Speech Recognition (SR) general info

TO THE TEACHER:

The Speech Recognition technology used in DynEd courses is state of the art. As a company dedicated to quality education, DynEd wants to ensure that teachers are aware of both the strengths and limitations of the technology at this time. The following points are important to keep in mind when guiding your students:

1. The Speech Recognition activities are extremely useful for speaking practice and for fluency development. However, the level of recognition is not sufficient for detailed pronunciation tutoring. Though the course may help to improve overall pronunciation, it is not a pronunciation course.
2. It may take some time for students to get used to the proper use of the microphone and to speaking within the time limits allowed. With practice there is no problem.
3. The microphone should be placed approximately 1 centimeter from the student's mouth. Too close or too far will reduce the performance.
4. The recognizer works best when it is working with a single student in a lesson. If several students take turns with the microphone, the recognizer may not function well, since the voice levels for each student vary. If students wish to switch, they should go to the speech pull-down menu and click on Reset or they should exit from the lesson and start again.

TEACHING SUGGESTIONS

If a student is having difficulty with speech, the following suggestions may be helpful:

1. Vary the speed of the sentence.
2. Try speaking in groups of words (e.g.,. "he's going" instead of "he is going"). If that doesn't work, try separating the words (e.g., "Where are you from?" instead of "Where're you from?").
3. Try varying the stress of troublesome words. For example, "fifteen" can be pronounced with two stress patterns: "fifTEEN" and "FIFteen." The second pronunciation can be confused with the word "FIFty."
4. Help students find patterns of words that cause them difficulty. For example, "her" and "are" both end with an "r" sound. Then students should listen to that "r" sound when spoken by the native speaker.
5. Please note that if the Expert level is checked in the Speech pull-down menu, the student may have more difficulty. At this level, the recognizer will be stricter. Typically, students score 10-15% less when the Expert level is checked.
6. A clear speaking native speaker will generally be recognized at an 85-95% rate when speaking at a moderate speed using the Expert level. For beginning students with poor pronunciation, the recognition rate may be 60-80% at first. With practice they will improve fairly rapidly at the non-expert level. When they feel confident, they should try the 'expert' level which will give them more detailed feedback.
7. With all Speech Recognition activities, students should listen to their own voices. This is recorded automatically. Click on the Headphone button to hear the student's recorded voice. Then click on the Repeat button to hear the native speaker. You may need to experiment with this at first until you learn how to do it.

TO THE STUDENT:

The Speech Recognition feature of this courseware is a fun and useful way to practice speaking. It will help you improve your speech articulation and fluency, while at the same time reinforcing important language structures and vocabulary. One way to prepare for Speech Recognition activities is to first use the Speech Record feature of the courseware in each of the presentation lessons. This allows you to record your voice, listen to the playback, and then compare it to the native speaker. If you are not confident with the language, the Speech Recognition activities may be difficult and frustrating, since they require a degree of speed and clearness for recognition to be successful. It is therefore important to master other lessons first.

Need more help? Get support and provide details about your issue.