Create video+audio frames for broadcasting over RTMP? Linux+Python preferably

Question

What technologies should be used?

I would want to write my own algo for this using libs available.

I am a CS student. Having knowledge of animation and front+backend development. Linux and Python, PHP, etc.

The link here is using APIs, yes. BUT how they are merging real-time data with visuals.

I want to work on similar real-time UI. Can we use HTML+CS+JS here? Or would we use OpenCL or Tkinter type of a thing?

What should be the flow if I want to broadcast it over Youtube. How much bandwidth and processing power does these things take? Thanks in advance.

Answer 1

You can do most, if not all, of what you are asking for with MoviePy, which under the hood uses FFMPEG and several other libaries. But under Limitations is the words "MoviePy cannot (yet) stream videos" - however you can use the same libraries to:

• Take a copy of a static image
• Overlay some text, (done with ImageMagick in MoviePy but there are other options)
• Generate a video stream from the current image and merge it with a audio stream (FFMPEG)
• Stream it out over RTMP with FFMPEG

Obviously you will need a reasonable specification machine to do all of the above but you may be pleasantly surprised at the required bandwidth as, since the video data is not changing rapidly it should be reasonably low bandwidth & you may find that the upload speed provided by a good broadband connection to be sufficient.

While all of the tools mentioned are cross platform as well as Gratis & Open Source I would not recommend trying something like this from a Windows machine as you do not have the necessary control over what other tasks might slow the performance. If the audio content is known in advance I would suggest looking into using something like an AWS instance as the streaming source.