Testing with Speech-to-Text Technology
Speech-to-text (STT) technology transcribes a student’s spoken words into text for question responses in TDS. Students with the appropriate accommodations may use STT assistive technology while taking tests. It is recommended that students use the embedded STT tool in TDS. The Texas Assessment Program can track and gather data about students’ embedded tool use (for more information about the embedded STT tool, refer to the Test Delivery System User Guide). If a third-party STT tool still needs to be used, please reference the following table.
Students using STT technology will need to use a headset while testing. Any wired headset with a 3.5mm or USB connection should be compatible.
The table below provides a list of third-party STT applications that can be used in TDS.
Third-Party STT Applications | ||
Product | System Requirements | Additional Details |
Dragon Naturally Speaking—Windows Supported Version: 16 Professional | Windows 8.1, 10; Server 2012 R2, 2016 R2 |
|
Windows Built-in Speech Recognition Supported Version: 8.0 | Windows 8.1, 10, 11; Server 2012 R2, 2016 R2 |
|
Supported Version: 5.4.29 | Windows 8.1,10, 11; Server 2012 R2, 2016 R2 |
|
Supported Version: 12.0.77 | Windows 8.1, 10. 11; Server 2012 R2, 2016 R2 |
|
Configuring Speech-to-Text Applications
Some applications listed in the table above require additional configuration to prepare for online testing. Necessary configurations are described in the following sections. Some applications send data to the cloud for processing by default. Where noted, this should be disabled to ensure the security of test data.
After you configure an application, CAI strongly recommends testing that application on a practice test administered through the Secure Browser prior to using it for operational testing.
Dragon Naturally Speaking 15 Home or Professional Individual for Windows
Necessary configurations for Dragon Naturally Speaking can be made from the Options dialog box, which is accessed from the Tools dropdown menu on the DragonBar.
From the Commands tab (refer to Figure 1), uncheck the following settings:
Enable launching from the Start menu
Enable launching from the desktop
Enable E-Mail and Calendar commands
Enable Cut shortcut commands
From the Miscellaneous tab (refer to Figure 2), uncheck Use the Dictation Box for unsupported applications.
In the Scheduled Tasks tab of the Administrative Settings window (refer to Figure 3), perform one of the following tasks:
In Dragon 16, uncheck Help us improve Dragon.
Setting Up User Profiles for Students in Dragon
Dragon Naturally Speaking requires each student to use a specific User Profile. You will need to create User Profiles and ensure that Dragon Naturally Speaking is set to the proper profile for each student prior to testing. When creating a profile, you can select the student’s age range, language, and accent, as well as set the audio input devices. Students will then read aloud prompts that Dragon Naturally Speaking uses to learn their voice. After creating profiles, you can select Profiles in the Dragon toolbar, then click Open User Profile to switch between User Profiles (refer to Figure 4). For more information about creating User Profiles, refer to the Dragon Naturally Speaking—Windows website.
Windows Speech Recognition
Prior to testing day, the Windows built-in Speech Recognition application must be set up on each testing device that will be used by students who require STT. The application can be set up through the Windows Control Panel. Users should set the device to not send data to Microsoft for improvement so that secure test data are not sent to the cloud. During setup, students must complete Speech Recognition Voice Training for optimal performance.
To prevent Windows from sending data to the cloud, go to Start > Settings > Privacy > Diagnostics & Feedback and mark the Basic radio button in the Diagnostic Data section. Then select the Speech tab and set the Online Speech Recognition toggle to Off.
While Windows Speech Recognition is processing speech, the testing site may slow down. If transcription is not needed for testing, ensure that Windows Speech Recognition is not listening in order to prevent this slowness from interfering with the test performance.
WordQ SpeakQ (WordQ5)
To minimize security risks, WordQ SpeakQ (also known as WordQ5) includes an Exam Mode feature, which can be enabled through the application’s settings. Exam Mode requires a time limit of 1–12 hours to be set. Please note, this does not eliminate all security risks, and once Exam Mode has been set, it cannot be disabled until the configured time has run out.
To turn on Exam Mode, click the Options icon, and select Exam Mode. In the dialog pop-up window that appears, you can allow and restrict the Word usage examples and Single words added by the user including topic words features. You can also set the exam time limit at the bottom of the window.
Read&Write for Windows
Read&Write has an Exam Mode that can be used to turn off features for a single student on his or her particular testing device. When Exam Mode is enabled, the student will have access to only the selected features on the toolbar and certain speech settings, including Voice Selection, Speed, Pitch, and Speak As I Type (the full settings menu will not be accessible).
To use Exam Mode, run Read&Write, click the Settings button in the top-right corner, and click Show more settings. In the Find a Setting field, type adminsettings. Enter and confirm a password to grant access on this computer. When logged into administrator settings, click the Select your features tab and select which features you would like to enable on the student’s toolbar. Enable the Use Exam Mode now toggle to start Exam Mode, then close the Read&Write menu to start the exam.