Open ronag opened 6 years ago
What is the theoretically lowest SDI->SDI throughput latency you can imagine, given GPUDirect etc.?
With black magic, 4 frames
Any lower, @jamesdwise ?
For the sake of the discussion https://github.com/CasparCG/server/issues/891 - current real-world end to end latency is currently said to be 8 or 9 frames regarding if low-latency is enabled or not.
Use shared OpenGL context to avoid the whole device->host->device ping pong