Question on jetson.utils.videoOutput


I want to do depth estimation inference using videoSource and videoOutput APIs from jetson.utils. I am streaming video file using videoSource and getting an output video with 15 FPS and 2 times longer than the input video, because my model’s throughput is 15 qps. My expectation was to write an output video file with 15 FPS, but with the same length as the input video (or in other words process and write only a part of the input frames). How can I achieve this through videoSource and videoOutput APIs?


Hi @tigran.khachatryann1, the videoOutput interface outputs/saves video at the same rate that the frames were submitted to it (so if your program is running at 15 FPS, the video will be saved at 15 FPS). Perhaps you could use some other GStreamer script to post-process the video at the correct framerate or use some other API to save it (like OpenCV)

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