GStreamer: How to use v4l2convert instead of videoconvert

Autonomous Machines Jetson & Embedded Systems Jetson Xavier NX
1

I’m using VideoCapture and GStreamer to capture frames from a camera and output using a VideoWriter. Below are my input and output pipelines:

#VideoCapture Input Pipeline
"v4l2src device=/dev/video2 io-mode=2 ! video/x-raw,format=YUY2,width=2560,height=720,framerate=60/1 ! nvvidconv ! \
        video/x-raw(memory:NVMM) ! nvvidconv ! video/x-raw,format=BGRx ! videoconvert ! video/x-raw,format=BGR ! appsink drop=0"

#VideoWriter Output Pipeline
"appsrc ! queue ! videoconvert ! video/x-raw, format=RGBA ! nvvidconv ! \
           video/x-raw(memory:NVMM) ! nvegltransform ! nveglglessink "

I noticed that the conversion using videoconvert is quite slow. After doing some reading, I found that some people using v4l2convert and v4l2h264enc as an alternative. How would I incorporate that in my pipeline. Note that I’m running this in a python script on a Jetson Xavier NX.

3

Hi,
For converting to BGR format you would need to use software converter videoconvert. Please refer to discussion in
[Gstreamer] nvvidconv, BGR as INPUT - #2 by DaneLLL

4

Hi,
Thanks for the response. I’m aware of using videoconvert however I understand that because it’s a CPU conversion it causes my frame rates to be much slower. I found this Raspberry Pi forum where the user used a v4l2encoder/decoder and v4l2converter to solve this issue. Would it be possible to replicate this in my pipeline on the Jetson NX?
https://www.raspberrypi.org/forums/viewtopic.php?t=245852

5

Hi,
The implemented plugins on Jetson platforms are nvv4l2h264enc, nvv4l2decoder, nvvidconv, so the pipeline is not applicable. Hardware devices are different.