When using the nvcompositor to compose three video streams into one the output video starts of as smooth video without lag, then gets progressively worse. The frame rate drops and the time difference between the input and output gets longer and longer (90 seconds of time delay, in one of my tests).
Would love some assistance.
Output of the line head -1 /etc/nv_tegra_release:
# R32 (release), REVISION: 3.1, GCID: 18186506, BOARD: t186ref, EABI: aarch64, DATE: Tue Dec 10 07:03:07 UTC 2019
Script:
SENSOR_W=1280
SENSOR_H=720
FLIP_METHOD=0
LEAKINESS=2
# Left overlap area
LEFT_END=$((SENSOR_W-184))
MID_START=200
# Right overlap area
MID_END=$((SENSOR_W-180))
RIGHT_START=180
# Width parameters
LEFT_W=$LEFT_END
MID_W=$((MID_END-MID_START))
RIGHT_W=$((VIDEO_WIDTH-RIGHT_START))
gst-launch-1.0 \
nvcompositor \
name=comp \
sink_0::xpos=0 \
\
sink_1::xpos=$LEFT_W \
\
sink_2::xpos=$((LEFT_W+MID_W)) ! \
\
nvoverlaysink sync=false \
v4l2src device=/dev/video2 ! \
"video/x-raw, format=(string)UYVY, width=(int)$SENSOR_W, height=(int)$SENSOR_H" ! \
nvvidconv flip-method=$FLIP_METHOD left=0 right=$LEFT_W top=0 bottom=$SENSOR_H ! \
queue leaky=$LEAKINESS ! \
comp. \
v4l2src device=/dev/video1 ! \
"video/x-raw, format=(string)UYVY, width=(int)$SENSOR_W, height=(int)$SENSOR_H" ! \
nvvidconv flip-method=$FLIP_METHOD left=$MID_START right=$MID_END top=0 bottom=$SENSOR_H ! \
queue leaky=$LEAKINESS ! \
comp. \
v4l2src device=/dev/video0 ! \
"video/x-raw, format=(string)UYVY, width=(int)$SENSOR_W, height=(int)$SENSOR_H" ! \
nvvidconv flip-method=$FLIP_METHOD left=$RIGHT_START right=$SENSOR_W top=0 bottom=$SENSOR_H ! \
queue leaky=$LEAKINESS ! \
comp.