Voice Recognition

Integrated Speech Recognition

Winscribe provides flexible front end speech recognition and back end server processing. Server-based recognition is carried out after the author has completed the dictation, while front-end recognition is processed on each individual PC, so the author can view and edit the recognized text as they dictate.

Both front-end (client) and back-end (server) speech recognition can be combined and deployed in many different scenarios to suit individual requirements. Both combinations also retain all the workflow management features that make Winscribe so powerful. Authors can review their own reports before and/or after they are transcribed, or have them reviewed by others.

Client-side Speech Recognition

Using Winscribe to process speech recognition locally allows authors to:

• Access and correct text immediately for ‘on the spot’ sign-off and printing or emailing of reports.
• Review and complete reports online, even when transcriptionists are not available.
• Watch the speech recognition process, or hide it until they wish to review it.
• Choose not to use speech recognition processing, in which case the dictation will be routed for manual typing.
• Format text as they dictate, or use voice commands to edit text.
• Make spoken changes and additions anywhere in the document with immediate recognition.
• Use speech recognition when offline, and synchronize with the server later.
• Mark the level of correction completed on each report – corrected, partially corrected, or uncorrected.
• Fully or partially self-correct their own dictation. Regardless of the level of correction the dictation can still be routed to a reviewer or transcriptionist and all reports remain available for the department’s review process.

Server-side Speech Recognition

The Winscribe server-side speech recognition solution is highly scalable and can be installed on any size site; from a small site with a single dictation server, to a large site with multiple servers. Speech recognition is processed centrally and all reports are routed to transcriptionists or back to the author. This means:

• Authors require no extra training.
• Speech recognition can be seamlessly integrated into your Winscribe solution.
• Transcription is handled in one place, by professional transcriptionists.
• Procedures can be streamlined to ensure maximum efficiency.
• Large reports can be batch worked on a server and corrected by internal or external transcriptionists, or rerouted to an author for later sign-off.
• There is minimal impact on client PC performance since all speech recognition processing is performed at the server, eliminating the need for expensive upgrades to client hardware.
• Authors have the flexibility of using any of the wide range of input devices that operate with Winscribe.

Additional Features

• Over 300,000 active words in the dictionary for unprecedented flexibility.
• Vocabularies available in US English, UK English, Australian English, Indian English, SE Asian English, German, Spanish, French, Italian, Dutch and more…
• Text to audio synchronization meaning that words in the text file are linked to words in the audio file and highlighted.
• Acoustic models representing how an author speaks, refined for each individual author. Context and acoustic adaptation to improve recognition rates.
• Personal dictionaries for the author to add, train, and manage to continually improve recognition rates.
• Macros to “shortcut”’ a commonly repeated set of words, using a single word or sentence to insert the text block.
• Microsoft Word templates for automatically formatting corrected text and linking to the job type.
• 160 words per minute recognition rate.
• Scaleable to over 120 million lines of recognized text per author.