User Manual


The following is a comprehensive and illustrated guide to using our Speaker Verification System, as you will see, it is very user-friendly and easy to use.

Running the Program
Our project files must be installed onto C:\ into a directory of the users choosing, (see Installation Guide for details on this). Once installed, the program is run by typing: java Client from a command prompt window.

The initial User Interface
The initial screen presents the user with three options: Logon, Become new User or Set Threshold. Obviously in a practical application of this project, these three functions would not appear together, but for simplicity and to show functionality, we have done so here. Here is the initial screen presented to the User:
Initial Client Screen

To use any of the functions, simply enter your username (or desired username as the case may be) and hit the button to the right.
Whenever incorrect information is supplied, such as: Incorrect username when logging in or setting threshold, No threshold set before logging in or Trying to create a new user when name already exists. An informative error message such as this is displayed:
Error Message


The User logon Process
Once a correct (i.e. existing) username is supplied to the system, the random sequence of words is generated and displayed to the user. Then the logon Screen below is displayed.
Logon Screen

The user must press the record button and say the text to its left. The speech is then recorded and sent to the server, the sample properties are calculated and compared to a threshold, access is then denied (as shown below on left) or granted (as shown below on right).
Unsuccessful Logon  Successful Logon !


The New User Process
Once a correct (i.e. not already existing) username is supplied to the system, the New User Screen below is displayed.
New User Screen

Each "Record Sample" button must be pressed and the text to its left must be spoken slowly and distinctly into the microphone. Once all 17 have been recorded, the "Finished" button is pressed and all 17 WAV files are sent to the server and the model is built.

The Set Threshold Process
Once a correct (i.e. existing user without threshold set) username is supplied to the system, the Set Threshold Screen below is displayed.
New User Screen

Each "Record Sample" button must be pressed and the text to its left must be spoken slowly and distinctly into the microphone. Once all 5 have been recorded, the "Finished" button is pressed and all 5 WAV files are sent to the server and the user threshold is set. Once this had happened, the user threshold is set and the option to change security levels are presented to the user, as below:
Set Personal Security Level

The user then clicks on their desired level of security from the list (these values represent what % chance there is of another person logging on as you), and their personal threshold is re-calculated and saved.