Nvv4l2decoder fail to decode the h264 3840x2560 video which encoded by nvv4l2h264enc

**• Hardware Platform (Jetson / GPU) NVIDIA GeForce RTX 2060
**• DeepStream Version 6.2.0
**• TensorRT Version 8.6
**• NVIDIA GPU Driver Version (valid for GPU only) 530.41.03

I use nvv4l2h264enc to encode 264 video, and use nvv4l2decoder to decode it. I found that if the video size is 1280-720, it can work. If the size is bigger like 3840-2560, the nvv4l2decoder will fail to work.

My test commands are here:

CAN work:

gst-launch-1.0 videotestsrc ! ‘video/x-raw, width=(int)1280, height=(int)720, format=(string)NV12, framerate=(fraction)30/1’ ! nvvideoconvert ! nvv4l2h264enc ! h264parse ! qtmux ! filesink location=filename_h264.mp4 -e
gst-launch-1.0 filesrc location=filename_h264.mp4 ! qtdemux ! h264parse ! nvv4l2decoder ! nvvideoconvert ! autovideosink sync=false

CAN NOT work:

gst-launch-1.0 videotestsrc ! ‘video/x-raw, width=(int)3840, height=(int)2560, format=(string)NV12, framerate=(fraction)30/1’ ! nvvideoconvert ! nvv4l2h264enc ! h264parse ! qtmux ! filesink location=filename_h264.mp4 -e
gst-launch-1.0 filesrc location=filename_h264.mp4 ! qtdemux ! h264parse ! nvv4l2decoder ! nvvideoconvert ! autovideosink sync=false

The error messages:

Setting pipeline to PAUSED …
Pipeline is PREROLLING …
nvbufsurface : Unsupported output format (36)
^Chandling interrupt.
Interrupt: Stopping pipeline …
ERROR: pipeline doesn’t want to preroll.
Setting pipeline to NULL …

More info:

  • the 3840-2560 video which encoded by nvv4l2h264enc can be played on other applications, such as ubuntu’s Videos application.
  • the nvv4l2decoder can decode other 3840-2560 video, which I encoded by other encoder. nvv4l2decoder can not work only on the video encoded by nvv4l2h264enc!

I need use nvv4l2h264enc/nvv4l2decoder to work on 3840*2560 video. Can someone give me solution to this issue?

I can not reproduce the failure. Can you send us the “filename_h264.mp4” file you generated?

Fiona,

Please try this filename_h264.mp4.

Can you install the correct dependencies for DeepStream 6.2? Quickstart Guide — DeepStream 6.3 Release documentation

I use the DeepStream docker to test the issue again. The problem is still the same.

The more info:
docker from: Quickstart Guide — DeepStream 6.3 Release documentation

command to download and run the container:

docker pull nvcr.io/nvidia/deepstream:6.2-triton
docker run --gpus “device=0” -it --rm -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY -e CUDA_CACHE_DISABLE=0 nvcr.io/nvidia/deepstream:6.2-triton

The components version in docker:

root@8e694d757581:/tmp# deepstream-app --version-all
deepstream-app version 6.2.0
DeepStreamSDK 6.2.0
CUDA Driver Version: 12.0
CUDA Runtime Version: 11.8
TensorRT Version: 8.5
cuDNN Version: 8.7
libNVWarp360 Version: 2.0.1d3

Only the driver version has different with OS Compatibility version: my driver version is 525.105.17, not the 525.85.12.

Can you install driver 525 as the instruction here: Quickstart Guide — DeepStream 6.2 Release documentation (nvidia.com)

I tested it in docker, and use driver 525.85.12. The problem is still same. 1280x720 is ok; 3840x2560 fail.

Can anybody else reproduce this issue?
I tested on multiple hardware and software combinations. All have this issue.

We don’t have RTX2060 now. I’ve tested with RTX 2080, RTX 4090 and A6000, the pipeline works well. Please make sure you installation follows the compatibility Quickstart Guide — DeepStream 6.2 Release documentation (nvidia.com)

It is very strange. I tested this issue on RTX2060, 3090, 4080.
Did you test the pipeline on 3840x2560 video? It is no problem on small resolution.

I am sure I have installed the compatibility Quickstart Guide — DeepStream 6.2 Release documentation

Yes, I tested with your pipelines with 3840x2560 resolution, the only change is to replace the autovideosink to nveglglessink.

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

I’ve borrowed one RTX 2060 for testing the case, the error can be reproduced. So it is the HW limitation. Please consider to use the DeepStream 6.2 GA with the recommended GPUs.Quickstart Guide — DeepStream 6.2 Release documentation

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