V3 Streaming Translation Is Here
We've made a major upgrade to VOXSHIFT's streaming translation system.
What's New
V3 Streaming Translation
This is the biggest change in this update.
We've rebuilt the streaming translation system to provide more natural and stable real-time translation.
With the introduction of V3, the previous V2 system has been discontinued.
V3 brings significant improvements over V2, including more natural timing for when translation begins.
Choose V1 / V3 Separately
You can now select V1 or V3 independently for your own voice and the other person's voice.
Lighter Translation and Voice Generation
We've reduced processing requirements while maintaining quality as much as possible, making VOXSHIFT more comfortable to use.
We plan to continue optimizing performance in future updates.
Our current long-term goal is to make VOXSHIFT usable even on systems with 4 GB of VRAM.
Automatic Model Selection
VOXSHIFT can now automatically select appropriate models based on your PC environment.
※ Available only in the paid version. Model quality remains fixed in the free version, as before.
Live Startup Progress
While Live mode is preparing, the overall startup progress is now shown as a percentage.
You can now easily see how far the preparation has progressed before Live is ready.
New Conversation UI
We've introduced a redesigned conversation view that separates your speech and the other person's speech into upper and lower sections.
Improved Stability
We've improved stability when starting Live mode and during speech playback, along with fixes for multiple issues.
For Existing Users
Your existing settings will be preserved after the update.
You can try the new V3 Streaming Translation without changing your current setup.
If you're currently using V1, we encourage you to give V3 a try.
For new users, V3 is now the default streaming translation setting.