This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA4VM: TDA4

Part Number: TDA4VM


I am currently trying to stream a 4MP video sequence, using custom built gstreamer pipeline at 30Fps. I purchased the TDA4 board and deployed my code, to realise that the board doesn't support that level of performance.

The code simply calls "videotestsrc", and makes an attempt to stream the 4MP 30FPS sequence. The output has 1-1.8 FPS result, and "htop" shows 80-95% consumption of the both the cores (the board only has 2 cores).

1. I would like to ask if the TDA4 board is built to support such tasks, and if so, are there any specific steps I should take to better utilize the "power" of this board.

2. I am also not sure if the board has GPU, and how to use it. For my task, I am using the "compositor" element, to composite multiple "videotestsrc". Even if I use the GPU memory, I am not sure if gstreamer provides an element which can handle that.

Any comments or suggestions on this would be greatly appreciate. Thank you for your time and patience.

  • Hello,

    Yes, the TDA4VM has a GPU. However, you mentioned gstreamer, which is for codecs.

    Can you elaborate on the "compositor" element that you are using?

    Regards,

    Erick

  • Hello Erick. Thank you for your response.

    I am using the "compositor" element from gstreamer. Essentially, I want to combine the feed from 5 cameras, into one single frame. So, I take feed from 5 cameras in 1920x1080 resolution (2MP). I resize the last four feeds into half (960x540 res), and then stack them in a square grid of 2MP. This is then stacked horizontally to the first 2MP camera feed. 

    For development purpose, I can simply replace the "camera" (v4l2src) with videotestsrc. Attached is an image for reference.

  • Hi Shubham,

    I think you can use Mosaic Node/gStreamer plugin for downscaling and for merging them into a single frame.

    Regards,

    Brijesh