Sunday, July 20, 2008

Next Level Internet Searching Is Coming Soon...

OK, so you go to Google, Yahoo, Live, etc and can search and find results that match your text within a website. Cool. Then you can search a word and find an image that matches the result of the word you want. Sweet.

Now get this. Programmers are now incorporating audio-to-text and video-to-text with online audio and video and your text-based query. FAR OUT !!!

EveryZing uses speech-to-text technology to search for audio and video around the web. It will even search for words in YouTube videos. Right now the search results are really nothing to write home about, but give these guys a year and EVERYONE will be searching like this.

Formerly PodZinger until May, 2007, Everyzing is a Cambridge, MA based audio and video search engine that leverages speech recognition technology developed by its parent company, BBN Technologies, to look for content inside audio and video files formats. Everyzing creates a text index from the audio and video feed and uses that index to find relevant terms within published audio and video files.

As a video search engine, Everyzing highlights the segment of an audio or video file by creating text snippets around the searched terms. The user can click anywhere in the snippet to listen to or watch just that snippet segment, they can listen to the entire file, and they can download the file or subscribe to feed. Everyzing is able to perform multimedia searches in both English and Spanish, and generates RSS alerts.

Zing Index is a forum dealing with podcasts which highlights the most talked about people, the most listened audio podcasts and the most viewed video podcasts. Blog Zinger is a topical blog with strong ties to Everyzing, which was created to promote open discussions and conversations on podcasts, search, and technology-related topics as an insight into the world of Everyzing.

Recently Thomas Wilde, the new CEO of EveryZing, acknowledged that Everyzing works 70% of the time on average when there is music, ambient noise or more than one person speaking. If newscast style speaking (one person, speaking clearly, no ambient noise) is available, that can rise to up to 93%.



[Google System]

No comments:

Post a Comment