This TRD includes the following jupyter notebooks:
base-trd-nb1.ipynb: Demonstrates videoplayback of a file source in rootfs of the target to the Jupyter notebook using the GStreamer multimedia framework.
base-trd-nb2.ipynb: Demonstrates streaming video from a v4l2 device on the target to the Jupyter notebook using the GStreamer multimedia framework
base-trd-nb3.ipynb: Demonstrates streaming video from a v4l2 device on the target to a HDMI monitor using the GStreamer multimedia framework.
base-trd-nb4.ipynb: Demonstrates two simultaneous streaming pipelines, one from file source and another from a v4l2 device onto two individual planes of a HDMI monitor using the GStreamer multimedia framework.
base-trd-nb5.ipynb: Demonstrates streaming video from a Mulit-Camera FMC module on the target to a HDMI monitor using the GStreamer multimedia framework.
base-trd-nb6.ipynb: Demonstrates the 2D filter accelerator kernels, both the PL and the AIE versions, inserted into the video pipeline of notebook 2.
base-trd-nb7.ipynb: Demonstrates using both 2D filter accelerator kernels in a time-multiplexed fashion in a multi-branch pipeline.
base-trd-nb8.ipynb: Demonstrates streaming audio/video from a v4l2 device or a file to a HDMI monitor/speaker using the GStreamer multimedia framework.
Note: MIPI sources in the notebook is platform specific, User is allowed to choose “mipi” in platform 1 and “mipi_quad” in platform 2 respectively, otherwise an exception would be thrown.
To run the notebooks, follow the below steps:
On the left pane of the browser, 8 notebooks are available under the folder Base TRD.
Double click to open the notebook
Select ‘Kernel’ → ‘Restart Kernel and Run All Cells’ from the top menu bar to run the demo. For nb1, nb2 and nb6 scroll down to the end of the notebook to see the video output. For all other notebooks the video output is displayed on the monitor.
Click the rectangular icon to interrupt the kernel and stop the video stream.
Select ‘Kernel’ → ‘Shutdown Kernel’ → close the notebook tab and move to the next notebook.
Note1: Rerunning a TRD notebook may have spurious behavior, this is because software commands are designed to run in order and may not behave as expected in iterative runs when the kernel is not restarted. In practice, there are some command sequences that can be regressed, but the safest thing is to restart the Jupyter Notebook kernel with ‘kernel’ → ‘Interrupt kernel’ , followed by ‘kernel’ → ‘Restart Kernel and Run All Cells’ when rerunning a notebook.
Note2: Jupyter-setup is expected to be started with root privilages ‘sudo systemctl start jupyter-setup’. If not started with root permission for the first atempt, then post install scripts will fail to execute and dot may throw an error creating a graph in png format. This can be resolved by executing ‘dot -c’ at the linux prompt .