Vowel reduction in forced alignment

[Qianchu L]

It seems to me that in order to select between the HMM branch for a full vowel and the branch for a reduced vowel, we need to do inference during forced alignment. So, the states that emit a full vowel and the states that emit a reduced vowel should be pre-trained using external data?

[Simon]

We certainly do need to train models of full and reduced vowels before we can used forced alignment to choose which one is the most likely label to place on a particular segment in the database.

We’ll look at how to train those models in the lecture.

[Author]

Posts