This is tough to watch. I think I understand where both @MajorPainTheCactus and @Azurfel are coming from. Azurfel is saying, why can’t Megatron match the exact R-G-B mix to create colour as an actual CRT that we’re trying to simulate and @MajorPainTheCactus is saying that it’s impossible to do so while still having accurate colours because the subpixel primaries are different for a CRT and a modern display, especially in HDR mode.
Somebody, please correct me if I’m wrong.
I’m thinking that if it was possible for the CRT Shader phosphor/Mask output simulation to match the actual CRT’s Phosphor mix and colour accuracy verbatim while in HDR Mode, that @MajorPainTheCactus would have implemented it already but it’s probably not so the current implementation is the best we can hope for and the best compromise at least for now.
@Azurfel, I hope you realize that you are opening an entire new layer of pedantry with the desire to have only the primary subpixels light up when the primary Phosphors light up and once we see something it cannot be unseen.
Are you saying that there was a point in time already, where the goal you seek was in fact achieved within Sony Megatron Colour Video Monitor or any other shader (while in HDR Mode) but then discarded or this is something new that should be strived for?
Do remember the challenges I showed you with the lack of purity of the “Phosphors” when using Colour Accurate Mode in HDR Mode on at least a few users’ LG WOLED displays, that didn’t and still don’t seem to occur on LCD Displays.
Also note that these have only been picked up by zoomed in photos of the display, which is one of the reasons the issue was so easy to identify in the first place since users were doing lots of photo comparisons to real CRTs in the early days of Sony Megatron Colour Video Monitor.
I think we should allow @MajorPainTheCactus to complete this update based on his vision and help him iron out any bugs within the framework of what he was trying to do, not necessarily what we think should be an improvement, at least for now. So a feature freeze. Get the thing stable and out the door, then do some more research and collaboration toward the next point or major update which might include a new mode or a better way of doing things.
I’m not sure if the examples you posted with the Sonic CRT photo for comparison show the same issue that the Mask Accurate Mode sought to correct.
In your Sonic example the image is still “Mask Accurate” I.e. it only consists of colours from within the simulated Phosphor/Mask structure and primaries. So only Red, Blue and Green from the mask that Sony Megatron created that looks like an Aperture Grille or Slot Mask.
The problem with the old Colour Accurate Mode was that the Mask was created in SDR Mode and was fine in SDR Mode, however when converted to HDR Colourspace and HDR Mode was enabled, additional subpixels, which were not a part of the original Phosphor/Mask primaries and structure were being lit at the display’s subpixel level, not just at the (emulated) Phosphor Level. So besides having intermediate colours which have no place in the mask, for example purple, yellow, orange imposing themselves on the Mask, this was like splashing random paint colours to deface a piece of artwork and destroyed or distorted the look, shape and structure of the CRT Phosphor Mask, we’re trying to recreate.
To me that defeats the entire purpose of subpixel accurate CRT emulation, which is why I have always been in the Mask Accurate camp and not in the Colour Accurate camp, at least with my LG 55OLEDE6P.
There could be various reasons why you never experienced or documented this anomaly. Maybe it’s because of the way your device is calibrated or differences in the panel design, OLED composition, software or tonemapping behaviour.
