Hear ye, search ye

In a move designed to stay ahead of its competitors, Convera has added optional phonetic audio search to its RetrievalWare search and categorization platform through a partnership agreement with Nexidia.

The agreement allows RetrievalWare to search 30 hours of audio information in less than one second--100,000 times faster than normal playback speed, reports Convera. The company adds that for enterprise areas with significant audio components--call centers, compliance monitoring, knowledge management, forensics, training, intelligence gathering, etc.--removing the stumbling block of speech to text conversion expedites access to real-time information.

Nexidia explains that its phonetic searching pinpoints spoken words in word or phrase formats by analyzing the phonetic content of indexed audio or video files based on the similarity of sound. Hence, RetrievalWare’s new audio search feature will enable customers to search for spoken words without converting the audio to text through speech-to-text software or human transcription.

RetrievalWare’s phonetic search capability identifies not just dictionary words, but also slang terms and code words. In addition to retrospective search, Convera will offer Nexidia technology for real-time monitoring of audio, matching capabilities already offered by Convera’s profiling feature for processing textual content.

Other RetrievalWare features for managing and searching audio and video data include:

• real-time capture and scene recognition within digital or analog video,

• video annotation and rough-cut editing, and

• retrieval based on transcribed text.

Both Convera’s existing video search capability and the new phonetically-based search from Nexidia allow users to go directly to the desired location within video or audio, eliminating the manual “fast forwarding” process to a specific part of the video or audio information.