skip to content
 

Dragon Naturally Speaking is assistive technology that transcribes what you say into text onto your computer. You may be awarded this through Disabled Students' Allowances, if it is appropriate for you.

Q. What Computer Should I buy for Speech Recognition Use?

Dragon NaturallySpeaking is a PC-only speech recognition application from Nuance. There are different versions of the software and the recommended machine specification varies between these versions. We currently advise upon and support Dragon NaturallySpeaking versions 7 to 10.

Dragon NaturallySpeaking v10

Dragon NaturallySpeaking V10 was launched in 2008. The minimum machine specification required to run this latest version of the software is:

  • Intel Pentium 4 or later, AMD Athlon 64, 1GHz or later - Faster processors will give faster performance speeds
  • Hard disk - 1 GB of space required for system and file storage
  • RAM - 1GB of RAM, more memory gives faster performance speeds
  • DVD-ROM drive - this is required for installation purposes
  • Sound card - Creative Labs SoundBlaster 16, or other soundcard supporting 16-bit recording
  • Microsoft Windows - Windows Server 2000 or 2003, Windows 2000 (SP4 or higher), Windows XP (SP2 or higher, 32 bit) and Windows Vista or Windows Vista (SP1, 32 bit)
  • Microsoft Internet Explorer 6 (or higher) - This is required for on-line activation of Dragon NaturallySpeaking
  • USB Microphone headset - We recommend the Plantronics range of microphones
  • During the installation process the software will check to ensure that your system meets these requirements. Installation may terminate if your system does not meet these minimum requirements.

  • Dragon NaturallySpeaking v9

    Dragon NaturallySpeaking V9 was launched in July 2006. The minimum machine specification required to run this version of the software is the same as the above specification for versions 7 & 8, except:

  • Intel Pentium processor (or equivalent AMD processor) - 1Ghz processor. Faster processors will give faster performance speeds
  • Hard disk - 1 GB of space required for system and file storage
  • During the installation process the software will check to ensure that your system meets these requirements. Installation may terminate if your system does not meet these minimum requirements.

    Dragon NaturallySpeaking v7 and v8

  • Intel Pentium processor III (or equivalent AMD processor) - 500MHz processor. Faster processors will give faster performance speeds
  • During the installation process the software will check to ensure that your system meets these requirements. Installation may terminate if your system does not meet the minimum requirements

  •  
  • RAM - We recommend 1GB of RAM as a minimum, more memory gives faster performance
  • Hard disk - 500 MB of space required for system and file storage
  • Sound card - Creative Labs SoundBlaster 16, or other soundcard supporting 16-bit recording
  • CD-ROM drive - this is required for installation purposes
  • USB Microphone headset - We recommend Plantronics 100 or 400
  • Microsoft Windows - XP (SP1 or higher), 2000 (SP4 or higher)
  • Microsoft Internet Explorer 5 (or higher) - This is required for on-line activation of Dragon NaturallySpeaking

Q. How long does it take to train Dragon NaturallySpeaking to understand my voice?

Dragon NaturallySpeaking is renowned for being easy to install, train and use.

Dragon NaturallySpeaking v7, v8, v9 and v10

  • Software installation - is complete within minutes providing the hardware specification of the machine meets the minimum specification required.
  • Create a user - Create a user profile name, select the sound source such as microphone headset.
  • Sound & quality check - select the sound source you are going to use and read aloud the on-screen passages to test the sound levels and quality output from the microphone.
  • General training - Creates personal speech, vocabulary and language models that contain data about your voice, pronunciation and use of words. These models allow the software to understand you when you speak. General training will take around 8-10 minutes depending upon how fast you speak.
  • Begin dictation - You are able to begin dictation directly after the general training has been completed.

If you are upgrading hardware or software you can prevent the need to re-train Dragon NaturallySpeaking by exporting or copying your old voice files to the new system. Details of how to do this can be found here: Copying User Files With Dragon NaturallySpeaking.

There are many variables that can influence the overall time it takes to become a proficient user of Dragon NaturallySpeaking. The most important factor to create maximum efficiency with the software is to correct any errors it makes as you go along. This will greatly increase the overall recognition accuracy achieved. There are various functions provided within the software which are designed to improve the recognition accuracy. Further details about these functions can be found here: How Can I Improve My Recognition Accuracy

Consistent usage of the system will increase a users speed and skills with the software.


Dragon NaturallySpeaking v9 & v10

Versions 9 & 10 also provide a 'no-train' facility which allows the user to begin dictating without completing the general training section. In our experience, it is still recommend that you perform the general training exercise to maintain good recognition and accuracy levels.

Q. Do I need to re-train Dragon NaturallySpeaking v8 if I have upgraded my hardware or software?

If you are upgrading hardware or software you can prevent the need to re-train Dragon NaturallySpeaking by exporting or copying your old voice files to the new system. Details of how to do this can be found here: How can I export or copy user's voice files from Dragon NaturallySpeaking.

Q. Exporting and Importing User Files with Dragon NaturallySpeaking

The Export Function

To export your user files using Dragon NaturallySpeaking v12:

  • Select Profile from the Dragon toolbar click Manage User Profiles 
  • In the Manage User Profile dialogue box select a user profile from the list and click the Advanced button.
  • Choose Export from the list
  • In the Browse for Folder dialogue box select the destination to export the user files to (Make a note of the location where the user files are being exported to for future reference).
  • Click OK to begin the export process. 

The Import function

To export your user files using Dragon NaturallySpeaking v12:

  • Select Profile from the Dragon toolbar click Manage User Profiles 
  • In the Manage User Profile dialogue box select a user profile from the list and click the Advanced button.
  • Choose Import from the list
  • In the Browse for Folder dialogue box navigate to where user files are located.
  • Click OK to begin the export process.

Voice File Location

Dragon NaturallySpeaking stores each users voice files in the following locations, depending upon the version of software you are using some of the folder names will be slightly different:

Dragon NaturallySpeaking 12:

  • C:\ProgramData\Nuance\NaturallySpeaking12\Users\  (Windows: Vista and 7)

Dragon NaturallySpeaking v 11:

  • C:\ProgramData\Nuance\NaturallySpeaking11\Users\  (Windows: Vista and 7)

Dragon NaturallySpeaking 10:

  • C:\ProgramData\Nuance\NaturallySpeaking10\Users\  (Windows Vista)

Q. How can I Improve Recognition Accuracy?

The key to good recognition accuracy is clear enunciation, speaking in phrases and consistent microphone placement. If you are still experiencing poor recognition accuracy Dragon NaturallySpeaking provides various tools to increase your accuracy.

Many of these functions can be accessed from the Accuracy Centre located under the Tools menu on the Dragon Bar.


Check your microphone

If you are not getting 95% accuracy, or above, and you are speaking clearly and steadily in phrases, then it is worth checking your microphone and soundcard combination. Check your microphone is properly positioned and correctly attached to the machine. Ensure you have the latest drivers for your soundcard.

A simple way to test the sound is to use the Sound Recorder facility that is built into the Windows operating system.
To start the Sound Recorder in Windows 7, click on the Start menu and chose All Programs -> Accessories -> Sound Recorder

How to test your voice:

  1. Click on the Start Recording button
  2. Talk into the microphone for roughly 5 seconds
  3. Pause for five seconds
  4. Talk again for a further 5 seconds
  5. Click on the Stop Recording  button to stop recording 
  6. In the Save As dialogue box save the file to the desired location
  7. Double click the file to play it back

Does the playback sound clear? You should get a good wave pattern, if not, then there is something wrong with either your microphone or your soundcard. During the part of your playback where you were silent can you hear any crackling or rustling sounds? These types of sounds during the silent time may be originating from faulty connections or noise being picked up from inside the machine. At this point it is worth testing with a new microphone and/or soundcard as this may be the cause of the loss of accuracy.

Return to top


Perform Additional Training

Carrying out additional training will help to improve your recognition accuracy. This can be carried out at any time but especially when

  • You have used Dragon for a few days and have become more accustomed to dictating
  • You move to an environment that is significantly quieter or nosier than your previous environment
  • You change your microphone and/or your soundcard

To access the Additional Training facility

  • Click on Audio on the Dragon Bar
  • Choose the Read text to improve accuracy option
  • Follow the instructions in the General Training wizard
  • Select one of the texts to read out, then follow the same guidelines as you did when you completed the initial training. 
  • Click Finish when you have finished reading the text.

Return to top


Add a single word or phrase to your vocabulary

Any word or phrase can be added to your vocabulary list by going to:

  • Vocabulary on the Dragon toolbar, choose Add new word or phrase. 
  • Type or say the word you want to add into the left hand box of the Add new word or phrase dialogue box.
  • If the word is pronounced differently from how it is spelt then enter a 'Spoken form' of the word into the right hand box of the Add new word or phrase dialogue box.
  • Make sure the check box for I want to train the pronunciation of this word or phrase is ticked.
  • Click Add.
  • Click Go  in the Train Words  box
  • Clearly speak the word or phrase you are trying to add

Return to top

 

Add Unknown Words from your Documents to the Vocabulary

To improve accuracy it is advisable to give Dragon samples of your existing documents. This will help it work out your writing style and the words that you use most often. Dragon will ask to do this process just after you have created your voice profile for the first time but you can launch it at any point thereafter.

To add words from existing documents to your vocabulary list

  • Click on the Vocabulary option on the Dragon Bar
  • Select Learn from specific documents from the list
  • Leave the default options as they are in the in the Learn from specific documents window
  • Click Next.
  • Use the Add Folder/Add Document buttons to navigate to the document or folder of documents that you want to add.
  • To add additional files repeat this process using the Add Document button.
  • Click Next for the scanning process to start.
  • Click Next to see a list of unknown words that Dragon can add into your voice files.
  • Use the Check all/uncheck all to select which words to add
  • Click Next
  • Select the words to train
  • Allow the process to Adapt to your Writing Style, click Next
  • Click Finish

NB: You should check carefully through the list of words for any incorrectly spelt words, if you find any, un-tick them to avoid added them to the vocabulary list.

Return to top


Import a List of Words or Phrases 

This is a useful way to add lists of words to Dragon. First, create a text (.txt) file containing each word or phrase you want to add. Each item should be on a separate line. Ensure that all words are spelt correctly.

  • Open the Vocabulary menu from the Dragon Bar
  • Click on the Import list of words or phrases option
  • Click Next to navigate to the location of the .txt file containing the new words you want to add
  • Click Add File to begin the process of adding the words

If you want Dragon to add words from more .txt files then repeat this process with each new file.