How to use nvtee in a pipeline to convert single to multiple streams , it has 3 always pad but no request pad to link 1 to N pipe linking

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson AGX)
**• DeepStream Version 4.0.1
**• JetPack Version 4.2.2

How to use nvtee in the pipeline to link source to perform inference, and optical flow separately by using nvtee to convert into multiple streams.

pipeline with tee:
gst-launch-1.0 filesrc location= /home/nvidia/deepstream/deepstream_sdk_v4.0.1_jetson/samples/streams/sample_720p.mp4 ! qtdemux ! h264parse ! nvv4l2decoder ! m.sink_0 nvstreammux name=m batch-size=1 width=1920 height=1080 ! nvof ! tee name=t ! queue ! nvofvisual ! queue ! nvvideoconvert ! videoconvert ! fakesink sync=false t. ! queue ! nvinfer config-file-path= /home/nvidia/users/vijaya/cplusplus/VOCA-Deepstream/server/config/dsmeta_pgie_config.txt ! nvvideoconvert ! nvdsosd ! nvvideoconvert ! nvv4l2h264enc insert-sps-pps=true ! h264parse ! queue ! rtph264pay mtu=65000 ! fpsdisplaysink text-overlay=false video-sink=“udpsink” host=10.201.1.191 port=5600 -v

I want to do similar with the help of nvtee, can anyone help me to solve this?

3

Hi,
In DeepStream SDK, we use tee plugin. You may see similar code in

/opt/nvidia/deepstream/deepstream-4.0/sources/apps/sample_apps

nvtee is specific to running nvgstcapture-1.0 app. May not work properly in DeepStream SDK. Please note this.

1 Like
4

Thank you DaneLLL