Text To Speech · published · build 22349401
New
Voice effects were revamped with a larger profile list, better-sounding processing, and a more efficient implementation.
Added Effect Intensity, so effects can be applied more subtly or more aggressively instead of always using a fixed full-strength version.
Added Voice Mood, which shifts the voice darker/warmer or brighter/happier.
Added Voice Energy, which changes drive, movement, and overall speech liveliness.
Mood and Energy influence each effect profile a little differently, so the same values can feel different depending on the selected effect.
Each effect profile now has recommended Intensity, Mood, and Energy defaults. The new Reset Profile button restores those recommended values.
Added effect intensity macro support with the format [effect:effectname,50].
Added mood macro support with the format [mood:0].
Added energy macro support with the format [energy:50].
UI
Reorganized the General and Voice settings tabs.
Added Mood, Energy, and Effect Intensity controls to clip preview/export as well.
Bug Fixes
@username no longer reads names from YouTube/Twitch.
Fixed an issue where the Piper downloader could fail on voices stored in subfolders.
Piper now uses an offline voice list to make voice loading faster.
Fixed a Piper file access error that could happen when changing voices.
Fixed WinRT being disabled on Windows versions where it should have been enabled.
Fixed Kokoro sometimes being detected as Piper.