Image of Petrana

Petrana Home







Watch a Movie with Petrana







Petrana Assistive Technologies







Build a Home Theater System for Petrana







Petrana Support and Community


About GFT Group Inc.





Buy Now

Share your discoveries of the best ways to use Petrana with other members of the Petrana community. Let Petrana know about problems, request new features, or just say hello.

Send Mail to Petrana

Petrana Users Group (Yahoo group)

Join the Petrana Users Group to receive the latest news and information for Petrana Users and to share information with other Petrana Users.

Subscribe to Petrana

Powered by

Buy "I watched a Movie with Petrana!" T-shirt and other gear


Petrana FAQ (Frequently Asked Questions)


The PetranaVoice Commands for DVD provide an easy-to-use conversational interface for playback, operation, and control of DVDs. The Petrana Voice Commands for DVD require Dragon NaturallySpeaking™ 7.3 Professional from ScanSoft Inc. and currently support the InterVideo WinDVD™ 5 and 4 DVD player software.


What can you use the PetranaVoice Commands for DVD for?


The commands may be used for:


1. Playback, operation, and control of DVD discs during presentations and classes. In this case, the user will probably use a wireless microphone system to control their computer remotely.


2.  Playback, operation, and control of DVD discs on a voice-operated computer with a wall display in the home or office.  A voice-operated computer with a wall display typically includes a wireless microphone system or a wide-area microphone that allows the user to move freely throughout the home or office. 


3.  Playback, operation, and control of DVD discs by persons with various disabilities.


What or Who is Petrana?


Petrana is an artificial personal assistant.  The Petrana Voice Commands for DVD are a component of the full Petrana system. Petrana provides an easy-to-use conversational interface for your computer that emulates a personal assistant to the extent feasible using current technology.

Petrana is a front-end for the Dragon NaturallySpeaking 7.3 Professional voice-recognition software from ScanSoft, Inc. Technically, Petrana is implemented as a set of additional commands for Dragon NaturallySpeaking 7.3 Professional.



Why Do You Need Petrana?


Petrana makes your computer act like a personal assistant and enables you to ask your computer to perform tasks using intuitive commands similar to spoken English and normal conversation.

Petrana makes it easier and more enjoyable to perform many common computer tasks such as DVD playback or making presentations to a large audience.


What Should You Expect?


Most of us are used to television shows and movies such as Star Trek in which computers can recognize and understand human speech perfectly. Talking to science-fiction computers is like talking to human beings.  These computers can understand high-level English commands and perform sophisticated tasks. While there has been significant progress in voice-recognition over the last several years, present-day voice-recognition technology has not achieved this level of performance.

The Dragon NaturallySpeaking 7.3 Professional voice-recognition software used by Petrana typically does not perform with 100% accuracy.  With a normal healthy speaker, a good sound system, and negligible background noise, as well as the influence of other factors — based on tests at GFT Group Inc. — the voice-recognition software may achieve an accuracy rate of about 95%, meaning that about 95% of the words spoken by the user will be transcribed correctly during general dictation.  If you experience an accuracy rate of 90% or less, you may need to reassess your system or you may have an atypical voice.

Modern-day computers have been designed for computer keyboards and pointing devices such as a mouse or trackball.  Modern-day computers and software are not designed for voice operation.  As you work with voice-recognition software, you will probably find that even with perfect accuracy in transcription of general dictation, it would still be difficult to perform many common computer tasks.  For example, computer keyboards have many special characters and action keys that are easy to press with your fingers but difficult to emulate by speaking.  Pointing devices such as the computer mouse or trackball are inherently difficult to control and emulate by voice. Thus, computer applications that rely heavily on special characters, action keys, or a pointing device can be difficult to control by voice.  Some computer applications such as drawing or computer animation inherently require a pointing device.

Unlike a human listener, Dragon NaturallySpeaking 7.3 Professional does not understand what you say.  It is designed primarily to transcribe what you say as accurately as possible given the limitations of current voice-recognition technology.  When we ask a human assistant to do something, the human assistant usually understands what we are asking. Similarly, to achieve the performance of computers in television shows such as Star Trek, the computer must understand or simulate understanding of spoken speech.  Perfectly accurate transcription of spoken speech is not sufficient.


How Does Petrana Improve the Performance of Voice-recognition Software?


GFT Group has developed Petrana to eliminate some of the limitations of current voice-recognition technology, partly by adding some emulation of understanding and intelligence.  However, Petrana cannot duplicate the capabilities of a human listener or science-fiction computer.

General dictation can be difficult with present-day voice-recognition technology, mainly because of limited accuracy which forces the user to break his or her train of thought and frequently correct the dictated text. However, the use of voice-recognition with Petrana can be used to perform a range of other computer tasks such as DVD playback both successfully and enjoyably. Typically these tasks require only light or moderate use of voice-recognition. Often in such tasks, the user will wait for the computer to respond to a request or command before continuing. In these cases, the user's train of thought is not broken if the computer does not recognize the spoken request or command. The user will simply repeat the spoken request or command, and then continue with what he or she was doing.

Many practical applications for a voice-operated computer with a wall display require light or moderate use of voice-recognition, unlike dictation or computer programming which require extensive use of voice-recognition. For these practical applications, occasional errors in recognition by the voice-recognition software are easy to tolerate and are offset by the many benefits of the system.  A Home Theater System using DVD player software is an example of such an application.


Can I Control My Mouse Pointer by Voice?


Dragon NaturallySpeaking 7.3 from ScanSoft Inc. includes Voice Commands to control the mouse pointer.  These include "mouse up/down/right/left <number of pixels >" to move the mouse pointer the specified number of pixels in the specified direction, "mouse move up/down/right/left < | fast | very fast>" to start the mouse pointer moving in the specified direction at the specified speech, a MouseGrid command to position a mouse pointer at a precise location on the screen, and several other useful mouse commands.

The Petrana Voice Commands for DVD include Voice Commands to quickly move the mouse pointer out-of-the-way while playing DVDs.  This prevents the mouse pointer from appearing in front of the video as it is playing and also avoids accidental clicking on highlighted DVD menu items which may otherwise occur.


I find headsets uncomfortable.  Do I have to wear a headset?


No.  You may wear a lavalier microphone that clips on your shirt or blouse.  For example, the Sony WCS-999 wireless microphone system used in our initial testing of the Petrana system comes with an inexpensive omni-directional lavalier microphone.  You should use a unidirectional Lavalier microphone. However, most inexpensive lavalier microphones are omni-directional microphones with less ambient noise rejection than unidirectional microphones or headsets.  For example, Dragon NaturallySpeaking has an error rate of 9-10% in our tests with the Sony lavalier microphone included in the box with the Sony WCS-999 wireless microphone system.  In our tests, this error rate is usually the very edge of usability for most users.  With an inexpensive headset an error rate of 5% is frequently possible, which is much easier to use.


Unidirectional lavalier microphones enable better performance than omnidirectional lavalier microphones, up to 95% accuracy in a quiet room.  Unidirectional lavalier microphones are usually more expensive than omnidirectional lavalier microphones, although more comfortable than headsets. We use the Sony ECM-DM5P unidirectional microphone (shown below) successfully at GFT Group.  

unidirectional lavalier microphone with wireless transmitter

A unidirectional lavalier microphone (the Sony ECM-DM5P unidirectional microphone) with wireless transmitter that can be clipped on your shirt, tie, or blouse.


You should choose the microphone with the right combination of comfort, performance, and cost for you.


Buy the Petrana Voice Commands for DVD Now


Copyright© 2004, GFT Group Inc., P.O. Box 66, Mountain View, CA 94042

Petrana Home Page       | Watch a Movie with Petrana | Petrana Assistive Technologies | Petrana Home Theater System | Petrana Support and Community | Buy Now | About GFT Group Inc.

© 2004 GFT Group Inc., P.O. Box 66, Mountain View, CA 94042. All rights reserved. Unauthorized redistribution of this document is prohibited.