Clustering speech utterances by speaker using Eigenvoice-motivated vector space models | IEEE Conference Publication | IEEE Xplore