Creating speech with totally different rhythms and pauses makes the sound extra human, in keeping with an evaluation by an AI skilled on speech from YouTube and podcasts.
Most AI text-to-speech techniques are skilled on datasets of spoken-to-speech, which may trigger the output to look contrived and one-dimensional. Extra pure speech usually displays all kinds of rhythms and patterns to convey totally different meanings and feelings.
Now, Alexander Rudnick at Carnegie Mellon College in Pittsburgh, Pennsylvania.
#skilled #YouTube #podcasts #talks #ums #ahs