the internet’s ubiquitous “@” symbol). EveryZing’s speech-to-text technology enables multimedia clips to be robustly indexed, increasing their “discoverability” by the web search engines and boosting online advertising opportunities. EveryZing uses its technology to ensure that every piece of audio and video from each client’s web site is wrapped in a rich layer of metadata, including a full text output of the spoken word track, so it can be searched and accessed easily and precisely by consumers, just like text… and, as a result, online advertisers can now place contextually relevant messages within and along side multimedia content, just like text.
EveryZing’s infrastructure is built to scale with the web. Our platform is capable of ingesting audio and video content in any format, include live streams. Content can be processed in near real-time, ensuring perishable content such as news and business information gets to your users first. Our architecture has the capacity to process millions of hours of content and millions of searches and page views in a fault-tolerant, high-availability environment.
EveryZing’s proprietary natural language processing technology (NLP) can be applied to any text source to identify key terms, concepts, people, places and other valuable entities to create automatic “tags” for your content. The NLP can run automatically against a document or can be directed using custom dictionaries provided by the customer and used to feed existing directories and taxonomies.
EveryZing’s scalable indexing technology can process hundreds of documents per second and search across collections with millions of documents.
May 12, 2008
May 7, 2008
April 28, 2008
