This update is mostly about voice. Japanese now reads aloud with a natural voice, Chinese can be read aloud for the first time, and every language plays back at a matching volume.
A natural Japanese voice
Japanese read-aloud now uses VOICEVOX rather than the general-purpose engine, so pitch accent and rhythm sound like spoken Japanese instead of a string of syllables. It is an optional download — the app asks first, and the previous voice keeps working until you fetch it.
Chinese read-aloud
Chinese can now be read aloud, covering Simplified Chinese. Like the Japanese voice, it arrives as an optional download.
Voices you can pick, at a volume that matches
- Each language now offers a small set of voices in Settings, and your choice survives a restart.
- Playback volume is evened out across languages. Chinese used to be noticeably quieter than English.
- The speak button now shows that it is working — the first play of a line takes a moment to synthesize, and the button no longer looks unresponsive while it does.
Steadier overlays
- Overlay boxes no longer change size on a screen that is not changing.
- Lines are no longer reordered between frames, which is what made short lines flicker and get translated again.
- Chinese source text now gets the same overlay sizing as Japanese and Korean.
Less waiting
Screen capture is faster, and lines now reach the Log while you are still reading them instead of waiting for the scene to change.
Also in this update
- Difficulty tags now cover Korean (TOPIK) and Traditional Chinese.
- Fixed Japanese screens with a lot of kanji being reported as Chinese.