Output display lags which is far from real time

• Hardware Platform (Jetson / GPU) - GPU
• DeepStream Version - 6.1
• JetPack Version (valid for Jetson only)
• TensorRT Version - 8.4.1
• NVIDIA GPU Driver Version (valid for GPU only) - 520.16.xx
• Issue Type( questions, new requirements, bugs) - question

My deepstream pipeline runs at 10 fps with four rtsp sources and a bodypose2d model. I see a lag in the output display which is far from real time. What I want to achieve is to make this output real time. Are there any suggestions to make this run at near real time?
I know there is a property interval in the nvinfer. But I cannot skip too many frames as it would affect the accuracy of other business logic in the downstream. Is there any other way to drop frames that are older than few millisec, so that only latest frames flow through the buffer.

Hi Team,

Any lead on this please?


I set both NVDS_ENABLE_LATENCY_MEASUREMENT and NVDS_ENABLE_COMPONENT_LATENCY_MEASUREMENT, but do not see any additional log in the terminal. Attached my pipeline dot for your reference as the converted png file size is too big to attach.
pipeline.swapped.dot (188.2 KB)

DeepStream SDK FAQ - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums

I’m still analyzing the measured latency. Will report the status here once I have a finding.

There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.

@arivarasan.e Can you let us know whether you still need support on this topic?

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.