Hi,
I am currently measuring the glass-to-glass latency of my system using a ffc-4p. I have two cameras outputting two h264 streams at 60 frames per second. The reported latency is approximately 25 milliseconds, which is as expected according to the documentation.
{'left': datetime.timedelta(microseconds=25649), 'right': datetime.timedelta(microseconds=25644)}
I am using a simple GStreamer pipeline to decode and display one of these streams. The total reported latency, including the previous one, is approximately 44 milliseconds (latency is measured in nanoseconds in this context). The pipeline latency is as follows:
Pipeline latency (live=True, min_latency=44692666, max_latency=18446744073709551615)
The GStreamer tracers have captured all of this data, with appsrc receiving the frames from the ffc-4p.
However, when I measure the glass-to-glass latency by framing a precise timer on a 160Hz screen, I find that the latency is approximately 115 milliseconds.
I am unable to account for the additional 70 milliseconds of latency. I would greatly appreciate any insights you might have regarding the potential causes of this discrepancy.
Additionally, I am interested in learning if there are any examples available that demonstrate how to display a frame with a latency of approximately 30 milliseconds.
Thanks
