I'm accessing eye gaze data in UE4 through the SRanipal plugin-in, but am having issues with repeated data. Since there seem to be two separate issues, I'll mention both.
First is that, according to logging in SRanipal_Eye_v2::UpdateData, not every frame of device data is received. From other threads (e.g. https://forum.vive.com/topic/5897-getting-verbosedata-at-the-fastest-rate-possible/) it seems like I should ignore the plugin and call GetEyeData_v2 directly from a worker thread operating at more than 120Hz, and use frame_sequence to identify unique frames.
However, even though I'm currently only receiving about 63 frames per second (the main thread tick rate?), the gaze data contains repeated values. In 9 out of 535 samples taken (excluding zeros - blinks?) gaze direction was repeated from the previous frame_sequence value, which in each case was smaller by 2. That is, the gaze direction in frame_sequence 3717 was identical to that in 3715. Is it to be expected that not every frame will have unique data? Should I mark such frames as containing no gaze data, as I assume this cannot arise by chance? If I manage to get every frame received, what fraction of repeated data can I expect? I assume in the observed cases that at least three frames in a row had the same gaze data, so presumably consecutive frames are more likely to contain repeated data.
Repeat gaze data
in VIVE Eye and Facial Tracking SDK
Posted
I'm accessing eye gaze data in UE4 through the SRanipal plugin-in, but am having issues with repeated data. Since there seem to be two separate issues, I'll mention both.
First is that, according to logging in SRanipal_Eye_v2::UpdateData, not every frame of device data is received. From other threads (e.g. https://forum.vive.com/topic/5897-getting-verbosedata-at-the-fastest-rate-possible/) it seems like I should ignore the plugin and call GetEyeData_v2 directly from a worker thread operating at more than 120Hz, and use frame_sequence to identify unique frames.
However, even though I'm currently only receiving about 63 frames per second (the main thread tick rate?), the gaze data contains repeated values. In 9 out of 535 samples taken (excluding zeros - blinks?) gaze direction was repeated from the previous frame_sequence value, which in each case was smaller by 2. That is, the gaze direction in frame_sequence 3717 was identical to that in 3715. Is it to be expected that not every frame will have unique data? Should I mark such frames as containing no gaze data, as I assume this cannot arise by chance? If I manage to get every frame received, what fraction of repeated data can I expect? I assume in the observed cases that at least three frames in a row had the same gaze data, so presumably consecutive frames are more likely to contain repeated data.
Thanks!
@Daniel_Y, @Marios Bikos