It seems that each time a new Synth V AI voicebank is released, its data is used to improve the performance of all Synth V AI voicebanks. For example, in the run up to Solaria's release, updates have improved the high range of previous English voices, and vocal fry has started appearing on certain notes. Multilanguage is the other big, obvious addition! Voicebanks are able to learn new languages from the data of native voices. This means that the more AI voices we have, the better the software will become.