Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU) Jenson TX2 and NX • DeepStream Version Deepstream 6.0 • JetPack Version (valid for Jetson only)
NX : 4.6.2 TX2: 4.6.0 • Issue Type( questions, new requirements, bugs)
The description of the problem is as follows: Using TX2 and NX to pull the same RTSP data stream, the frame rate basically does not change during the TX2 inference process, and the frame rate changes greatly during the NX inference process.
Set GST_DEBUG=2, compare the output of the two, and find that the following error is output during NX inference:
No clock-rate in caps!, dropping buffer cannot get clock-rate for pt 8
. I would like to ask, have you encountered any related problems and have any solutions?
thanks for the reply! In any case, within the same LAN, after many tests, the streaming media of the TX2 is normal, and the NX will skip frames almost every time. Can you provide some improvement suggestions to solve this problem? I will also read the rtpjitterbuffer plugin to find the answer.
Thank you for your reply, I used deepstream-app to do the following test。On the NX platform, close nvinfer, only pull the stream, and find the following errors, there is no similar error on TX2。
Thank you for your answer. Through subsequent tests, we found that the stream pulled from the camera does not have this problem, but the data pushed through easydarwin has problems.
