For  Listen and Practice (listen and repeat) modes to work properly, ReadAloud will mark up your model audio so that it pauses and highlights the text currently being spoken. 


When using TTS (machine voice) as model audio


Model audio mark up is done automatically when using most TTS (text to speech) voices. 

(NB However some voices do not support automatic mark up. Those voices have a ! mark after them in the list.)


ReadAloud will guess at the correct locations to put break points. But you can tweak the break points and set them manually from the model audio tab. To do this, visit the the activity's model audio tab.


To tweak the break point, tap the space or word before the space to toggle the break point on or off.




When using a recorded speaker as model audio


Model audio mark up is also done automatically when you are using an audio-recording/uploaded-audio as model audio. However there will be a delay while that audio is transcribed. After recording or uploading wait 5 mins and refresh the page to see the automatically created breaks. 


When the break timing is not working correctly

Sometimes the break timing is not correct, and the audio will continue after the break, or stop too soon. If the audio is fast and there is not much silence between words this can happen. Adding commas and periods is generally a good idea when creating ReadAloud passages. It creates small pauses in the audio. 


But sometimes it is really out of sync. In that case try:

i) Regenerate the model audio breaks THEN

ii) Save the model audio


If that does not work, go back to the activity settings page, and change the voice and save the activity. Poodll will regenerate the audio and breaks. After that you can go back to the original voice if you wish.


If none of that works you may need to set the break times manually, as described below.


Setting the break times manually


Use "manual break timing" if:


a)  the sync between breaks and the audio location is not precise enough. 

OR

b) the model audio breaks are not created automatically.

This can happen if the TTS voice does not support time stamps. This is currently true for some languages, including Hebrew, Tamil and Telegu.



This way requires the teacher to play the audio and while playing tap the location in the passage that matches the audio playback point, to toggle the breaks. To begin this, 

  1.  Check "manual break timing" 
  2.  Begin playback by pressing the audio player play button
  3. Tap the break point when the audio reaches it
  4. Tap the next break point when the audio reaches it ..


Finally do not forget to "Save Model Audio Mark Up"



This video explains how to set up model audo breaks manually.

https://www.kapwing.com/videos/6415af1c1dd1fe030e288d09