video streaming does not consist of encoding frames to JPEG and sending them. that’s the lowest form of compression. you should look into WebRTC. that’s for live video streaming.
by the way, HTTP as well as websockets are binary-safe, and HTTP requests and responses can be multi-part, so you can deliver both text and binary data in the same request/response.
I won’t entertain hypotheticals. you need to investigate first. then you can present your quantitive findings that suggest spending time improving specific parts of your situation. then you can ask for others to spend their brain time solving actual issues.