Lawfer wrote:As Fudoh once mentioned here most CRTs were basically made for 480i as the main resolution, which follows the DVD standard at the time 720x480, however 720x480 is technically not 4:3, but actually 3:2 and 3:2 has a higher length than 4:3, 320x240 and 640x480 are both 4:3.
Here is your first misconception: A CRT does not have a notion of discrete "pixels" in the horizontal direction, it only has discrete lines in the vertical direction.
The analog TV standards date from a time when the pixel had not been invented yet, so there are no discrete pixels in an analog TV signal - instead, the length of a line is defined by the time between the end of one horizontal sync pulse and the start of the next one (ignoring blanking for simplicity). During this time, the electron beam of a CRT simply sweeps from left to right over the entire width of the tube and the video signal can change its level to indicate how intense the electron beam should be at this time, resulting in a brighter or darker spot on screen (also ignoring color for simplicity). There is no requirement for the analog signal to change its level at some discrete time within a line, instead it can change at any time since it's analog. There are limitations on the fastest possible brightness changes within a line, but I'm currently struggling to come up with a way to explain that without invoking differentials or frequency-domain considerations.
The digital video resolutions you keep referring to divide this horizontal line length into discrete time intervals and allow brightness changes only at the boundaries of these time intervals. A CRT doesn't care about this distinction however - it just sees the timing of the horizontal syncs and adjusts its electron beam sweep speed so it covers the entire width of the tube. With identical timing, the CRT sees absolutely no difference between a signal with 640 or 720 horizontal pixels - it's just that in one case the signal is able to change a bit more often during a line than in the other case.
You may have noticed that the entire discussion above is just about a single horizontal line, without any consideration for the other lines in a field/frame. This is because the number of vertical lines in a video signal does not influence the horizontal sweep range of a CRT. Your attempts to calculate an aspect ratio from a pixel count assumes that there is such a relation, but there actually is not.
The square pixels you assume are a relatively recent thing - they simplify some things, but there is no technical requirement for pixels to be square. In the PC space, most graphics modes used non-4:3 pixel counts (VGA's 640x480 is one of the earlier examples of a 4:3 pixel count mode), but basically all of them were meant to be displayed full-screen without borders on a 4:3 screen - even ones like the 720x348 of a Hercules card or 640x200 from CGA. The Commodore Amiga even supports a hires mode with 1280x240 with TV-compatible timings which was meant for display in a 4:3 aspect ratio, resulting in very rectangular pixels.
(Fun fact: Plasma screens with a 16:9 display aspect ratio and 1024x768 pixel resolution exist. Feeding them a 720p signal results in no black borders.)