Getting Started

<< Back to the GSI Demo page

Tutorial - Getting Started

Below is a brief walkthrough of the functionality provided by GSI Demo. In this tutorial you will learn how to

  • Set up a Speech Client
  • Use GSI Player to load an application mapping

Please note: Steps 1 - 4 are needed upon first use, steps 5 and above are needed each time the computer is restarted.

1) Before starting (and if you have not already done so), make sure that you have downloaded and installed all of the required components mentioned in the download instructions. Plug in your Diamond Touch table and make sure that it is working (try using the merldt.exe utility).

2) Train a speech profile by opening the Control Panel, selecting Speech

3) Make sure that the Microsoft English Recognizer is selected, then create and train a new speech profile by pressing on the New button.

4) Next, open the GroupLab Collabrary Shared Dictionary Monitor. This can be done by opening the Start menu \ All Programs \ GroupLab \ Collabrary \ Shared Dictionary monitor or c:\Program Files\Common Files\GroupLab\CollabrarySdMonitor.exe. It is advisable to create a shortcut on your desktop or start menu for easy access.

5) Create a new shared dictionary by typing in 'gsidemo' in the Textbox and pressing the Start New button. You can double click on the dictionary to see its contents (should be empty right now, but that will change once a speech client is added).

6) Now we can run our Speech Client, open the Start Menu \ All Programs \ Ilab \ GSI Demo \ Speech Client. This should open a new speech client, it already has a number of words that it is set to recognize, click on one of the words to view the result.

The speech client includes a check box to disable speech recognition, it also shows the current speech volume. A threshold is provided if you want all speech recognition events under a certain threshold to be ignored.

There is also a configuration utility at the bottom of the window that allows you to select the input device and speech recognition profile (this is particularly useful if you want to have multiple speech recognizers on the same computer with multiple sound cards). Also, there is an ip address option on the bottom right of the window that allows you to select which computer speech input will be sent to. This allows one to run speech recognizers across multiple computers if they choose. Any changes in the configuration is saved locally and will be retained when the speech client is opened.

7) To run the GSI Player, open the Start Menu \ ... \ GSI Demo \ GSI Player shortcut.

Click the Load Commands Button and select a mapping file to load (e.g., VirtualSurgery.config). This will allow you to load the mapping for the Virtual Surgery Application (open http://edheads.org/activities/knee/swf/index.swf in Internet Explorer). Notice that the list of available speech and gesture commands are written in the GsiPlayer listbox. You will also notice that the commands in the Speech Control panel have changed. Say "Scratch Pad" to enter a multi user annotation mode, say "Scratch Pad" to exit this mode.

With the GSI Player you can set a Speech Prefix (e.g., "Computer"), you can also set turn off the whistle upon successful speech recognition. If there is a multimodal command (e.g., Create table here [point]) you can change how much time is permitted between the speech and gesture components in order for it to be considered a recognized event. Also, you can change the Floor Control Policy that deals with multiple people issuing speech and gesture commands simultaneously.

8) Feel free to try some of the other mappings for existing applications