I want to get the captured YUV data from the pipeline through GStreamer;
1.If I use the following method with nvvidconv-nvmm,it can’t get valid yuv data,the target-file will only crease 10 or 11k each time;
How about without nvvidconv element?
it must have nvviconv element，otherwuse it can’t get valid yuv data,the target-file will only crease 10 or 11k each time;
Possible the filesink can’t take memory:nvmm as input.
the element:appsink is in the same condition as filesink is；
I think Nvidia is one of the greatest companies in the world，i confirm it has the ability to deal it well，Why not improve it
What’s your purpose?
Why not just use second command?
if i use the second command，the cpu-occupy will increate 15 percent;
if i use the first command,the cpu-occupy only 3percent
As shown in the figure below,queue-2-thread was 18 percent in second mode;
Any possible add nvv4l2h264enc to the pipeline for it.
Not sure if nvivafilter element help on your case.
the nvivafilter-element’sink is also nvmm-format ，it can‘t generate valid target-file。
and i can’t add any encoderenc because i have anthoer problem：the high-cpuload when use encoding high resolution yuv;
I don’t think there an solution for your case, even we implement filesink support nvmm that could increase CPU load by the element.
