Speech

In addition to the emotional AI, the R2-D2 Vocalizer contains a speech AI powered by a dataset containing containing every vocalization uttered by R2 on film.

The speech AI ensures that every vocalization is unique, yet stylistically indistinguishable from R2's vocalizations in the films.

It also actively avoids vocalizing repeating sounds in quick succession. Even when vocalizing the same emotion repeatedly, the software steers clear of blatant repeats of the same voicing.

Improvisational vs Canonical Mode

In improvisational mode, the speech AI creates unique vocalizations in response to stimuli.

In canonical mode, the software will only produce vocalizations that have appeared, verbatim, in a Star Wars film. Instead of operating on a phoneme-level of granularity, the speech AI will switch to operating on a "sentence"-level. It will still chain canonical sentences where appropriate.

Improvisational mode is recommended unless strict canonical compliance is required for your application.