Research Application for WSRToolkit

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic

   I am new to the world of speech recognition, so I am hoping some of the members of this message board can steer me in the right direction.
I am currently finishing my Ph.D. in educational psychology at the University of California, Riverside.  I am conducting a study involving the application of speech recognition technology to a certain type of educational testing.  For my study, I need the software to facilitate multiple voice profiles, each of which can be quickly trained to a child's voice pattern (as young as age 8).  WSR allows multiple profiles, but the training process is too difficult for a young child to complete (the vocabulary is too high and the passages are too long).  Nuance's software seems to accommodate young children, but it does not allow multiple voice profiles.  Would the WSRToolkit provide the functionality I need to adapt WSR to the needs of my study (via the "scan" / "train to text" function)?  For this particular application, I need the software to distinguish between correct and incorrect pronunciations of certain words.  It works fine for my own voice on my own profile, but the "out of the box" configuration for a new user is inadequate.

Any thoughts or suggestions would be greatly appreciated!

Thanks,
Robert

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
rgcrosby Members Profile Find Members Posts Member Joined: 22/Sep/2011 Location: United States Status: Offline Points: 1	Post Options Post Reply Quote rgcrosby Report Post Thanks(0) Quote Reply Topic: Research Application for WSRToolkit Posted: 23/Sep/2011 at 6:59pm
	I am new to the world of speech recognition, so I am hoping some of the members of this message board can steer me in the right direction. I am currently finishing my Ph.D. in educational psychology at the University of California, Riverside. I am conducting a study involving the application of speech recognition technology to a certain type of educational testing. For my study, I need the software to facilitate multiple voice profiles, each of which can be quickly trained to a child's voice pattern (as young as age 8). WSR allows multiple profiles, but the training process is too difficult for a young child to complete (the vocabulary is too high and the passages are too long). Nuance's software seems to accommodate young children, but it does not allow multiple voice profiles. Would the WSRToolkit provide the functionality I need to adapt WSR to the needs of my study (via the "scan" / "train to text" function)? For this particular application, I need the software to distinguish between correct and incorrect pronunciations of certain words. It works fine for my own voice on my own profile, but the "out of the box" configuration for a new user is inadequate. Any thoughts or suggestions would be greatly appreciated! Thanks, Robert

mmarkoe_admin Members Profile Find Members Posts Admin Group Joined: 16/Jul/2008 Status: Offline Points: 331	Post Options Post Reply Quote mmarkoe_admin Report Post Thanks(0) Quote Reply Posted: 24/Sep/2011 at 11:27am
	Robert, Dragon NaturallySspeaking does allow for multiple profiles. However, the license for the software as per user and does not legally permits multiple users RE license on the same software package. The Windows Speech Recognition although it will allow multiple users to create individual profiles, the software was developed as a business speech model. I’m not sure it will accommodate young voices with any degree of accuracy. The WSRToolkit has some features that should meet the requirements you have. The Manage Profiles feature allows you to easily switch between profiles and create new ones. You can certainly use this scan and train functions to increase accuracy. You can also insert any text into the train feature. You can also have a standard document to insert using the Scan feature. The WSR toolkit is licensed per user at $25.99. If you’re going to have more than 10 users, contact us and we’ll work out a discount plan. Marty Markoe, eMicrophones, Inc.