Ein interessanter Post aus deinem Link:
"Jitter comparison is meaningless.
The IMUs provide a motion vector as an input to a macro motion estimator. With Lighthouse the estimate is updated with position locks as they arrive (usually 100Hz) so jitter is the result of a delta between the estimate and the sampled positions.
With inside-out there’s still IMU-based motion estimation but the estimate is updated with positions obtained by correlating markers found in the camera images.
When the headset isn’t moving, there is no data coming from the IMU so output positions are coming exclusively from the respective absolute positioning systems.
With Lighthouse the causes of jitter are the same regardless of what the headset is doing so you’ll see some while it’s idle. With camera tracking there’s little reason for the position to change while the camera images are the same.
But when the headset is moving - ie. the IMUs are actually doing something useful and you’re not just standing there doing nothing - the story changes. Now jitter will be higher in whichever headset has the less accurate absolute positioning system.
And that is definitely not Lighthouse.
Correlating point clouds is not nearly as accurate as timing lasers. It’s always an estimation and there’s always some wiggle room so jitter is reduced through averaging which smooths out the motion in exchange for latency and accuracy.
That's for the headset. Rift S controllers are outside-in and again the IMU estimates are updated with data from the cameras, but this time the distance and vector are calculated, and the pose estimated, based on camera views of the LEDs. Again, estimations upon estimations which are then smoothed, resulting in less jitter.
If you want to perform real tests, buy a robot arm. Then you can test things that matter, like latency and positional accuracy."
@mospider
Ich glaube die Leute, die Probleme hatten, lagen bei 69. Aber gut, daß es bei dir noch klappt.