Voice assistants like Alexa convert written words into speech using text-to-speech systems, the most capable of which tap to verbalize from scratch rather than string together prerecorded snippets of sounds. Neural text-to-speech systems, or NTTS, tend to produce more natural-sounding speech than conventional models, but arguably their real value lies in their adaptability. They’re able to mimic the prosody of a recording, which refers to its shifts in tempo, pitch, and volume.
In a (“”) presented at this year’s Interspeech conference in Graz, Austria, Amazon scientists investigated prosody transfer with a system that enabled them to choose voices in recordings while preserving the original inflections. They say that it significantly improved on past attempts, which generally haven’t adapted well to input voices they haven’t encountered