v1.4.1 patch for v1.4.0 CPU users.
Replace the v1.4.0 files with the ones in this .zip.

For just the v1.4.1 update specifically, delete the <.exe>/resources/app/assets folder first, and use the new one freshly included in this patch (not critical).

Just the xVASynth tool - download the separate voice models from the Optional section below. If you already have this from other games, you only need to download the voice models.

CPU only version

See the changelog in the description for news.

Voice model - Female Altmer, Bosmer, Dunmer
Model: FastPitch1.1+HiFi
Voice ID: ob_femalealtmer_bosmer_dunmer

Voice model - FemaleAltmerBosmerDunmer
Voice ID: ob_femalealtmer_bosmer_dunmer

Voice model - FemaleArgoniansKhajiit
Model: FastPitch1.1+HiFi
Voice ID: ob_femaleargonianskhajiit

Voice model - FemaleArgoniansKhajiit
Model: FastPitch1.1+HiFi
Voice ID: ob_femaleargonianskhajiit

Voice model - FemaleBretonsImperials
Model: FastPitch1.1+HiFi
Voice ID: ob_femalebretons_imperials

Voice model - FemaleBretonsImperials
Voice ID: ob_femalebretons_imperials

Voice model - FemaleNords
Model: FastPitch1.1+HiFi
Voice ID: ob_femalenords

Voice model - FemaleNords
Voice ID: ob_femalenords

Voice model - FemaleRedguards
Model: FastPitch1.1+HiFi
Voice ID: ob_femaleredguards

Voice model - FemaleRedguards
Voice ID: ob_femaleredguards

v1.4.1 patch for v1.4.0 GPU+CPU users.
Replace the v1.4.0 files with the ones in this .zip.

For just the v1.4.1 update specifically, delete the <.exe>/resources/app/assets folder first, and use the new one freshly included in this patch (not critical).

Just the xVASynth tool - download the separate voice models from the Optional section below. If you already have this from other games, you only need to download the voice models.

GPU+CPU version

See the changelog in the description for news.

Voice model - Male Altmer Bosmer Dunmer GSaint
Bespoke Tacotron2 pre-training
Bespoke HiFi-GAN vocoder
Voice ID: ob_malealtmer_bosmer_dunmer_gsaint

Voice model - Male Argonians Khajiit
Bespoke Tacotron2 pre-training
Bespoke HiFi-GAN vocoder
Voice ID: ob_maleargskhaj

Voice model - Male Redguards
Model: FastPitch1.1+HiFi
Voice ID: ob_maleredguard

Voice model - MaleAltmerBosmerDunmerGSaint
Voice ID: ob_malealtmer_bosmer_dunmer_gsaint

Voice model - MaleBreton
Model: FastPitch1.1+HiFi
Voice ID: ob_malebreton

Voice model - MaleBreton
Voice ID: ob_malebreton

Voice model - MaleRedguard
Bespoke Tacotron2 pre-training
Bespoke HiFi-GAN vocoder
Voice ID: ob_maleredguard

2.0.2 patch. See the changelog for what's changed/fixed.

Copy and replace these files over the top of v2.0.0+, by moving the "resources" folder into the folder where the .exe file and the "resources" folder are.

2.0.3 patch. See the changelog for what's changed/fixed.

Copy and replace these files over the top of v2.0.0+, by moving the "resources" folder into the folder where the .exe file and the "resources" folder are.

v2.0.4 patch. See the changelog for what's changed/fixed.

Copy and replace these files over the top of v2.0.0+, by moving the "resources" folder into the folder where the .exe file and the "resources" folder are.

v2.0.6 patch. See the changelog for what's changed/fixed.

Copy and replace these files over the top of v2.0.0+, by moving the "resources" folder into the folder where the .exe file and the "resources" folder are.

APPLY THE UPDATE PATCH FOR FIXES

Main app. Unified GPU and CPU build. Extract anywhere. Check your file paths in settings if models don't show up.

Download voices from the optional section. Read through the description for instructions and more details

For version v1.1.0 and up. You don't need this if you have the two WaveGlow .pt files downloaded already.
Install this with either xVA version you use. The two .pt files need to go:

<the .exe>/resources/app/models/<here>