OVR Overlay Translator · beta · first seen
Version 2.4.10 is now released. This update includes the following improvements and new features:
[Update Details]
New Features
Added a "Preload Model" switch to reduce initial recognition wait time.
Image translation and speech translation can be enabled simultaneously, with independent control of the two switches.
Speech captions support color schemes and text outlines, with independent settings for desktop and VR.
Screenshot frame adjustment mode supports joystick panning.
Speech models can each specify endpoints and protocols, separating recognition and synthesis.
Optimizations
Optimized OCR recognition speed, shortening the time from trigger press to seeing the translation.
Optimized image transmission and character block depth calculation, with more significant improvements in long text scenarios.
Improved the fit stability when the tilt plane has few text blocks.
Double-clicking the trigger and wrist panel tabs only switch input focus, making operation clearer.
Bug Fixes
Fixed the issue of the translation floating in front of the original text and the offset becoming more obvious with distance.
Fixed the issue of individual text blocks jumping out and being closer to the original text.
Fixed the issues of the volume bar not moving and no longer recognizing text after device sleep reconnection, headphone plugging/unplugging, and the built-in mode selection translation engine failing after restarting.
Fixed issues of startup crashing upon exit, being forcibly killed after closing the debug console, and residual child processes.
| Depot | New manifest | Previous manifest |
|---|---|---|
| 4304621 | 7511017619718069935 | 5033434841057006050 |