There's a synchronous and instantaneous nature you don't find in modern designs.
The image is not stored at any point. The receiver and the transmitter are part of the same electric circuit in a certain sense. It's a virtual circuit but the entire thing - transmitter and receiving unit alike - are oscillating in unison driven by a single clock.
The image is never entirely realized as a complete thing, either. While slow phosphor tubes do display a static image, most CRT systems used extremely fast phosphors; they release the majority of the light within a millisecond of the beam hitting them. If you take a really fast exposure of a CRT display (say 1/100,000th of a second) you don't see the whole image on the photograph - only the most recently few drawn lines glow. The image as a whole never exists at the same time. It exists only in the persistence of vision.
It doesn’t begin at the transmitter either, in the earliest days even the camera was essentially part of the same circuit. Yes, the concept of filming a show and showing the film over the air existed eventually, but before that (and even after that, for live programming) the camera would scan the subject image (actors, etc) line-by-line and down a wire to the transmitter which would send it straight to your TV and into the electron beam.
In fact in order to show a feed of only text/logos/etc in the earlier days, they would literally just point the camera at a physical object (like letters on a paper, etc) and broadcast from the camera directly. There wasn’t really any other way to do it.
>>> The image is not stored at any point.
The very first computers (Manchester baby) used CRTs as memory - the ones and zeros were bright spots on a “mesh” and the electric charge on the mesh was read and resent back to the crt to keep the ram fresh (a sorta self refreshing ram)
Yeah it super weird that while we struggle with latency in the digital world, storing anything for any amount of time is an almost impossible challenge in the analog world.
> The image is not stored at any point.
Just wanted to add one thing, not as a correction but just because I learned it recently and find it fascinating. PAL televisions (the color TV standard in Europe) actually do store one full horizontal scanline at a time, before any of it is drawn on the screen. This is due to a clever encoding used in this format where the TV actually needs to average two successive scan lines (phase-shifted compared to each other) to draw them. Supposedly this cancels out some forms of distortion. It is quite fascinating this was even possible with analogue technology. The line is stored in a delay line for 64 microseconds. See e.g.: https://www.youtube.com/watch?v=bsk4WWtRx6M