This page was last updated on 09 January 2024, 5:31PM
Changelogs
Version 3.0
updated FemaleChild and MaleChild (English) to version 3.0 trained with XVA trainer
Version 2.0
For version 2.0 I will re-train the models with the updated Fastpitch 1.1. This makes them compatible with the new features in xVA-Synth 2.
The fist updated voice is FemaleChild
I highly recommend you to use the ARPAbet pronunciation when you encounter words that are not correctly pronounced. This greatly helps to improve the quality in my experience.
Version 1.4
updated English versions. should be less "thin" and less mispronounced words.
Version 1.3
added French models to optional files
updated German models
Version 1.2
added German voice models to optional files
Version 1.1
retrained voices using SSE High Quality voices
added the HiFi-GAN modells
added the preview audio files (right-click on name to hear preview)
