Abstract: We introduce a novel speech synthesis system, called NAUTILUS, that can generate speech with a target voice either from a text input or a reference utterance of an arbitrary source speaker.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results