Why is there a difference between the fast and slow pronounciation?
As I hear it: nouns ending with ة are pronounced -atun when said fast. But then when you press the turtle, you suddenly only hear the -a ending. Why is that?
The point is that Arabic words have pausal forms, to be used before pauses, and full forms, to be used when other words follow. So, the full form of words ending in ة ends in -atun, or -atan, or -atin, or -atu, -ata, -ati without nunation, according to grammar rules which are not explained in this course, at least not yet. The pausal form of ة-words ends always in -a.
In fast mode, the TTS engine tries to guess the full forms: it often gets the vowels wrong but the -t- is audible. In “turtle” mode, it puts a stop after each word, forcing the use of the pausal form.
Hope this clarifies the situation.
My impression is that the normal speed TTS is pronouncing the whole phrase, and adding the case endings it thinks should be there, whereas the turtle is pronouncing each word individually, so exactly what is there.