2023-10-17 11:17:44 +02:00
2023-07-25 20:05:24 +02:00
2023-10-17 11:17:44 +02:00
2022-11-05 22:10:57 +01:00
2022-12-04 19:04:17 +01:00
2022-11-05 22:10:57 +01:00
2022-12-17 20:15:44 +01:00
2023-10-05 14:14:53 +02:00
2023-06-21 19:31:21 +05:00
2022-11-28 22:35:10 +01:00
2023-07-25 20:03:51 +02:00
2023-10-05 14:14:53 +02:00
2022-11-07 21:44:29 +01:00
2022-11-05 22:10:57 +01:00
2023-10-12 15:18:00 +02:00

WiVRn

License: GPL v3 CI

OpenXR streaming application

WiVRn lets you run OpenXR applications on a computer and display them on a standalone headset.

Installation

Server (PC)

Dependencies

WiVRn requires ffmpeg (for AMD/Intel), cuda (Nvidia), x264 (software) for video encoding. Other dependencies include avahi-client, eigen3, libpulse, nlohmann_json.

Note that some distributions don't ship h264/h265 support in their mesa builds. On Fedora this can be solved by replacing mesa:

dnf swap mesa-va-drivers mesa-va-drivers-freeworld

Compilation

From your checkout directory

cmake -B build-server . -GNinja -DWIVRN_BUILD_CLIENT=OFF -DCMAKE_BUILD_TYPE=RelWithDebInfo
cmake --build build-server

# Set WiVRn as the active OpenXR runtime, delete ~/.config/openxr/1/active_runtime.json after you are done using WiVRn
mkdir -p ~/.config/openxr/1/
ln --relative --symbolic --force build-server/openxr_wivrn-dev.json ~/.config/openxr/1/active_runtime.json

Client (headset)

At this early stage of development, we do not have stable releases. Download apk from Releases. Install with adb (headset connected to PC), developer mode must be activated on Quest.

adb install WiVRn.apk

Usage

On the computer, run wivrn-server, from checkout directory

build-server/server/wivrn-server

Then, on headset, launch WiVRn from the App Library, in "unknown sources" section.

You will briefly have a "Waiting for connection" screen, followed by "Waiting for video stream". Now on your computer you can run an OpenXR application, and it will show on your headset, enjoy!

TODO

  • Sound support (recording)
  • Improve sound playback latency
  • SteamVR support with OpenComposite
  • Latency improvement
  • Application launcher
  • Support more headsets

Credits

WiVRn uses the following software:

Languages
C++ 90.8%
QML 2.9%
CMake 2.7%
Python 1.8%
GLSL 0.8%
Other 0.8%