1 paper
Lantian Li, Dong Wang, Chao Xing +2
The popular i-vector model represents speakers as low-dimensional continuous vectors (i-vectors), and hence it is a way of continuous speaker embedding. In this paper, we investiga…