Speech-to-text perfected

Digital Syphon Sonic Extractor

Speech-to-text perfected from from any source.

With Sonic Extractor you can generate multiple transcriptions from audio content in 1/4, 1/8 or up to 1/16 of the time it would normally take you to process a given set of files.

All one has to do is direct Sonic Extractor appliance or Cloud Service to the appropriate Folder path, stream url location or select a live input source to begin processing of a given transcription.

Once Sonic Extractor has a stream the system will begin to analyse and listen for any potentially responsive content, such as alert terms or keywords that one may want to be notified of with their appropriate time stamps. Additionally the XML file will contain a breakdown of all terms transcribed with the likely word and duration of such terms, as well as other metadata such as speaker ID, gender Identification. One can also employ Named Entity Recognition for the generation of additional metadata within the transcription process to provide additional Media Intelligence Gathering capabitlies for indexing into a Search Engine or employing a workflow process to utilise as a alert interceptions desired against any for exmaple telephony devices.

Sonic Extractor accomplishes all of this through the utilization of large vocabulary Markov Models, and an advanced acoustic methodoloy to create a robust solution to quickly and accurately transcribe all spoken audio. The system is capable of being configured to transcribe most European, and Asian languages depending on the needs based on the geographic location of each corporate office.



Benefits:
  • Start getting a powerful ROI immediately
  • Listen, Process, Read for productivity
  • No end user agents to install
  • Easily adaptable to any environment
  • Internationalize for all major languages
  • Never worry if something is missed
  • Search in any language
  • Asian and Arabic solution available as well
  • Easily adaptable to any environment


Key features:
  • Automated Alerting (Live version)
  • Can Process 4, 8 or 16 hours of audio in 1 hr
  • Language Independent
  • Comes in configurations to process streams, live, and file base content
  • Generate XML files or RDBMS option
  • Appliance model


Appliance Specifications:
  • Native 64 bit computing
  • 8GB of Memory
  • Network-attached appliance
  • 1TB RAID storage
  • Smart Card Technology for DSP/licensing
  • Redundant 400-600 watt power supplies
  • 4 Channel Encoding (Live version)


Supported languages:
  • ARABIC, ENGLISH, SPANISH,FARSI, RUSSIAN, DANISH, SWEDISH, CATALAN, MANDARIN, JAPANESE, KOREAN, GERMAN, DUTCH, PORTUGUESE, ROMANIAN TURKISH, GREEK, FRENCH, ITALIAN


Audio Types
  • The System can process MMS (wma,wmv,mp3), Shoutcast, MP3(m3u), RTMP(flash-mp3-aac3), and Icecast streams


REPORTING AND AUDITING
  • Log of what streamed
  • Log of what is processed by date



To speak with a Digital Syphon Sales Manager, please contact us at 469-854-3528.
Contact our Sales team by email at sales@digitalsyphon.com