Medical Transcription

Since its first beginnings in the 1980's, Dragon Naturally Speaking has evolved and been owned by 4 different companies. The first research prototype of the speech recognition software we know today was invented by Dr. James and Dr. Janet Baker (husband and wife). In 1982 they founded the company known as Dragon Systems. It was not until 1997 that the first version of Dragon Naturally Speaking hit the market place.

Today this speech recognition software package is now produced by Nuance Communications for use with all PC's that run Windows. This particular software package has 3 primary areas of function, and these are shown below.

1. Dictation This is the function where the spoken word is then converted into text.

2. Commands This allows for the spoken word to be recognized as a command. Say for example, when a person says stop or pause.

3. Text to Speech This software program allows for the written word to be converted into an audio stream (although it is synthesized).

In the initial stages, in order for the software to be used correctly, a person would need around about 40 minutes training the software to recognize their voice. However, by Version 9 which is the latest of all versions of Dragon Naturally Speaking, this time is no longer required to train it to recognize a user's voice.

It is claimed by the producers of this software that a person should be able to produce a 900 word essay in around 6 minutes. This is considerably less time than it would take for someone to type such an essay if they could type at 40 wpm.

However, even though the latest version of Naturally Speaking is advertised as being able to be around 99% accurate, many people are still complaining that this product is too cumbersome and too fussy to use. In fact, a lot of people become extremely frustrated with the program as they need to be continually correcting the recognition errors that it is making. In fact, you will find that the most widely discussed topic on many forums relating to speech recognition software concerns the accuracy of these systems.

The problem with regard to the accuracy of Dragon Naturally Speaking software is that although the producers - Nuance Communications - are claiming that a person using this system can expect around a 98 to 99% accuracy, this has not as yet been independently substantiated.

For Sure, one of the best ways to ensure that you get the highest levels of recognition accuracy out of Naturally Speaking is through ensuring that you have a good quality input signal and also that you have a very good quality computer system also. In fact, most people who are already using this type of software program will suggest the first thing that you do after purchasing yours is to buy a better quality headset.

Medical Transcription | Privacy Policy |