Opengl-webrtc-streamer

The proj is based on cloud rendering.(云渲染)

1 Set Up

For the detailed enviroment setup, please refer to SETUP.md. Ubuntu and macos is recommended. And please note that the server file is from libdatachannel repo.

git submodule update --init --recursive
cmake -B cmake-build-debug
cd cmake-build-debug
make

before running the program, please make sure the server is set up.

cd client
python3 signaling-server.py
python3 -m http.server --bind 127.0.0.1 8080

then just run the program ./ors, and open http://127.0.0.1:8080/ in your browser, and press start button.

2 Parameters

Parameter	Function
-docker	Use in docker mode
-d	Dump video to check video
-gpu num	Use hardware for accelerate, e.g. -gpu 4 is videotoolbox. (0-cpu, 1-nvenc, 2-vaapi, 3-qsv, 4-videotoolbox)

3 Funcs

Lots of work still need to be done.

• Record OpenGL app screen and encode to H.264
• Rtmp Streamer, which can push the H.264 raw frame in buffer to server
• Webrtc
• Terminal controls transfer to server

3 Next Steps

• Use librtc(libdatachannel is the origin name) to send memory video.
• Enable browser control.
• Accelerate encode.(partial)
• Integrate with webgl rendering.
• Task scheduling on local and remote machine

Please note: the repo's license is MIT, but the 3rd_party/librtc is GPL.