They are waiting for your support, it looks like you are passing fields rather than frames
However, this doesn’t resolve the rest of the issues. Please read back what I reported and needs fixing on your side.
Hi Rudoplh,
We can observe the issue of H264i decoding and make a new topic for tracking:
https://devtalk.nvidia.com/default/topic/1066497/
For the issue ‘the disableDB=true is producing problems too’, is it also specific to H264i stream? Please share more information about this issue so that we can reproduce it. Thanks.
Hi Dane,
Testing video can be a complicated thing. I’m helping Jiang on the FFmpeg-jetson project. I’m developing a series of scripts that I will be happy to share with you, so the release of this SW is tested before is released.
Onething I’m still waiting respond from you is to release this open-source code into nVidia’s GitHub or GitLab sites, so we can help you with test and bug reporting and to avoid the costly waiting.
I will push my scripts here in a moment
Test_streams.zip (2.32 KB)
The previous script, will create a series of test streams, I have tested one by one using proffesional tools from Tektronix, Intel, and other to validate they are valid test.
Like this, I want to help you to create scripts that will test each parameter of the GStreamer components, This way we can have a clear picture of memory leaks, DTS/PTS issues, stream complaints etc…
Hi,
For memleak in decoding interlaced h264 stream through nvv4l2decoder, please apply attached patch to gst-v4l2 source, rebuild/replace libgstnvvideo4linux2.so, and try again.
0001-gst-v4l2dec-Fix-MemLeak-Fast-Playback-for-Interlace-.zip (1.68 KB)
Has this issue been fixed/included in jetpack 4.3? I also need to decode some interlaced .ts(h264 byte-stream alignment=nal) and have yet to successfully accomplish this.
Hi,
The fix is not in Jetpack 4.3. Please download gst-v4l2 source code to apply the patch and build libgstnvvideo4linux2.so
https://developer.nvidia.com/embedded/dlc/r32-3-1_Release_v1.0/Sources/T210/public_sources.tbz2