Is There Any Potential For A Raster Based (Instead of Sub-Frames) CRT Beam Sim?

[FENIKS]

While this is mostly a question for Chief Mark, everyone is free to share their thoughts.

To be clear, I have very little understanding of the technical details, so this is probably pie-in-the-sky/wishful thinking, but I've been wondering if it's at all feasible to have an implementation of CRT Beam Simulation that doesn't rely on shader sub-frames (necessitating excessive refresh-rate headroom for best results)?

If such a thing could ever work, I'd imagine it would only be suitable for emissive displays like OLED, so I'm always loosely thinking about the possibilities and limitations. For example, it's been explained that modern OLED panels (after the LG CX & C1) aren't capable of sub-refresh BFI because they lack the electronics to turn pixels on AND off within a refresh cycle (at least I think that's how it works). But I often wonder, is it possible for OLEDs to change the luminance of pixels within a refresh cycle, without turning them off? And if they could, would it be possible to dim the pixels to a near-black (but still active) state, similar to how Plasmas would display near-black/very dim versions of a frame with it's sub-field drive and still achieve low MPRT? Essentially what runs through my head is this: lines of pixels that are at full luminance at (and just after) the raster line and as the pixel stays on screen within that refresh cycle, it slowly dims to a state where it is very dim (but still on) to the point where it doesn't contribute to eye-tracking/persistence blur.

Could something along these lines be achievable on recent OLEDs, at least in theory? Very interested in hearing the thoughts of Chief and everyone else.

Thank you.

[RealNC]

Such OLED displays do exist actually. Or did exist. They use rolling scan, so output is very similar to a CRT. But they don't display one scanline at a time, but rather a scan bar that is dozens of lines tall. They could only display one scanline if they wanted, but the issue is that this would the image so dim, it's unusable. What you suggest would be a tweak to the rolling scan implementation and sounds feasible to me.

Anyway, I've seen such an OLED display years ago. OLED rolling scan seems to have fallen out of favor now so I don't think there's any consumer OLED displays anymore that have it. I could be wrong though.

[FENIKS]

Such OLED displays do exist actually. Or did exist. They use rolling scan, so output is very similar to a CRT. But they don't display one scanline at a time, but rather a scan bar that is dozens of lines tall. They could only display one scanline if they wanted, but the issue is that this would the image so dim, it's unusable. What you suggest would be a tweak to the rolling scan implementation and sounds feasible to me.

You're definitely right about the brightness levels. If it's true that excited CRT phosphors could briefly output ~30,000 nits, only to drop off for the average brightness level to be ~100 nits, then nothing but maybe MicroLED could hope to replicate it. However, I'd imagine that it wouldn't necessarily have to be that granular (or bright) to be useful on modern OLEDs. For instance, you mentioned a rolling "scan bar that is dozens of lines tall". What if that scan bar was just tall enough to take advantage of the 10% APL brightness algorithms in modern displays, with a gradual fade out to near-black (but again, the pixels would still be active as to cheat the ON/OFF problem within 1 refresh cycle). Of course, the ideal would be to have the height of the "scan bar" be user adjustable for the clarity/brightness trade-off.

Anyway, I've seen such an OLED display years ago. OLED rolling scan seems to have fallen out of favor now so I don't think there's any consumer OLED displays anymore that have it. I could be wrong though.

I wouldn't expect display manufacturers to help us out on this one. My hope would be that something like this could be implemented in software, or hardware in the form of "box-in-the-middle" solutions as Chief calls it (like the RT4K and OSSC Pro). Like I mentioned in the OP, I suppose the lynchpin for something like this to work would be whether or not pixel luminance can be addressed/changed within a single refresh cycle, and whether or not software (or external hardware) can exert that level of control.