I'm honestly surprised that this nut hasn't been cracked already.
A 320x240 image on an arcade monitor is comprised of 76,800 individual pixels.
Multiplying this value by 3, to represent the RGB components of the CRT dot mask, comes out to a 230,400 pixel image of individual RGB subpixels.
This doesn't account for blank scanlines, so multiplying this value by 2 in order to skip lines brings us to 460,800 pixels required for a "sharp" representation (no glowing effects).
A 1080p display has 2,073,600 pixels available. If my math is correct (it may not be) about 518,400 of those pixels are discarded due to the difference in aspect ratio. This still leaves 1,555,200 pixels for the 4:3 representation, or 3.375x what is required.
A 4K display obviously does much better with 6,220,800 for the 4:3 representation, or roughly 13.5x the number required for the sharp dot mask representation, leaving room for effects.
I would think that all which would be necessary are good shaders, extremely low latency OLED and variable refresh rates for a near perfect CRT replacement.
Or has my math failed me?
*edit* these 4:3 pixel numbers are probably wrong on the low side....I'm working on it
)
*edit 2* Okay, I think it's right now.....even better numbers than before!
*edit 3* The more I think about it, the more I believe that a shader with RGB Triad dot mask and glow effects is possible on a 4k display. Sheesh, that's a bunch of edits