OK so I have a real cool picture to share but I need to explain some stuff first.
OK SO: originally TV was done over CRTs. CRTs are analog devices made out of curved glass vacuum tubes, so they have interesting properties with how they draw pictures.

One of which is overscan: the problem of parts of the image being outside the viewable area.
Basically you can't assume all TVs are consistently manufactured and calibrated, so the whole screen probably won't be visible.

So there's a set of areas which were considered in TV production, including:
title safe: which is definitely viewable on all TVs
action safe: which is viewable on most TVs
Fullscan: the total amount broadcast, which might be viewable on some TVs

so if you are putting text on screen, it needs to be in title safe.
things which should be viewable but you can lose some edges? those go in action safe.
and fullscan is the total area you need to worry about: most viewers can't see all of it, but some can

so you gotta keep things like boom mics and stagehands out of fullscan, just in case someone sees it.

this was not always followed, so sometimes they are visible.

but let's think about the shape of a CRT tube (yes, that's two tubes. shut up)
It's really one big piece of glass, going from a thin neck at the back, funneling up to a rectangular-ish panel at the front.

because it's one big piece of glass it's not like an LCD where you can have real sharp right-angle corners.
it's naturally rounded, a bit.

But look at a CRT TV or monitor (at least after the 50s or maybe 60s) and it'll look rectangular.
So how's that work?
SIMPLE! it's a bezel. It's just a cover hiding part of the tube.

So over the tube itself there's a piece of plastic (or wood/metal) that makes it a bit more rectangular looking.
This both helps hide any mistakes in that fullscan area and it also makes the TV look better, since it's more clearly rectangular, instead of very rounded.

but the tube is still drawing in those parts hidden by the bezel, because it doesn't know not to.

Tubes are not smart. All they are is an electron beam that's going left to right, and top to bottom, at a certain frequency.
The intensity of the beam is varying, so it goes up for bright parts and down for dark parts.

and the signal that's being sent has to include the parts where it's moving back to the left and from the top to the bottom
there's no smarts in there to say "ok hang on I'll stop drawing pixels now while I reset the beam". NOPE, the signal has to go black for a while.

and this all being analog is why adjustments had to be made, and later TVs had this as little dials, earlier TVs you had to get out a screwdriver and adjust potentiometers on the PCB.

because get the voltages a little off and it's too wide, too tall, too left, too up, it rolls, etc.
And all of these are parameters that can vary from TV to TV and as tubes (and other components) age and warm up.
So there was no exactness of TV calibration.
Thus, overscan.

which wasn't just an analog broadcast TV thing.
You'd sometimes see this in digitally generated pictures, like from the NES: Mario 3 famously has attribute glitching on the right side of the screen, but many CRT TVs would have hidden this from the player.

OKAY so now that we know about how overscan works, first we have to think about different shapes of TVs.
In the late CRT era we'd basically standardized on 4:3 rectangles for TV/Monitors but there's no reason CRTs have to be that shape.

and a lot of early ones were not that shape.
After all, making a glass rectangle is not an easy thing to do. You have to blow this as glass, remember?
So a circle is much easier.
And a lot of early displays were circular, like this radarscope.

And in fact one of the first games ever made, Tennis for Two, was played with an oscilloscope display, which was of course circular.

now a lot of these circular displays are not standard rectangular raster-scan displays, they're instead polar/vector displays, and there's some complexity there that I don't want to go into.

But imagine you've got a circular display: what does the overscan look like?

it ends up looking, for lack of a better diagram, like the flag of japan.
You've got a circle in the middle that is visible, and a big area outside it that isn't.

But here's the fun bit: Remember how I mentioned bezels just hide part of the display? They don't stop the screen from drawing.
The tube doesn't "know" there's parts of the raster scan that it can't light up. It does anyway, you just can't see them.

So imagine you have a circular tube, and you take off the bezel, and then you pump a standard rectangular picture into it, in a dark room where you can clearly see everywhere that gets lit up.
What would that look like?
WELL, HOLD ON TO YOUR HATS

This here is an 8" RCA TM-9AN B&W display. They were used in TV production back in the 50s, and this one has the bezel removed, and already there's hints of how awesome the next picture will be..

GET READY

You turn off the lights and MY GOD IT'S BEAUTIFUL

the rectangular image is WRAPPING AROUND THE OUTSIDE OF THE TUBE!

So yeah, that's the awesome picture I wanted to share but had to ramble for like 50 posts before I could get to it

So yeah, these pictures of the round tube and the info on it are from this site:
bunkerofdoom.com/crt10sp4/index…

and if you want more info on how analog TV/CRTs worked, @TechConnectify has a great series of videos:

mentions@TechConnectify One question I have though is: why is it lighting up there?
CRTs light up because an electron beam excites a phosphor. It's not like an electron beam is a flashlight, making light anywhere it hits. No phosphor = no light.

mentions@TechConnectify So there's got to be phosphor there, and there really shouldn't be. It's not visible.

So my guess is whatever process they used to coat the front of the tube (on the inside) with phosphor caused nearby areas to get coated too, and that wasn't considered a problem.

mentions@TechConnectify one random side-note about an alternative way to do CRT monitors: VECTOR DISPLAYS!
Vector displays are the same basic technology (an electron beam exciting phosphors) but they're not controlled in the same way.

mentions@TechConnectify So you can't aim the electron gun in a CRT display, it's static.
So instead you aim the electron beam, using magnetic fields.
A deflection yoke is an electromagnet wrapped around the neck of the tube and it creates a magnetic field that'll bend the beam

mentions@TechConnectify You have one set of electromagnets that bend it up and down, and another set that bends it left and right.
So you can position it anywhere on the screen, and you can adjust the intensity.

mentions@TechConnectify And like I said before, you want the beam to move from left ot right, top to bottom.
You do that using a flyback transformer.

mentions@TechConnectify You have to go top to bottom at about 60hz (in NTSC), but to go left to right you have to do it for every line of the picture drawn, so it's very very fast. It has to draw 262.5 scan lines, and it has to do it 60 times a second.
So the left-to-right oscillation is 15 kilohertz.

mentions@TechConnectify This, by the way, is why CRT TVs make that high pitched whine noise.
It's switching a magnet on and off at 15,000 times a second, and that causes vibration through magnetostriction

mentions@TechConnectify And the human hearing range is something like 20-20khz, but that rapidly shrinks as you age, starting as young as eight.
So by the time you're 40, most people can't hear 15khz anymore.

mentions@TechConnectify in any case, vector displays:
So what if instead of having the display moving left-to-right, top-to-bottom, and just varying the intensity, you instead gave the computer full control over where it could move?

mentions@TechConnectify So instead of careful timing and one analog signal, you now have 3 analog signals.
intensity, X, and Y.
So now it can move it around and draw whereever.
The result is a vector display.

mentions@TechConnectify There are no pixels here, there's no resolution.
The only limiting factors are how fast the computer can tell it to move, how fast the magnetic fields can change, and the quality of the digital-to-analog-converters in the system.

mentions@TechConnectify so if you turn off the beam and move it somewhere, then turn it on, you get a dot.
if you leave it turned on when you move, you get a line.
So the world can be built up of lines and dots, instead of pixels.

mentions@TechConnectify This was inherently more expensive than a standard CRT monitor but it definitely had its uses. Oscilloscopes traditionally used them, and they made their way into a number of arcade games, like Tempest (shown above), Asteroids, and Battlezone.

mentions@TechConnectify and of course the Vectrex home console.

mentions@TechConnectify An interesting thing about vector graphics is the limitation they have with regards to time:
You can only draw so many points/lines before they start flickering, and it turns out this is because you're taking too long to redraw them.

mentions@TechConnectify To explain that, let's back to regular CRTs.
There's two effects causing you to see a picture:
1. Phosphor persistence
2. Persistence of vision

mentions@TechConnectify phosphor persistence is the effect where an excited phosphor doesn't instantly go black when the electron beam is no longer hitting it. It instead fades back to black, at some speed determined by the materials it's made of.

mentions@TechConnectify Normal CRT displays need to be refreshed about 30 times a second, in other words they fade to black in about 33 milliseconds.
You want it to be fast so you don't get smearing, but not so fast you have to refresh it faster than the signal can handle.

mentions@TechConnectify A fun side-note: Tektronix made some displays which would had persistence measured in MINUTES. These were used for Storage Tubes, which are a complex subject I don't want to get into now.
But they had a thing where they could force a clear by flashing it with a special color

mentions@TechConnectify and persistence of vision is the other effect used by CRTs to display an image:
This one's in your eyes & brain. If you've ever moved a sparkler around and seen a line instead of a bright dot, you know what this one is.

mentions@TechConnectify it's caused by both chemistry in your eye and re-amplification in your neurons, but the tl;dr version is that you continue to "see" a bright thing for a little bit after you stop actually getting photons from it hitting your eye.

mentions@TechConnectify So for a regular CRT, they have the phosphors set up to decay at the right rate so they'll be "dark" by the time the beam comes around again to re-draw it. So you just redraw the screen at 60hz (or 30hz) or whatever and it's fine.

mentions@TechConnectify you can't "draw too much", really. Either you have a framebuffer and trying to draw "too much" causes the last frame to be drawn again, or you don't and it causes nothing to be drawn.
It's going to happen every 30/60/70/whatever hertz whether you like it or not.

mentions@TechConnectify because it's an intrinsic property of how the transformers in the CRT are set up.
But vector displays work differently. Instead of the TV generating its own raster scan pattern and just doing it automatically forever, the pattern is under control of the computer.

mentions@TechConnectify So instead of refreshing the screen every 33 milliseconds like clockwork, you instead "refresh" it at a rate entirely determined by how fast your drawing happens.

mentions@TechConnectify so if you try to draw a few lines/dots very fast and don't wait, you'll get afterimages because when you draw frame 2, frame 1 hasn't finished fading away.

mentions@TechConnectify and if you do the opposite and take too long drawing frame 1, it means parts of frame 1 have already started fading away before you can redraw them for frame 2

mentions@TechConnectify So the end result is that if you put too many lines or dots on a vector display, you get flicker. The phosphors can't maintain the image long enough, and the persistence of vision in your visual system doesn't maintain the image long enough.

mentions@TechConnectify But this is entirely dependent on what phosphors you're using, and the contrast level in the room! In a bright room, persistence of vision doesn't last as long. in a dark room, it lasts longer.

mentions@TechConnectify And phosphors that go dark faster mean you can draw more... but now you're limited by how fast you can move things without smearing, because the phosphors are going to persist whether you want them to or not.

mentions@TechConnectify And if you get faster-darkening phosphors, you can avoid that kind of afterimage, but now you have less time to draw before it starts flickering.

It's an interesting trade-off and shows a completely different way to think about how you'd manage those systems.

mentions@TechConnectify Like as a games programmer you're always thinking about framerate, but you're thinking about it from the perspective of a CRT/LCD with a framebuffer.
If you take too long drawing, your framerate goes down, and it feels less smooth.

mentions@TechConnectify But with a vector display, your framerate is changing all the time as the number of lines/dots changes, and not only does that affect the smoothness, but your display gets darker and flickerier as your take longer and longer.

mentions@TechConnectify A final fun fact and then I'll shut up:
Phosphors are named after the element phosphorus (obviously)
But phosphorus is not a phosphor!

mentions@TechConnectify Phosphorus glows when exposed to oxygen, and it's this property after which phosphors were named.
But phosphors are materials which exhibit phosphorescence or fluorescence, which are types of photoluminescence.
Basically they absorb photos and re-emit them

mentions@TechConnectify with fluorescence it's immediate (or nearly so: scales of nanoseconds). You hit them with photons and photons come out.
With phosphorescence it's slower, and takes milliseconds.

mentions@TechConnectify but the important part is that it's photons in, photons out.
This is why CRTs work: The electron beam hits the phosphors, they then emit visible photons and slowly fade back to black.

mentions@TechConnectify but they're named after phosphorus, which DOESN'T DO THIS.
It instead has chemiluminescence: a chemical reaction that gives off photons (but without any photons needed to excite it)

mentions@TechConnectify I love that sort of thing in science, where an effect or property is named after something that doesn't actually demonstrate it.

mentions@TechConnectify Anyway, that was a long thread and as always, if you want me to continue this sort of rambling, feel free to support me on ko-fi:
ko-fi.com/fooneturing

mentions@TechConnectify or patreon:
patreon.com/foone

mentions@TechConnectify and sorry to Technology Connections for tagging you on half the thread.
The new twitter interface doesn't seem to let me untag people sometimes, or I would have.