Qt World Summit: Register Today!

Video Rendering Application with Overlay

  • Hi,

    I'm developing an application for static video analysis of multiple uncompressed video footage (5GB+ per video) in parallel, while displaying the extracted information as an overlay on top of the composed video footage.

    This software already exists with some hacky OpenCV and a yaml file for configuration. I'm looking forward to porting it to Qt with a GUI for a better configuration management and possible rendering previews.

    I'm after reading some tutorials on QML I still do not understand the benefits of using QML, or if it's even applicable to my application. The current "problem" is displaying multiple plots and other changing information in a OpenGL based view and then exporting these frames in a h264 encoded video container.

    I know that it is possible via basic OpenCV, with the development cost of a OpenCV specific plotting library, but I've hoped to leverage the Qt ecosystem for this task, as the cross-platfrom development with ffmpeg/gstreamer/opencv is already cumbersome.

  • you need QML only for UI part such as main window, UI elements etc
    all your backend code should be concentrated around QAbstractVideoFilter subclass, where you can use opencv or whatever you want

  • Thank you for your answer! I've yet to find a good introduction into QML for C++ developers, but I was a little put off by the somewhat convoluted approach as different tutorials either don't show any QML or only use it soley to develop a very simple application.

    The main problem I'm struggling is with QML is the ability to generate a GUI based on some set preferences. In my case the amount of e.g tabs in the "Options" pane should be defined by the amount of video files which are currently loaded. Is this easily possible to programatically extend QML GUIs?